阅读说明：本技术 基于网络分发历史从数字媒体的储存库生成对应于预测的分发类别的数字媒体群集 (Generating digital media clusters corresponding to predicted distribution categories from a repository of digital media based on network distribution history ) 是由 赵跃 王宇 威廉·塞谬尔·贝利 马里·谢布莱 于 2020-03-30 设计创作，主要内容包括：本公开涉及用于基于网络分发历史从数字媒体的储存库中生成对应于预测的分发类别的准确数字媒体群集的系统、非暂时性计算机可读介质和方法。例如,数字媒体聚类系统可以(基于用户的网络账户的网络分发历史)在远程服务器处应用机器学习模型,从而为未来的电子通信生成预测的分发类别。远程服务器可以向用户客户端设备提供预测的分发类别,用于对存储在客户端设备处的数字媒体进行安全的本地分析。基于预测的分发类别和存储的数字媒体,客户端设备可以建议数字媒体项经由网络系统进行分发。因此,所公开的系统可以在不向远程服务器提供关于存储在客户端设备处的数字媒体项的任何信息的情况下展示数字媒体内容。(The present disclosure relates to systems, non-transitory computer-readable media, and methods for generating accurate digital media clusters corresponding to predicted distribution categories from a repository of digital media based on network distribution history. For example, the digital media clustering system may apply a machine learning model at a remote server (based on the network distribution history of the user's network account) to generate predicted distribution categories for future electronic communications. The remote server may provide the predicted distribution category to the user client device for secure local analysis of digital media stored at the client device. Based on the predicted distribution category and the stored digital media, the client device may suggest that the digital media item be distributed via the network system. Thus, the disclosed system can present digital media content without providing any information to a remote server regarding digital media items stored at a client device.)

1. A system, comprising:

at least one processor; and

non-transitory computer memory including instructions that, when executed by the at least one processor, cause the system to:

identifying a network system profile, the network system profile including a network distribution history corresponding to a network account of a user;

generating, based on the network distribution history, a predicted distribution category for future electronic communications corresponding to the network account of the user using a distribution classification model; and

providing the predicted distribution category to a client device of the user such that the client device provides, for display, a digital media cluster corresponding to the predicted distribution category from a repository of digital media items on the client device.

2. The system of claim 1, further comprising instructions that when executed by the at least one processor cause the system to: generating the predicted distribution category and providing the predicted distribution category to the client device without receiving information from the client device regarding a repository of digital media items on the client device.

3. The system of claim 1, further comprising instructions that when executed by the at least one processor cause the system to:

receiving, from the client device, a user indication of a digital media item from the digital media cluster; and

in response to receiving the user indication of the digital media item, distributing the digital media item to a plurality of client devices of a plurality of users via the network system.

4. The system of claim 1, further comprising instructions that when executed by the at least one processor cause the system to:

identifying a plurality of short-lived digital content items provided by the client device;

generating a suggested short-time digital content item from the plurality of short-time content items; and

providing the suggested short-time digital content item to the client device for display with the digital media cluster corresponding to the predicted distribution category from a repository of digital media items on the client device.

5. The system of claim 1, further comprising instructions that when executed by the at least one processor cause the system to:

identifying the network distribution history by identifying a plurality of historical posts of a social networking system, wherein the plurality of historical posts comprise a plurality of historical digital media items; and

generating the predicted distribution category from a plurality of distribution categories based on the plurality of historical digital media items.

6. The system of claim 1, further comprising instructions that when executed by the at least one processor cause the system to:

identifying an additional network account of the user corresponding to the network system;

determining an additional network distribution history corresponding to an additional network account of the user corresponding to the network system; and

generating, using the distribution classification model, additional predicted distribution categories for future electronic communications corresponding to the additional network accounts of the user based on the additional network distribution history.

7. The system of claim 6, further comprising instructions that when executed by the at least one processor cause the system to: providing the additional predicted distribution categories corresponding to the additional network accounts of the user to the client device of the user such that the client device provides, for display, additional digital media clusters corresponding to the additional predicted distribution categories from a repository of digital media items on the client device.

8. The system of claim 1, further comprising instructions that when executed by the at least one processor cause the system to:

determining suggested digital media style modifications based on the network distribution history; and

providing the suggested digital media style modification to the client device of the user such that the client device modifies a set of digital media items with the suggested digital media style modification and provides the set of digital media items for display in relation to the digital media cluster.

9. A non-transitory computer-readable storage medium comprising instructions that, when executed by at least one processor, cause a computer system to:

generating a plurality of digital media clusters from a digital media repository on a client device of a user having a network account of a network system;

identifying a predicted distribution category of future posts of the network account of the user, the predicted distribution category based on a network distribution history of the network account;

selecting a digital media cluster from the plurality of digital media clusters using the predicted distribution category;

providing, via a user interface of the client device, the digital media cluster for display; and

providing the digital media item to one or more remote servers for distribution of the digital media item to client devices of other users of the network system based on user interaction with the digital media item from the digital media cluster.

10. The non-transitory computer-readable storage medium of claim 9, further comprising instructions that, when executed by the at least one processor, cause the computer system to: identifying the predicted distribution category by receiving the predicted distribution category from the one or more remote servers without transmitting information about the digital media repository on the client device to the one or more remote servers.

11. The non-transitory computer-readable storage medium of claim 9, wherein the plurality of digital media clusters correspond to a plurality of distribution categories, and further comprising instructions that, when executed by the at least one processor, cause the computer system to: selecting a digital media cluster from the plurality of digital media clusters by comparing the predicted distribution category to the plurality of digital media clusters.

12. The non-transitory computer-readable storage medium of claim 9, further comprising instructions that, when executed by the at least one processor, cause the computer system to generate the plurality of digital media clusters by:

generating a first cluster corresponding to a first set of digital media items from the digital media repository based on time;

generating a second cluster corresponding to a second set of digital media items from the digital media repository based on location; and

generating a third cluster based on the one or more public individuals identified in the third set of digital media items from the digital media repository using a facial recognition algorithm.

13. The non-transitory computer-readable storage medium of claim 9, further comprising instructions that, when executed by the at least one processor, cause the computer system to:

identifying a suggested digital media pattern modification, wherein the suggested digital media pattern is based on the network distribution history;

modifying a set of digital media items by applying the suggested digital media style; and

the set of digital media items is provided for display in relation to the digital media cluster.

14. The non-transitory computer-readable storage medium of claim 9, further comprising instructions that, when executed by the at least one processor, cause the computer system to:

identifying an additional network account of the user corresponding to the network system;

identifying an additional predicted distribution category for future electronic communications corresponding to the additional network account of the user, the additional predicted distribution category based on an additional network distribution history for the additional network account; and

selecting an additional digital media cluster from the plurality of digital media clusters using the additional predicted distribution category.

15. The non-transitory computer-readable storage medium of claim 9, further comprising instructions that, when executed by the at least one processor, cause the computer system to:

identifying a suggested short-term digital content item from a plurality of short-term digital content items previously distributed from the network account of the user; and

providing the suggested short-time digital content item for display with the digital media cluster.

16. A computer-implemented method, comprising:

identifying a network system profile comprising a network distribution history corresponding to a network account of a user;

generating, based on the network distribution history, a predicted distribution category for future electronic communications corresponding to the network account of the user using a distribution classification model; and

providing the predicted distribution category to a client device of the user such that the client device provides, for display, a digital media cluster corresponding to the predicted distribution category from a repository of digital media items on the client device.

17. The computer-implemented method of claim 16, further comprising generating the predicted distribution category and providing the predicted distribution category to the client device without receiving information from the client device regarding a repository of digital media items on the client device.

18. The computer-implemented method of claim 16, further comprising:

receiving, from the client device, a user indication of a digital media item from the digital media cluster; and

in response to receiving the user indication of the digital media item, distributing the digital media item to a plurality of client devices of a plurality of users via the network system.

19. The computer-implemented method of claim 16, further comprising:

identifying the network distribution history by identifying a plurality of historical posts of a social networking system, wherein the plurality of historical posts comprise a plurality of historical digital media items; and

generating the predicted distribution category from a plurality of distribution categories based on the plurality of historical digital media items.

20. The computer-implemented method of claim 16, further comprising:

identifying a plurality of short-lived digital content items corresponding to the network account;

generating a suggested short-time digital content item from the plurality of short-time digital content items; and

providing the suggested short-time digital content item to the client device for display with the digital media cluster corresponding to the predicted distribution category from a repository of digital media items on the client device.

Background

In recent years, significant technological advances have occurred in hardware and software platforms for transferring digital content stored in different media repositories (reposititors) corresponding to a plurality of client devices over a computer network. For example, conventional digital content delivery systems may select and transmit digital images or digital videos stored at client devices to multiple recipient client devices along with various electronic messages (e.g., social networking posts or instant messaging). In addition, conventional digital content delivery systems may organize digital media repositories (e.g., based on date) to allow previously captured digital files to be selected and delivered to other client devices (e.g., as part of electronic messages tailored to other client device users).

Despite these advances, conventional digital content transmission systems still suffer from a number of problems in terms of efficiency, accuracy, security, and operational flexibility. As one example, conventional systems are inefficient in that they require an excessive amount of interaction with various user interfaces to identify and transfer digital content from a large digital media repository. In fact, it is now common for personal client devices to store thousands of digital media items (digital media items). Identifying particular digital content items from such repositories using current digital content delivery systems requires a significant amount of time and user interaction to sort, scroll through, and view the individual digital files.

In addition, conventional digital content transmission systems are also inaccurate. In fact, conventional digital content transmission systems often identify and transmit digital content that does not correspond exactly to the user's intent and/or corresponding electronic message. In fact, users of client devices often seek to generate electronic communications conveying particular concepts, and conventional systems often result in inaccurate misalignments between the identified digital content and the particular concepts conveyed in the digital messages. In some cases, inaccuracies and inefficiencies of conventional systems result in client device users abandoning efforts to generate and transmit electronic messages (and/or leaving digital content transmission systems as an alternative system).

Conventional systems also have deficiencies with respect to digital security. In fact, many conventional digital content transmission systems operate on remote servers. Thus, conventional systems access and analyze digital content stored at various client devices through these remote servers. However, in recent years, the number of digital attacks and security holes has steadily increased on servers that maintain client information. Thus, conventional systems may introduce digital privacy and security issues with respect to sensitive client information accessed by a remote server on a client device.

In addition, conventional systems are not flexible. For example, conventional digital content delivery systems provide a rigid way to identify and deliver digital content to users. To illustrate, conventional digital content delivery systems provide the same classification and analysis tools (e.g., by date) for all users, all accounts, and all client devices. This rigorous approach further exacerbates the inaccuracy and inefficiency of the conventional systems discussed above.

These and other problems exist with conventional digital content transmission systems.

SUMMARY

Embodiments of the present disclosure provide benefits and/or address one or more of the foregoing or other problems for systems, non-transitory computer-readable media, and methods for generating an accurate digital media cluster (cluster) corresponding to a predicted distribution class (distribution class) from a repository of digital media based on a network distribution history. For example, the disclosed system may expose (surface) digital media that accurately depicts the moment a user attempts to distribute to other client devices. In particular, the disclosed system utilizes a machine learning model to identify digital content suggestions that accurately identify what digital media items (and styles) a client device user seeks to share based on network distribution history.

Further, in one or more embodiments, the disclosed system improves privacy and security by performing an analysis of the network distribution history at a remote server, analyzing a private digital media repository at a client device, and then aligning the two analyses at the client device to expose accurate digital media suggestions. For example, the disclosed system may apply a machine learning model (based on previously shared information) at a remote server to generate a predicted distribution category. The remote server may provide a predicted distribution category for secure local analysis of digital media stored at the client device. Thus, the disclosed system can present digital media content without providing information about the digital media item to a remote server or overloading computing resources at the client device. In this manner, the disclosed system may reduce the time and interaction barriers to sharing digital content, accurately present digital media, improve security and user privacy, and flexibly customize digital content suggestions for particular users and accounts.

Additional features and advantages of one or more embodiments of the disclosure will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of such exemplary embodiments.

Brief Description of Drawings

The detailed description provides one or more embodiments with additional details through the use of the accompanying drawings, as briefly described below.

FIG. 1 illustrates a schematic diagram of an environment in which a digital media clustering system (digital media clustering system) operates, in accordance with one or more embodiments.

FIG. 2 illustrates a schematic diagram of generating suggested digital media clusters in accordance with one or more embodiments.

FIG. 3 shows a diagram of generating suggested digital media clusters in accordance with one or more embodiments.

FIG. 4 illustrates a diagram of generating suggested ephemeral digital content in accordance with one or more embodiments.

FIG. 5 shows an illustration of providing suggested short-term digital content and other suggested digital media items for display in accordance with one or more embodiments.

FIG. 6 illustrates training a distribution classification model to generate predicted distribution categories in accordance with one or more embodiments.

FIG. 7 illustrates a computing device having a user interface depicting digital media clusters and digital media items in accordance with one or more embodiments.

Fig. 8A-8B illustrate a computing device having a user interface depicting digital media aggregation and short-lived digital content in accordance with one or more embodiments.

9A-9C illustrate a computing device having a user interface for selecting digital media items, drafting electronic messages, and distributing digital media items and electronic messages in accordance with one or more embodiments.

FIG. 10 illustrates a block diagram of a digital media clustering system in accordance with one or more embodiments.

FIG. 11 illustrates a flow diagram of a series of acts for generating and providing a predicted distribution category based on network distribution history in accordance with one or more embodiments.

FIG. 12 illustrates a flow diagram of a series of acts for generating a digital media cluster based on a predicted distribution category in accordance with one or more embodiments.

FIG. 13 illustrates a block diagram of an example computing device for implementing one or more embodiments of the present disclosure.

FIG. 14 illustrates an example network environment for a network system (networking system) in accordance with one or more embodiments.

FIG. 15 illustrates a social graph in accordance with one or more embodiments.

Detailed Description

One or more embodiments of the present disclosure include a digital media clustering system that generates digital media clusters corresponding to predicted distribution categories from a digital media repository based on network distribution history. For example, the digital media clustering system may analyze public information (e.g., previous network posts) at the server while securely analyzing the digital media repository at the client device. The digital media clustering system may combine the analysis performed at the server with the analysis performed at the client device to effectively expose suggested digital media clusters without communicating information about sensitive client digital media from the client device.

For example, in one or more embodiments, the digital media clustering system identifies (at a remote server) a network system profile that includes a network distribution history corresponding to a network account of a user. Based on the network distribution history, the digital media clustering system may utilize a distribution classification model (e.g., a machine learning model such as a neural network) to generate predicted distribution categories for future electronic communications corresponding to the user's network account. Further, the digital media clustering system may provide the predicted distribution categories to the client device (e.g., from a remote server).

At a client device, a digital media clustering system may generate a plurality of digital media clusters from a digital media repository. In particular, in one or more embodiments, the digital media clustering system applies media clustering models (e.g., one or more machine learning models) to a digital media repository stored on a client device to generate the digital media clusters. The digital media clustering system can then identify and utilize the predicted distribution category (the predicted distribution category generated at the remote server) to select a digital media cluster (from among the digital media clusters). In one or more embodiments, a digital media clustering system provides digital media clusters for display and digital media items for distribution via a network system based on user interaction with the digital media items from the digital media clusters.

As just mentioned in the foregoing example, the digital media clustering system may analyze the network distribution history corresponding to the user's account at a remote server while analyzing sensitive client information only at the client device. In practice, by generating distribution categories at a remote server (based on public network distribution history) and then analyzing the digital media repository at the client device with the distribution categories, the digital media clustering system can expose digital media clusters without communicating information about sensitive digital media from the client device.

As described above, in one or more embodiments, the digital media clustering system utilizes a distribution classification model to generate predicted distribution categories. In particular, the digital media clustering system may utilize a distribution classification model that includes one or more machine learning models to analyze network distribution history and generate predicted distribution categories for future electronic communications. For example, the digital media clustering system may analyze previous social network posts from user accounts and utilize machine learning to predict a classification of future electronic communications (e.g., a network account category indicating digital media items most likely to be distributed in the future). In this manner, the digital media clustering system may determine a predicted distribution category that reflects topics, structures, styles, and trends (trend) from previous content distributions of the user's network account.

In addition, as described above, the digital media clustering system may also generate digital media clusters from the digital media repository at the client device using the media clustering model. In particular, the digital media clustering system may generate digital media clusters using media cluster models that themselves include various machine learning methods (e.g., facial recognition, object recognition, temporal and geographic clustering, composition quality models). In some embodiments, the digital media clustering system generates digital media clusters corresponding to different distribution categories. For example, the digital media clustering system may generate digital media clusters that depict digital images of particular subjects that are also reflected in the distribution classifications (e.g., generate digital image clusters that depict dogs corresponding to the dog distribution classifications).

The digital media clustering system may utilize the digital media clusters along with the predicted distribution categories to suggest digital media clusters that the client device user is most likely to distribute to other client devices. In fact, the digital media clustering system may align predicted distribution categories (which reflect historical trends from previous content distributions) with digital media clusters generated from digital media repositories stored on the client devices to suggest those digital media clusters that are most relevant to the user account. For example, in one or more embodiments, the digital media clustering system ranks digital media clusters based on predicted distribution categories and selects digital media clusters to be presented at the client device based on the rankings.

Once one or more suggested digital media clusters are identified, the digital media clustering system may provide the suggested digital media clusters for display via a client device of the user. For example, the digital media clustering system may select a representative image (e.g., a representative card) for each suggested digital media cluster and provide the representative image for display. Once the user selects the representative image, the digital media clustering system may provide each digital media item in the digital media cluster for display. A user may select one or more digital media items for distribution via a network system (e.g., posting to a social networking system).

In addition to suggesting digital media, the digital media clustering system may automatically edit digital media based on network distribution history. For example, the digital media clustering system may determine from the network distribution history previous digital media style modifications that the user has applied when distributing digital content via the network system. Digital media style modifications may be automatically applied when generating and providing digital media clusters.

The digital media clustering system may also suggest short-term digital content that the user account has previously distributed via the network system. In fact, in some embodiments, the digital media clustering system may re-present (resurface) the short-time digital content previously distributed by the user, which is no longer publicly available. For example, after a triggering event (e.g., after a short-time digital content is viewed or after a threshold period of time), the user's account may share (e.g., publish) the short-time digital content hidden from public view (public view) and the user client device. The digital media clustering system can analyze the historical short-term digital content (e.g., at a remote server) and utilize a distribution category model to predict the short-term digital content that the user is most likely to distribute in future electronic communications. The digital media clustering system may provide the predicted short-term digital content to the client device (e.g., for display with other digital media clusters identified at the client device).

As explained above, digital media clustering systems offer many advantages, benefits and practical applications over conventional systems. For example, digital media clustering systems have improved efficiency over conventional systems. In particular, by generating suggested digital media clusters based on network distribution history, the digital media clustering system may reduce the time and user interaction required to locate digital media items within a large digital media repository. For example, the digital media clustering system may generate a user interface that depicts digital media clusters that the user is most likely to utilize and distribute in future messages. Thus, a user may identify and distribute digital media items with only a small number of interactions in much less time.

In addition, the digital media clustering system improves accuracy over conventional systems. In fact, by generating digital media clusters that include digital media items that the network account is most likely to distribute, the digital clustering system may more accurately align the digital media with the relevant communications. In fact, the digital clustering system can identify digital media corresponding to user intent and electronic communications, which improves electronic messaging and avoids forgoing implementation of the system.

In addition, the digital media clustering system also improves the security of implementing computing devices without overloading the computing resources of individual client devices. As described above, the digital media clustering system may analyze previously shared information on a remote server to generate a predicted distribution category. The digital media clustering system may then analyze the digital media repository at the client device using these predicted distribution categories to select related digital media clusters. Thus, the digital media clustering system may securely identify digital media clusters at a client device without sending sensitive client information to a remote server. In addition, by performing the analysis of the network distribution history at the remote server, the digital media clustering system can reduce computational overhead at the client device and avoid overburdening local computing resources.

In addition, digital media clustering systems provide increased flexibility over otherwise rigid traditional systems. As described above, the digital media clustering system may suggest digital media clusters that are uniquely customized for individual users and/or accounts. In fact, the digital media clustering system may even suggest different digital media clusters for the same user between different network accounts and devices. For example, the digital media clustering system may expose a first set of digital media clusters for a first account (having a first focus) and a second set of digital media clusters for a second account (having a second focus) of the user. Thus, the digital media clustering system can customize suggested digital media clusters to historical distributions and digital media of individual users, accounts, and devices.

As the foregoing discussion shows, the present disclosure utilizes various terms to describe features and advantages of digital media clustering systems. Additional details regarding these and other terms used herein are now provided. For example, as used herein, the term "digital media item" (or "digital media") refers to a file that includes digital media. In particular, digital media items include digital images, digital audio, or digital video files. For example, the term "digital media" includes digital images having the following file extensions: JPG, TIFF, BMP, PNG, RAW or PDF. Similarly, "digital media" includes digital video with the following file extensions: FLV, GIF, MOV, QT, AVI, WMV, MP4, MPG, MPEG, or M4V.

As used herein, the term "digital media cluster" (or "cluster" or "cluster of digital media items") refers to a set of one or more digital media items. In particular, a digital media cluster may include a set of digital media items having a common characteristic. For example, a digital media cluster may include digital media items depicting common objects, people, actions, or locations; they share a common temporal characteristic (e.g., captured on the same day or time period); and/or share a common color scheme or mood (mood).

Further, as used herein, the term "repository of digital media" refers to a collection of multiple digital media items. In particular, the repository of digital media may include a memory, database, or log of digital media items. For example, the repository of digital images may include a camera roll on a client device. Similarly, the digital image repository stored on the client device may include a collection of thumbnails reflecting other digital media items managed (e.g., stored) on a third-party server.

As described above, the digital media clustering system may also suggest "short-time digital content. As used herein, "ephemeral digital content" (or "ephemeral digital content item") refers to digital content temporarily created and/or distributed with respect to one or more client devices. In particular, the ephemeral digital content may include digital media items that are distributed to client devices via a network system, but are hidden, removed, and/or deleted after a triggering event (e.g., from a distributing client device, a recipient client device, or a public server).

As discussed, the digital media clustering system may analyze the network distribution history of the user's network account when generating digital media clusters. As used herein, the term "network distribution history" refers to digital content previously distributed via a network system. In particular, the network distribution history may include digital media, electronic messages, social media posts, or other digital content distributed via the network system (e.g., by the user or the user's network account). For example, the network distribution history may include previous posts of the user to the social networking system, including digital media items, titles, tags (hashtags), or comments corresponding to the previous posts.

As used herein, the term "network account" refers to a collection of permissions, information, and data corresponding to a user of a multi-user computer system (e.g., a network system, as described below in connection with FIGS. 14, 15). In particular, the network account may include a set of permissions corresponding to users of the network system (e.g., rights to access content), electronic communications, posts, connections with other users (e.g., "friends" on the social networking system). As described in more detail below, a single user may have multiple accounts with different permissions, purposes, and data (e.g., a first account for distributing digital content about pets to a first group of users and a second account for distributing digital content about households to a second group of users).

In some embodiments, the digital media clustering system maintains a network profile for each user and/or network account. As used herein, a "network profile" includes a collection of data about a user and/or a network account (according to user privacy settings and preferences). For example, the network profile may include network distribution history, user preferences, interaction history (e.g., clicks, comments, views, reactions), location, and/or demographic information about the user.

As discussed, the digital media clustering system may utilize a distribution classification model to generate distribution categories for users/accounts based on network distribution history. As used herein, the term "distribution category" refers to a category or category corresponding to a user or network account that distributes digital content. In particular, the distribution category may include a classification (from multiple classifications) of a network account, a user, or a digital media item. For example, the predicted distribution category may include a predicted classification of the user/network account (or a predicted classification of future electronic messages to be distributed by the user/network account). For example, the digital media clustering system may determine that the network account corresponds to a "sports (sports)" category, where future electronic messages are likely to include digital media items depicting sports. Similarly, the digital media clustering system may determine that the network account corresponds to an "animal" category, where future electronic messages are likely to include digital media items depicting animals.

Additionally, as used herein, the term "distribution classification model" refers to a computer algorithm for generating distribution categories. In particular, the distribution classification model may include a machine learning model that predicts distribution categories based on network distribution history. As used herein, the term "machine learning model" refers to a computer algorithm that can learn from and predict data. In general, machine learning can operate by building models from example inputs (such as training network distribution history) to make data-driven predictions or decisions. The machine learning model may include one or more neural networks.

As used herein, the term "neural network" refers to a machine learning model that can be adjusted (e.g., trained) based on inputs to approximate an unknown function. In particular, the term neural network may include a model of interconnected neurons that communicates and learns to approximate complex functions and generate outputs based on a plurality of inputs provided to the model. For example, the term neural network includes an algorithm (or set of algorithms) that implements a deep learning technique that utilizes a set of algorithms to model high-level abstractions in data using supervised data (e.g., transcription training data) to adjust parameters of the neural network. For example, the neural network may include a convolutional neural network, a recurrent neural network (e.g., LSTM), or an antagonistic neural network (e.g., a generative antagonistic neural network).

Additional details regarding the digital media clustering system will now be provided in connection with the illustrative figures that depict example embodiments and implementations of the digital media clustering system. For example, FIG. 1 shows a schematic diagram of one embodiment of an example environment 100 in which a digital media clustering system 108 may operate in the environment 100. As shown, the environment 100 includes one or more servers 106 (examples of which are described in more detail below with reference to FIG. 13) connected to a plurality of client devices 102a-102n via a network 104.

Although fig. 1 shows a particular arrangement of client devices 102a-102n, network 104, and server 106, various additional arrangements are possible. For example, the client devices 102a-102n may communicate directly with the server 106, bypassing the network 104. Or alternatively, the client devices 102a-102n may communicate directly with each other.

Similarly, although environment 100 of fig. 1 is depicted as having various components, environment 100 may have additional or alternative components. For example, the environment 100 may be implemented on a single computing device having a digital media clustering system 108. In particular, the digital media clustering system 108 may be implemented in whole or in part on the client device 102 a.

As shown, environment 100 may include a server 106. The server 106 may generate, store, receive, and transmit various types of data, including distribution categories or selected digital media items. For example, server 106 may receive data from a client device (such as client device 102a) and send the data to another client device (such as client devices 102b, 102c, and/or 102 n). Server 106 may also transmit electronic messages between one or more users of environment 100. In an example embodiment, the server 106 is a data server. The server 106 may also include a communication server or a web hosting server. Additional details regarding server 106 will be discussed below with reference to FIG. 12.

As shown, in one or more embodiments, the server 106 may include or implement all or a portion of the network system 110 and/or the digital media clustering system 108. The network system 110 may include a social networking system or a digital communication networking system. Additional details regarding network system 110 are provided below (e.g., in connection with fig. 14 and 15).

The network system 110 and/or the digital media clustering system 108 may include an application running on the server 106, or a portion may be downloaded from the server 106. For example, the digital media clustering system 108 may include a web-hosted application that allows client devices 102a-102n to interact with content hosted at the server 106. For example, in one or more embodiments of environment 100, one or more client devices 102a-102n may access a web page supported by server 106. In particular, the client device 102a may run a web application (e.g., a web browser) to allow a user to access, view, and/or interact with web pages or websites hosted at the server 106.

As described above, in one or more embodiments, client devices 102a-102n store a repository of digital media. The digital media clustering system 108 may utilize the server 106 and the client devices 102a-102n to suggest digital media clusters and distribute selected digital media items to other users via the network system 110. For example, the digital media clustering system 108 may utilize the server 106 to generate distribution categories by analyzing the network distribution history of network accounts. The digital media clustering system 108 may transmit the distribution categories from the server 106 to the client device 102a and generate suggested digital media clusters with the client device 102 a. In particular, the client device 102a can analyze a repository of digital media items, generate a set of digital media clusters, and select a digital media cluster to suggest based on a distribution category from the server 106. Based on the digital media item selected by the user from the suggested digital media clusters, the client device 102a may transmit the digital media item to the server 106 for distribution to the client devices 102b, 102n via the network system 110.

As just described, the digital media clustering system 108 may be implemented in whole or in part by the various elements 102a-106 of the environment 100. It should be appreciated that although certain components or functionality of the digital media clustering system 108 are described in the foregoing examples with respect to particular elements of the environment 100, various alternative implementations are possible. For example, in one or more embodiments, the digital media clustering system 108 is implemented on the client device 102 a. Similarly, in one or more embodiments, the digital media clustering system 108 may be implemented on the server 106. Moreover, the different components and functionality of the digital media clustering system 108 may be implemented separately between the client devices 102a-102n, the server 106, and the network 104.

As described above, the digital media clustering system 108 can utilize a distribution classification model and/or a media clustering model to generate suggested digital media clusters. In particular, FIG. 2 illustrates generating suggested digital media clusters in accordance with one or more embodiments. In particular, FIG. 2 illustrates a digital media clustering system that uses a distribution classification model 208 to generate suggested digital media clusters 216 based on network distribution history 204 corresponding to a user profile 202.

As shown in FIG. 2, the digital media clustering system accesses a user profile 202 that includes a network distribution history 204. In particular, user profile 202 includes details regarding a network account corresponding to a user (e.g., a user of network system 110). For example, the network distribution history 204 indicates previous digital content, such as social network posts, distributed from a network account corresponding to the user via the network system 110.

As shown in fig. 2, the network distribution history 204 also includes short-time digital content 206 previously distributed via the network system 110. For example, the ephemeral digital content 206 may include social networking posts or other electronic communications distributed by the user/user account that are automatically removed after a triggering event. Additional details regarding the digital content over short-time are provided below (e.g., in connection with fig. 4).

As shown in FIG. 2, the digital media clustering system 108 analyzes the network distribution history through a distribution classification model 208. The distribution classification model 208 may include one or more machine learning models (and/or heuristic models). For example, the distribution classification model 208 may include a facial recognition model (to identify individuals depicted in the digital media from the network distribution history 204) and an object recognition model (to identify objects depicted in the digital media from the network distribution history 204). Further, the distribution classification model 208 may include a classification model that identifies distribution categories based on network distribution history. Additional details regarding the distribution classification model 208 are provided below (e.g., in conjunction with fig. 3-6).

As described above, the digital media clustering system 108 may analyze the digital media repository to suggest digital media clusters. In particular, as shown in FIG. 2, the digital media clustering system 108 analyzes a digital media repository 210 at the client device 102 a. The digital media repository 210 includes digital media items 212 (e.g., digital images and digital videos captured by the client device 102 a). The digital media clustering system 108 analyzes the digital media repository 210 using a media clustering model 214.

The media clustering model 214 may include various machine learning models (and/or heuristic models). For example, the media cluster model 214 may include a facial recognition model (to recognize individuals depicted in the digital media item 212) and an object recognition model (to recognize objects depicted in the digital media item 212). Further, the media clustering model 214 can include a machine learning model for grouping digital media items. For example, the media clustering model 214 can generate digital media clusters based on common characteristics of the digital media items 212. Additional details regarding the media clustering model 214 are provided below (e.g., in conjunction with fig. 3-6).

As shown in FIG. 2, the digital media clustering system 108 may generate suggested digital media clusters 216 based on the distribution classification model 208 and the media cluster model 214. For example, the digital media clustering system 108 may analyze digital media clusters generated by the media clustering model 214 according to distribution categories generated by the distribution classification model 208 to generate suggested digital media clusters 216.

In conjunction with FIG. 2, the digital media clustering system 108 analyzes the digital media repository 210 and presents suggested digital media clusters 216 on the client device 102 a. Thus, the digital media clustering system 108 may analyze the digital media repository 210, generate suggested digital media clusters 216, and display the suggested digital media clusters 216 without transmitting information about the digital media repository 210 from the client device 102 a. In some embodiments, with the appropriate permissions of the user, the digital media clustering system 108 may also analyze the digital media repository 210 at a remote server using the network distribution history 204 to generate suggested digital media items.

Turning now to fig. 3, additional details regarding generating suggested digital media clusters and distributing digital media items are provided in accordance with one or more embodiments. In particular, fig. 3 shows that server 106 stores network distribution history 302, and client device 102(a) stores digital media repository 308. The digital media clustering system 108 utilizes the server 106 to generate and transmit the predicted distribution category 304 to the client device 102 a. The digital media clustering system 108 utilizes the client device 102a to analyze the digital media repository 308 and the predicted distribution categories 208 to generate suggested digital media clusters 316. In addition, the digital media clustering system 108 utilizes the client device 102a and the server 106 to distribute the user-selected digital media items to other client devices (e.g., via the network system 110).

As shown in fig. 3, the digital media clustering system 108 utilizes the distribution classification model 208 to generate a predicted distribution category 304 (or multiple predicted distribution categories). The distribution classification model 208 may include various separate machine learning models (and/or heuristic models). For example, in one or more embodiments, the distribution classification model 208 includes a convolutional neural network. In particular, the distribution classification model 208 may include a convolutional neural network that includes multiple neural network layers that analyze features of the network distribution history 302 at different levels of abstraction. For example, the distribution classification model 208 may include a plurality of interconnected convolutional layers, pooling layers (pooling layers), ReLu layers, and fully connected layers.

The digital media clustering system 108 may convert the network distribution history 302 into a form compatible with the inputs of the distribution classification model 208. For example, the digital media clustering system 108 may generate an input feature vector that reflects the network distribution history 302. For example, the digital media clustering system 108 may convert digital media items, text, or other content of previous electronic communications (e.g., social networking posts) into feature vectors. For example, the digital media clustering system 108 may generate feature vectors that include R, G, B color channels for individual pixels in the digital media. Similarly, the digital media clustering system 108 may generate vector representations of text, such as one-hot encoding (one-hot encoding) of individual text characters. In some embodiments, the digital media clustering system 108 generates a semantic vector representation that reflects the semantics of the text. For example, the digital media clustering system 108 may apply a word to vector algorithm (word to vector algorithm) that maps text input to a semantic feature space, where locations within the semantic feature space reflect the relative meaning of the text. The digital media clustering system 108 may also generate feature vectors that reflect historical distribution timing (e.g., frequency or time of historical distribution).

The distribution classification model 208 may analyze the inputs generated from the network distribution history 302 to generate a predicted distribution category 304. In particular, where the distribution classification model 208 includes a convolutional neural network, the various layers may apply learned weights and parameters (which are learned through training iterations) to the input features to accurately determine a predicted distribution category that the user may select in future electronic communications. Additional information regarding training the distribution classification model 208 is provided below (e.g., in connection with FIG. 6).

Although some of the foregoing examples describe implementations of the distribution classification model 208 as a convolutional neural network, in some embodiments, the digital media clustering system 108 utilizes additional or alternative machine learning models. For example, in some embodiments, the digital media clustering system 108 implements the distribution classification model 208 as a recurrent neural network, such as a long-short term memory neural network ("LSTM"). In particular, the distribution classification model 208 may utilize a recurrent neural network that includes a layer divided into sequential elements that analyze items in sequential order. These sequential elements may together analyze the sequential input data and the feature vectors from previous sequential elements to generate a prediction. In this manner, the distribution classification model 208 may take into account sequential trends in the data.

For example, the distribution classification model 208 may analyze digital content distributed from network accounts over time as sequential inputs to a recurrent neural network. Each sequential element of the recurrent neural network can analyze individual distributions, predict distribution categories, and pass feature vectors about the timing and content of previous distributions to other sequential layers. In this manner, based on the sequential timing and structure of previous distributions, the distribution classification model 208 can learn to predict distribution categories for network accounts over time. For example, the distribution classification model 208 may learn that network accounts post related to the "outdoor" category on saturday, but post related to the "friends" category on friday. Similarly, the distribution classification model 208 may learn that the web account followed the next day with a "self-portrait" category post (follow) "friend" category post.

Digital media in addition to recurrent neural networksThe distribution classification model 208 may also be implemented by the body clustering system 108 using other machine learning models. For example, the digital media clustering system 108 may utilize an antagonistic neural network, a decision tree, a random forest algorithm, a naive Bayes algorithm (c/e: (a/d)), andbayes algorithms) and/or support vector machines to generate predicted distribution categories 304.

The distribution classification model 208 may also include other machine learning models to help generate predicted distribution categories. For example, the distribution classification model 208 may include a facial recognition model and/or an object recognition model. For example, the distribution classification model 208 may include a you only need to look at (you-only-look-once, "YOLO") object detection model or another object classifier that identifies objects depicted in digital media. Similarly, the distribution classification model 208 may include a face recognition algorithm, such as deep face, that employs a nine-layer neural network to recognize faces depicted in digital media.

In some embodiments, the distribution classification model 208 utilizes objects and individuals identified in digital media from the network distribution history 302 to determine a predicted distribution category 304. For example, digital media clustering system 108 may provide the identified objects and individuals as inputs to a neural network (e.g., the convolutional neural network or the recurrent neural network described above). The neural network then generates a predicted distribution category based on the identified objects and individuals 304.

In some embodiments, the distribution classification model 208 may utilize heuristics (or predefined rules) to determine the distribution category. For example, the distribution classification model 208 may determine the predicted distribution category 304 by applying heuristics to identified objects and/or individuals depicted in the digital media from the network distribution history 302. For example, once a threshold number of digital media depicting dogs are identified, the distribution classification model 208 may determine a predicted distribution category "dogs (dog)". Similarly, once a threshold number of digital media depicting mountains or water are identified, the distribution classification model 208 may determine a predicted distribution category "nature".

Similarly, the distribution classification model 208 may analyze text (e.g., tags, titles, or comments) to determine a predicted distribution category 304. For example, the distribution classification model 208 may identify from the tags that the user repeatedly distributes content for a trending challenge (e.g., # ice bucket challenge (# icebucktchallonge)) or a trip (e.g., # california)). The distribution classification model 208 may determine a predicted distribution category 304 (e.g., generate a "challenge" category or a "travel" category) from the text.

As shown in the foregoing example, the digital media clustering system 108 may use a variety of different categories for the predicted distribution category 304. In some embodiments, the digital media clustering system 108 generates the predicted distribution category 304 from a set of predefined categories. For example, the predefined categories may include an animal category (e.g., dog, cat, pet, or horse), an outdoor/natural category (e.g., mountain, lake, hiking, mountain climbing), a sports category (e.g., basketball, football, or golf), a food category, a challenge category, a travel/travel category, a friends category, a holiday category, or a self-portrait category (e.g., a front-facing camera or a self-portrait (selfie) "category). In some embodiments, the predefined categories include categories defined based on timing. For example, the digital media clustering system 108 may utilize a "photos on this day" category (e.g., for accounts that repeatedly post historical photos from the same day in the last few years), and/or a "then-and-now category" (e.g., for accounts that post digital media depicting locations or people within different time periods), and/or any other distribution category described herein. In some embodiments, the set of predefined distribution categories may include one or more mood or color categories. For example, digital media clustering system 108 may determine that a network account distributes digital content (e.g., a color profile) with a particular mood or color.

As shown in fig. 3, the distribution classification model 208 may analyze the network distribution history 302 and generate a predicted distribution category 304 (e.g., from a plurality of predicted distribution categories). In some embodiments, the distribution classification model 208 generates a relative ranking (e.g., confidence score) for a set of distribution categories. For example, where the distribution classification model 208 includes a convolutional neural network, the distribution classification model 208 may output a probability (e.g., a confidence score) for each possible distribution category. The distribution classification model 208 may also rank the possible distribution categories and select a predicted distribution category 304 based on the ranking (e.g., select the highest ranked distribution category or the distribution category that satisfies a confidence threshold). Although fig. 3 shows a single predicted category 304, the digital media clustering system 108 may generate multiple predicted distribution categories. For example, in some embodiments, the digital media clustering system 108 provides a ranking (e.g., according to corresponding confidence scores) of a plurality of distribution categories.

As shown in fig. 3, the digital media clustering system 108 may provide a predicted distribution category 304 (or a set of ranked predicted distribution categories) to the client device 102 a. The client device 102a can analyze the predicted distribution categories 304 and the digital media repository 308 to generate a suggested digital media cluster 316 (or multiple suggested digital media clusters).

In particular, as shown, the digital media clustering system 108 utilizes the media clustering model 214 to analyze the digital media repository 308 at the client device 102 a. Similar to the distribution classification model 208, the media cluster model 214 may include various machine learning models. For example, the media cluster model 214 may include a k-means clustering algorithm (or other clustering/grouping algorithm) that takes into account various input features and generates digital media clusters. For example, the media clustering model 214 may include a k-means clustering algorithm that generates clusters based on time, individuals depicted in the digital media, and/or objects depicted in the digital media.

The media cluster model 214 may also include a face recognition algorithm and/or an object recognition algorithm (as described above). For example, the media clustering model 214 may utilize facial recognition algorithms and/or object recognition algorithms to generate features of media items in the digital media repository 308. The media clustering model 214 may then utilize a clustering algorithm to generate clusters based on the features. Thus, for example, the media clustering model 214 can generate digital media clusters that group common individuals depicted in digital media (e.g., the same set of friends depicted in a digital image). The media clustering model 214 can generate digital media clusters that group common objects depicted in digital media (e.g., groups dogs depicted in digital images). The media clustering model 214 may also generate media clusters based on the time at which the digital image was captured.

In other embodiments, the media cluster model 214 may include other machine learning models, such as the neural network described above. For example, in some embodiments, the media cluster model 214 may include a convolutional neural network that classifies the digital media categories (relative to a set of predefined categories). The media clustering model 214 then generates clusters by grouping the digital media based on categories.

Similar to the actions described above in connection with the distribution classification model 208, the digital media clustering system 108 may generate input features from the digital media repository 308. By way of example, the digital media clustering system 108 may generate a color channel (e.g., R, G, B map (map)) that reflects the visual appearance of digital media items in the digital media repository 308. In addition, the digital media clustering system 108 may encode the time of capture, the location of capture, or other metadata from the digital media (e.g., by generating a feature vector that includes time, date, and geographic coordinates). The digital media clustering system 108 can then provide this data as input to the media clustering model 214.

As described above, the digital media clustering system 108 may also generate other features as input. For example, the digital media clustering system 108 may apply facial recognition algorithms and/or object recognition algorithms to determine individuals, objects, or locations depicted in the digital media.

The media clustering model 214 can analyze the characterizing representations reflecting the digital media repository 308 and determine a category corresponding to each digital media item. For example, by applying the learned weights and parameters to the internal parameters of the convolutional neural network, the media clustering model 214 may determine that a first digital media item corresponds to an animal category, a second digital media item corresponds to a sports category, and a third digital media item corresponds to a travel category.

In some embodiments, the digital media clustering system 108 may determine that a single digital media item corresponds to multiple categories. For example, the digital media clustering system 108 may determine that the digital images depicting the hiking grouping correspond to a "friends" category and an "outdoor" category. In fact, the media clustering model 214 may generate confidence scores for a plurality of predefined categories and assign media items based on the confidence scores (e.g., assign media items to any category that satisfies a confidence threshold).

In addition to k-means clustering and convolutional neural networks, the digital media clustering system 108 may utilize other machine learning models. For example, digital media clustering system 108 may implement media clustering model 214 as a generative neural network (generative neural network), a recurrent neural network, a random forest algorithm, a support vector machine, and/or another machine learning method described herein.

Further, media clustering model 214 can include one or more heuristic models to identify clusters. For example, the media cluster model 214 may identify a "travel (trip)" cluster by applying heuristics corresponding to the number, time, and location of digital media items. For example, the media cluster model 214 may identify an increased number or percentage of digital images over a period of time at a particular location. In response, the media clustering model 214 may identify the digital images as "travel" clusters.

Similarly, the media clustering model 214 may apply heuristics for objects or individuals depicted in the digital images to identify clusters. For example, the media cluster model 214 may identify a set of digital images depicting an animal and include the set of digital images in an "animal" cluster. Similarly, the media clustering model 214 may identify a set of digital images having a large number of individuals and include these digital images in a "friends" cluster.

In some embodiments, the media cluster model 214 may infer relationships between individuals depicted in the digital media items. For example, the media clustering model 214 may identify friends and/or family relationships. For example, in some embodiments, the client device 102a includes thumbnail images of "friends" (e.g., connected users with sharing permissions to access and/or share digital content) on the network system 110. The media clustering model 214 may utilize facial recognition algorithms to match visual features from the thumbnail images to visual features of the digital media repository 308. The media clustering model 214 may then identify images that include these visual features (e.g., the faces of friends) in the "friends" cluster. In this manner, the digital media clustering system 108 may identify a "friend" cluster without any sensitive identifying information (e.g., name or ID) of any of the user's friends.

As shown in FIG. 3, the digital media clustering system 108 utilizes the media clustering model 214 to generate clusters 312 of media items based on the digital media repository 308. The digital media clustering system 108 then generates a suggested digital media cluster 316 based on the predicted distribution categories 304 and the media item clusters 312 using the digital media cluster selection model 314. In particular, the digital media cluster selection model 314 aligns the media item clusters (as generated at the client device 102a) reflecting the digital media repository 308 with the predicted distribution categories 304 (as generated at the server 106) reflecting the network distribution history 410 to generate suggested digital media clusters. By utilizing the predicted distribution categories 304 and media item clusters, the digital media cluster selection model 314 can generate suggested digital media clusters 316 to accurately reflect digital media on client devices that the network account user is likely to distribute in future electronic communications.

As shown, the digital media cluster selection model 314 can compare the media item clusters 312 to the predicted distribution categories 304 to generate suggested digital media clusters 316. For example, as described above, the predicted distribution category 304 may include a "pets" category. Additionally, the media item clusters 312 may include a "pet" category, a "sports" cluster, and an "outdoor" category. The digital media cluster selection model 314 may determine a correspondence between the "pet" category and the "pet" cluster and select the pet cluster as the suggested digital media cluster 316.

As illustrated by the foregoing example, in some embodiments, the categories/clusters used by the media clustering model 214 directly match the distribution categories used by the distribution classification model 208. For example, the distribution classification model may utilize a set of predefined distribution categories including an animal category, a sports category, and an outdoor category. Similarly, the media cluster model 214 can utilize digital media categories including animal categories/clusters, sports categories/clusters, and outdoor categories/clusters.

In some embodiments, the categories/clusters used by the media clustering model 214 do not directly match the distribution categories used by the distribution classification model 208. Instead, the digital media clustering system 108 can utilize the digital media cluster selection model 314 to intelligently align distribution classifications with predicted distribution classifications 304. For example, the predicted distribution category may indicate that the network account in question is an athletic account. The media item clusters 312 generated by the media clustering model 214 may include food clusters, outdoor clusters, and exercise clusters. The digital media cluster selection model 314 can utilize one or more heuristics that associate distribution categories with digital media clusters. For example, the digital media cluster selection model 314 may include heuristics indicating that sports accounts are relevant for outdoors and exercise. Thus, the digital media cluster selection model 314 can select outdoor clusters and exercise clusters as suggested digital media clusters.

The digital media clustering system 108 may also take into account digital media quality and composition when generating suggested digital media items. For example, in some embodiments, the digital media clustering system selects (e.g., filters) digital media items based on quality (e.g., clarity or fuzziness) and composition (e.g., arrangement of visual elements). For example, where the digital media repository includes a large number of similar digital images, the digital media clustering system 108 may filter the similar digital images based on quality and composition (e.g., a threshold number or a threshold quality/composition level). The digital media clustering system 108 may then suggest only selected/unfiltered digital media items.

Further, as described above, in some embodiments, the predicted distribution categories 304 are provided to the client device 102a as a ranking of multiple predicted distribution categories (e.g., based on confidence scores for the respective distribution categories). The digital media cluster selection model 314 can utilize the ranking (e.g., confidence score) to generate suggested digital media clusters 316. For example, consider a case where the predicted distribution categories include the following rankings (and confidence scores): pet category (0.8), current and present category (0.75), and travel category (0.65). Further, the media item clusters include a food cluster, a travel cluster, an exercise cluster, and a current and present cluster. Based on the ranking and confidence scores of the distribution categories, the digital media cluster selection model 314 may utilize pet clusters (i.e., clusters corresponding to the highest ranked predicted distribution categories) as suggested digital media clusters 316.

Although fig. 3 (and the preceding examples) indicate a single suggested digital media cluster 316, the digital media cluster selection model 314 may also generate multiple suggested digital media clusters. In fact, the digital media clustering system 108 may generate a predefined number of suggested digital media clusters or a variable number of suggested digital media clusters (e.g., clusters that meet a confidence threshold). For example, in the foregoing example of different rankings based on confidence scores, the digital media cluster selection model 314 may select clusters that satisfy a confidence threshold of 0.7. In such an example, the digital media cluster selection model 314 would select the "pet" cluster (confidence score of 0.8) and the "then and now" cluster (confidence score of 0.75) as the suggested digital media clusters.

Once the suggested digital media clusters 316 are identified, the digital media clustering system 108 may provide the suggested digital media clusters 316 for display. For example, the digital media clustering system 108 may provide digital media clusters 316 for display in a user interface to select and distribute digital media items for distribution via the network system 110. Additional details regarding example user interfaces used by the digital media clustering system 108 are provided below (e.g., in conjunction with fig. 7-9C).

As shown in FIG. 3, the digital media clustering system 108 may identify a user-selected digital media item 318. In particular, based on user interaction with the digital media selection user interface, the digital media clustering system 108 may identify one or more user-selected digital media items from the one or more suggested digital media clusters.

Further, as shown, the digital media clustering system 108 may transmit a user-selected digital media item 318. In particular, the client device 102a can send the user-selected digital media item 318 to the server 106, and the server 106 (via the network system 110) can distribute the user-selected digital media item 318 to other client devices (e.g., client devices 102b, 102 n). For example, the server may distribute the digital media items as part of an electronic communication to other users of the network system 110.

As described above, the digital media clustering system 108 may also analyze and suggest short-term digital content items. Fig. 4 illustrates suggesting an in-time digital content item using the server 106 and the client device 102 a. For example, fig. 4 shows that the digital media clustering system 108 utilizes the server 106 to analyze the short-time digital content 402 generated on the client device 102a, generate suggested short-time digital content clusters 416, and provide the suggested short-time digital content clusters 416 as well as other suggested digital media clusters (e.g., suggested digital media clusters 316).

As shown in fig. 4, the digital media clustering system 108 generates short-time digital content 402. For example, the digital media clustering system 108 (via the client device 102a) may capture and include digital images or digital videos as part of a short-time electronic message (e.g., an electronic message designated to be automatically deleted, removed, or hidden upon a triggering event).

As shown, the digital media clustering system 108 may transmit the short-time digital content 402 from the client device 102a to the server 106. Further, the server 106 (via the network system 110) may distribute the short-time digital content 402 to other client devices. For example, the digital media clustering system 108 may publish the short-lived digital content 402 as a temporary dynamic (store) on a social networking system or transmit the short-lived digital content as an instant message to other client devices.

The digital media clustering system 108 may also remove short-term digital content. For example, as shown, the digital media clustering system 108 may delete, remove, or hide short-time digital content (e.g., short-time digital messages) from the client device 102 a. In addition, the digital media clustering system 108 can delete, remove, or hide the short-time digital content from other public sources (e.g., from other client devices or public servers). For example, the digital media clustering system 108 may automatically remove the short-lived digital content in response to one or more triggering events. For example, in response to detecting that the recipient client device has displayed the short-time digital content, the digital media clustering system 108 may remove the short-time digital content from the recipient client device. Similarly, based on the threshold time period, the digital media clustering system 108 may remove or hide the short-time digital content. For example, the digital media clustering system 108 may post the short-lived digital content (e.g., as short-lived dynamic) as part of a social networking system within a threshold period of time and then remove the short-lived digital content when the period of time elapses.

Although the digital media clustering system 108 may delete or remove digital content from one or more client devices (or other public facing resources) according to consumer privacy settings and retention preferences, the digital media clustering system 108 may also retain a copy of the short-term digital content (e.g., on the server 106). For example, the digital media clustering system 108 may retain previously distributed short-term digital content as part of the network distribution history 410 for the user/network account, depending on the user privacy settings and retention preferences. Further, as shown in fig. 4, the digital media clustering system 108 may analyze the short-time digital content 402 when it is suggested to the user's client device.

For example, as shown in FIG. 4, the digital media clustering system 108 utilizes the server 106 to analyze a network distribution history 410 of users/network accounts that include the short-time digital content 402. In particular, the digital media clustering system 108 utilizes the distribution classification model 414 to analyze the network distribution history 410 and the short-term digital content 402 (along with other short-term digital content corresponding to the user/network account) to generate a suggested short-term digital content cluster 416 that includes one or more short-term digital content items. The distribution classification model 414 may utilize the actions and algorithms described above (e.g., in conjunction with the distribution classification model 208 and the media clustering model 214) to generate the short-time digital content clusters 416.

In contrast to the digital media repository shown in fig. 3 (which is stored on the client device 102a), the short-time digital content 402 has been transmitted to the server 106 according to user instructions and privacy settings. Thus, the digital media clustering system 108 can analyze the short-time digital content 402 using the other network distribution history 410 to select a suggested short-time digital content cluster 416 (i.e., rather than separate the analysis between the client device and the server). In addition, the digital media clustering system 108 may also analyze other information stored on the server 106, such as network accounts and network distribution histories corresponding to affiliations (e.g., friends) of users on the network system 110.

In some embodiments, the server 106 may provide the short-time digital content 402 to the client device 102a for analysis (e.g., as part of the digital media repository 308 shown in fig. 3). Thus, although fig. 4 illustrates analysis of the short-time digital content 402 on the server 106, in some embodiments, the digital media clustering system 108 analyzes the short-time digital content 402 on the client device 102a and generates a suggested short-time digital content cluster 416.

In connection with the embodiment shown in fig. 4, once the suggested short-time digital content cluster 416 is generated, the digital media clustering system 108 provides the suggested short-time digital content cluster to the client device 102 a. In turn, the digital media clustering system 108 utilizes the client device 102a to provide the short-time digital content clusters 416 for display. As shown in fig. 4, the digital media clustering system 108 can provide short-term digital content clusters 416 along with other suggested digital media. For example, the digital media clustering system 108 may provide the short-time digital content clusters 416 along with the suggested digital media clusters 316.

The digital media clustering system 108 may also distribute digital media based on short-term digital content clusters and/or suggested digital media clusters 316. For example, as shown in FIG. 4, the digital media clustering system 108 identifies a user-selected digital media item 420. The user-selected digital media items 420 may include digital short-time digital content 402 from the short-time digital content cluster 416 and/or digital media items from the suggested digital media cluster 316. The digital media clustering system 108 provides the user-selected digital media items 420 to the server 106 (via the client device 102a) and distributes the user-selected digital media items 420 to other client devices according to user instructions.

As described above, the digital media clustering system 108 can generate distribution categories and suggested short-time digital content at the server 106 based on network distribution history. In addition, the digital media clustering system 108 may also determine suggested digital media style modifications based on network distribution history. Fig. 5 illustrates an example embodiment of generating (at the server 106) a distribution category, a suggested short-time digital content, and a style modification to generate (at the client device 102a) a suggested digital media cluster.

In particular, in conjunction with FIG. 5, the digital media clustering system 108 utilizes the distribution classification model 208 to analyze a network distribution history 502 that includes short-time digital content 504. As described in connection with fig. 3 and 4, the distribution classification model 208 generates distribution categories 506a-506b (i.e., "dogs" and "travel") and ephemeral digital content items 506c (i.e., "ephemeral content l").

In addition, the digital media clustering system 108 generates suggested digital media style modifications 506d based on the network distribution history 502. In particular, the suggested digital media style modification 506d (i.e., "Boomerang") includes modifying the digital media item to a brief video representation that repeats a forward and backward loop. The digital media clustering system 108 determines suggested style modifications based on analyzing the network distribution history and identifying previous digital media items (which include suggested digital media style modifications 506d) distributed via the network system 110.

The digital media clustering system 108 may monitor and track various suggested style modifications. For example, the digital media clustering system 108 may monitor the application of various filters; correcting the quality; or a change in color, style, quality (e.g., to fix unclear eyes or blurry regions), hue (hue), or texture in a digital media item distributed via the network system 110. The digital media clustering system 108 may include these modifications in the network distribution history 502. Using a similar approach as described above (in conjunction with fig. 3, 4), the distribution classification model 208 may analyze the network distribution history and generate suggested style modifications. For example, the distribution classification model 208 may utilize an LSTM neural network to predict one or more style modifications that a user may apply to a digital media item in future electronic communications.

As shown in FIG. 5, the digital media clustering system 108 also utilizes the media clustering model 214 to generate a plurality of digital media clusters 510 (at the client device 102 a). Digital media clusters 510 include a "this day" cluster (including digital media items from this day in a previous time period), "location" cluster (including digital media items from a particular location), "dog" cluster (including digital media items depicting a dog), "holiday" cluster (including digital media items on a holiday), "friend" cluster (including digital media items of a user friend), "self portrait" cluster (including self portrait), "mood"/"color" cluster (including digital images with a common color scheme), and "travel" cluster (including digital images of recent travel).

The digital media clustering system 108 analyzes the digital media clusters 510 and the distribution categories 506a-506b, the short-time digital content items 506c, and/or the suggested digital media style modifications 506 d. As shown, the digital media cluster selection model 314 generates suggested digital media clusters 512a-512 c. In particular, the digital media cluster selection model 314 compares the distribution categories 506a-506b to the digital media clusters 510 and identifies a correspondence between the distribution category 506a (the "dog" category) and the digital media cluster 510a (the "dog" cluster). In addition, the digital media cluster selection model 314 also identifies a correspondence between the distribution category 506b (the "travel" category) and the digital media cluster 510b (the "travel" cluster). In response, the digital media clustering system 108 utilizes the digital media clusters 510a, 510b to generate suggested digital media clusters 512a, 512 b.

However, as shown, the digital media clustering system 108 can modify the digital media clusters identified by the media clustering model 214 when generating the suggested digital media clusters 512. For example, based on the suggested digital media pattern modification 506d, the digital media clustering system 108 modifies the digital media cluster 510a to generate the suggested digital media pattern 506 d. In particular, the digital media clustering system 108 modifies one or more digital media items in the digital media cluster 510a and transforms the one or more digital media items with the suggested digital modifications (e.g., such that the digital media items are now "boomerangs" that are repeatedly played back forward and backward).

Although FIG. 5 illustrates generating a particular digital modification, the digital media clustering system 108 can apply any suggested digital modification to the digital media items. For example, the digital media clustering system 108 may apply various filters or corrections identified based on the network distribution history 502.

3-5 illustrate generating suggested digital media clusters for a particular network account of a user and a particular client device, digital media clustering system 108 may generate suggested digital media clusters for a plurality of client devices corresponding to a user and a plurality of network accounts of a user. For example, the digital media clustering system 108 may analyze a second network distribution history for a second network account of the user to generate a second set of distribution categories. The digital media clustering system 108 may compare the second set of distribution categories to digital media clusters at the client device and generate a second set of suggested digital media clusters (specific to a second network account).

Similarly, the user may utilize a second client device (with an additional repository of digital media items), and the digital media clustering system 108 may generate an additional set of suggested digital media clusters. For example, the digital media clustering system 108 may analyze additional repositories of digital media items to generate an additional set of digital media clusters. The digital media clustering system 108 may analyze a set of digital media clusters having a distribution category (e.g., the predicted distribution category 304) to generate an additional set of suggested digital media clusters. Similarly, as the digital media repository changes (or is based on a second digital media repository), the digital media clustering system 108 may generate further suggested media clusters. Thus, the digital media clustering system 108 may generate unique suggested digital media clusters for different users, different network accounts, different client devices, and/or different digital media repositories.

In some cases, the user may not wish to see one or more suggested digital media items. For example, if a suggested digital media item depicts a lost pet, the user may not wish to see the digital media item in future suggestions. In some embodiments, the digital media clustering system 108 provides user feedback control to remove a digital media item or several types of digital media items. Upon receiving user interaction via the user feedback control, the digital media clustering system 108 may remove digital media items and similar digital media items from future suggestions.

To reduce the time and computational power to generate suggested digital media items, the digital media clustering system 108 may limit its analysis to particular digital media items. For example, after generating the digital media cluster, the digital media clustering system 108 may limit its analysis to new digital content (e.g., add the new digital content to the cluster previously generated by the digital media clustering system 108, rather than re-analyze the entire digital content repository). Similarly, the digital media clustering system 108 may generate suggested digital media items in advance (e.g., behind scenes) before the user accesses the digital content for transmission to other client devices. For example, upon launching a network system application at a client device, the digital media clustering system 108 may analyze a repository of digital media items and generate suggested media items. The digital media clustering system 108 may then present the suggested media items at a later time (e.g., when the user accesses a camera roll).

As described above, the digital media clustering system 108 may train the distribution classification model to generate predicted distribution categories. FIG. 6 illustrates training the distribution classification model 208 in accordance with one or more embodiments. In particular, FIG. 6 illustrates training the distribution classification model 208 based on a training network distribution history 602 and a ground truth (ground route) distribution category 606.

The digital media clustering system 108 may identify or generate a training network distribution history 602. For example, the digital media clustering system 108 may monitor user interactions with the network system 110 to identify and/or generate the training network distribution history 602. By way of example, the digital media clustering system 108 may identify previous posts or electronic communications provided via a network account of the network system 110.

As shown in FIG. 2, the digital media clustering system 108 analyzes the training network distribution history 602 using the distribution classification model 208 to generate a predicted distribution category 604. For example, the digital media clustering system 108 may analyze the training network distribution history 602 for network accounts and predict a distribution category that indicates a "dog account" indicating that future posts will likely include digital media items depicting dogs.

The ground truth distribution category 606 may be accessed by the digital media clustering system 108. The ground truth distribution category 606 includes observation data indicating the actual distribution category corresponding to the user account. For example, the ground truth distribution category 606 may include observed social media posts or electronic communications that include digital media items that the training account actually distributed after training the network distribution history 602. For example, the ground truth distribution category 606 may include social media posts of outdoor pictures. Similarly, the ground truth distribution category 606 can include a label (e.g., a user label) that identifies the network account as corresponding to a particular category (e.g., a user selection that a particular account is a "travel log" account). Further, the ground truth distribution category 606 may include negative (negative) user input or positive (positive) user input as a ground truth indicator (e.g., user interactions via user feedback control indicating that suggested digital media items are not desired may be used as negative ground truth data points).

As shown, the digital media clustering system 108 may train the distribution classification model 208 by comparing the predicted distribution categories 604 with the ground truth distribution categories 606. In particular, the digital media clustering system 108 may utilize a loss function 608 to compare the predicted distribution category 604 and the ground truth distribution category 606 and determine a measure of loss (e.g., a difference between predicted and actual). Based on the measure of loss, digital media clustering system 108 may modify (as indicated by arrow 610) internal parameters of distribution classification model 208. For example, the digital media clustering system 108 may backpropagate the measure of loss through the layers of the neural network to modify parameters and weights and reduce the measure of loss between the predicted distribution category 604 and the ground truth distribution category 606.

For example, the digital media clustering system 108 may determine a difference (e.g., a measure of loss) for the predicted distribution category 604 "animals" and the ground truth distribution category "outdoor". The digital media clustering system 108 may then modify internal parameters of the distribution classification model 208 to reduce the measure of loss and more accurately predict distribution categories. The digital media clustering system 108 may repeatedly perform the process shown in fig. 6 (e.g., repeatedly generate predicted distribution categories and compare the predicted distribution categories to ground truth distribution categories) to train the distribution classification model 208.

The digital media clustering system 108 may perform a similar training process for the media clustering model 214. For example, the digital media clustering system 108 may analyze the training digital media repository with the media clustering model 214. The media clustering model 214 may predict one or more media clusters, and the digital media clustering system 108 may compare the predicted media clusters to ground truth media clusters (e.g., media clusters identified, approved, or generated by a user). The digital media clustering system 108 may train the media clustering model 214 by similarly comparing ground truth clusters to predicted clusters.

This training method is applicable to the various machine learning models discussed above. For example, digital media clustering system 108 may train a convolutional neural network or an LSTM neural network using the method described in fig. 6.

As described above, the digital media clustering system 108 may generate one or more user interfaces for displaying suggested digital media items. For example, fig. 7 illustrates a computing device 700 having a display 702 that includes a user interface 704 in accordance with one or more embodiments. The user interface 704 includes a cluster representative image 706 (i.e., representative of a card) and a digital media display area 716 that includes a plurality of digital media items.

The cluster representative image 706 corresponds to a suggested digital media cluster generated by the digital media clustering system 108. For example, the cluster representative image 706 represents a "face" cluster. Based on user interaction with the user interface 704, the digital media clustering system 108 may display different cluster representative images corresponding to various suggested digital media clusters. For example, in response to a swipe gesture (swipping capture), the digital media clustering system 108 may display the second cluster representative image 712 in the user interface 704. Similarly, in response to a swipe gesture in the opposite direction, the digital media clustering system 108 may display the third cluster representative images 710 and/or the fourth cluster representative images 708.

Each cluster representative image 706-712 depicts a digital image from a corresponding suggested digital media cluster. The digital media clustering system 108 generates suggested digital media clusters as described above (e.g., in conjunction with fig. 2-5). In particular, the digital media clustering system 108 analyzes digital media repositories on the computing device 700 and distribution categories that reflect network distribution history to determine a plurality of suggested digital media clusters. The digital media clustering system 108 then provides 712 suggested digital media clusters for display as cluster representative images 706.

The digital media clustering system 108 selects a cluster representative image 706 from the digital media items within the corresponding digital media cluster. The digital media clustering system 108 may select representative images using a variety of methods. In some embodiments, the digital media clustering system 108 selects the highest quality digital image as the cluster representative image (e.g., based on quality or composition). In one or more embodiments, the digital media clustering system 108 selects a cluster representative image (e.g., the digital media item with the highest confidence score from the media clustering model 214) based on the confidence scores. In some embodiments, the cluster representative images 706 vary over time. For example, the digital media clustering system 108 can loop through the digital media items in the respective digital media clusters and display each digital media item sequentially over time.

In connection with FIG. 7, each of the cluster representative images 706-712 is also a selectable user interface element. In practice, based on user interaction with the cluster representative images 706-712, the digital media clustering system 108 can display digital media items from the respective digital media clusters in the user interface 704. In fact, FIG. 7 shows the digital media items 714a-714n within the digital media display area 716 of the user interface 704 when the cluster representative image 706 is selected.

Each of the cluster representative images 706-712 also includes textual information. In particular, the cluster representative image 706- > 712 includes a textual description and one or more tags associated with the corresponding digital media cluster. For example, the cluster representative image 706 includes a text description 706a ("face") and tags 706b (# me), # self-timer (# selfield), and # take me back (# takeback)). The digital media clustering system 108 may determine this textual information based on various factors.

For example, the digital media clustering system 108 may determine the textual description 706a from the corresponding cluster. For example, the digital media clustering system 108 may determine the textual description "face" from the title or descriptor of the cluster (or a block or descriptor corresponding to the distribution category of the cluster). Similarly, the digital media clustering system 108 may determine the textual description 706a based on objects identified in the digital media (or the time or capture location of the digital media). For example, for a particular moment or travel cluster, the digital media clustering system 108 may include text indicating the location and date (e.g., a date range of travel) at which the digital media was captured. In this manner, the digital media clustering system 108 generates textual descriptions 706a-712 a.

Additionally, the digital media clustering system 108 may determine the label 706b from predefined labels or labels corresponding to previously distributed digital content. For example, the digital media clustering system 108 may utilize predefined, user-generated tags that map to particular clusters or categories. In some embodiments, the digital media clustering system 108 determines tags from the tags applied to previous social network posts. For example, the system identifies a threshold number of particular tags (e.g., from multiple network accounts) associated with a particular distribution category, and the digital media clustering system 108 may represent the tags in the images with clusters selected based on the distribution category. In some embodiments, the digital media clustering system 108 may generate tags based on the particular environment or situation of the client device. For example, upon detecting that the current day is Thursday, the digital media clustering system 108 may use the label "return Thursday" for clustering. In this manner, the digital media clustering system 108 generates the tags 706b-712 b.

As described above, the digital media clustering system 108 may provide suggested digital media clusters and suggested short-time digital content. Fig. 8A-8B illustrate a user interface for displaying suggested digital media clusters and suggested short-time digital content in accordance with one or more embodiments. In particular, fig. 8A illustrates a computing device 800 having a display 802, the display 802 illustrating a user interface 804. The user interface 804 includes a cluster representative image 806 and a plurality of digital media items 808 from a digital media cluster (e.g., "face") corresponding to the cluster representative image 806.

The user interface 804 includes a first selectable element 810 (labeled "Photos") and a second selectable element 812 (labeled "dynamic (Stories)"). Based on the user's interaction with the first selectable element 810, the digital media clustering system 108 provides for display one or more suggested digital media clusters. Based on the user's interaction with the second selectable element 812, the digital media clustering system 108 provides one or more suggested short-time digital content items for display. The digital media clustering system 108 may also transition between digital media clusters and short-lived digital content items based on alternative user interactions (e.g., a left swipe or a right swipe on the user interface 804).

For example, FIG. 8B shows the user interface 804 when the second selectable element 812 (or left swipe gesture) is selected. As shown in FIG. 8B, user interface 804 now includes an in-time digital content representative image 814. The user interface 804 also includes a plurality of short-time digital content items 816a-816n in a short-time digital content area 818.

Using the user interface 804, the digital media clustering system 108 can thus present various suggested digital media items. Further, the digital media clustering system 108 may allow a user to scroll through and view suggested digital media items. For example, the digital media clustering system 108 may provide an interactive timeline scrolling element (to skip to digital media items at a particular date or time). The digital media clustering system 108 may also provide magnified digital images for viewing based on user interaction with the digital media items (e.g., pressure touch or long selection gestures).

As described above, the digital media clustering system 108 may also generate a user interface for selecting and distributing digital content via a network system. For example, FIG. 9A illustrates a client device 900 that includes a display 902, the display 902 illustrating a user interface 904, the user interface 904 including a cluster representative image 906 and a plurality of digital media items 908a-808 n. As shown, based on the user's interaction with the digital media item 908n, the digital media clustering system 108 provides an indication that the digital media item 908n has been selected by the user (e.g., selected for distribution).

As shown in FIG. 9B, upon selection of the "Next" user interface element, digital media clustering system 108 also generates a user interface 910. The user interface 910 includes a text entry area for a user to enter text 912 as part of a digital message. The user interface 910 also includes elements for tagging (tag) (e.g., flag and notification) other users, adding locations identified in the digital message, and sharing the digital message via various social networking systems.

Upon identifying an indication that sharing of the digital communication is desired (e.g., a user selection of a "Share" user interface element), digital media clustering system 108 distributes the selected digital media items. For example, digital media clustering system 108 may generate social media posts and provide the social network posts for display. In particular, fig. 9 illustrates the digital media clustering system 108 generating and providing a user interface 914 that displays social media posts 920 (e.g., as part of a social network information flow (feed) to a network account of a user). As shown, the social network post includes a digital media item 908n and text 912.

Although fig. 9C illustrates distributing digital content related to social network posts, it should be understood that digital media clustering system 108 may distribute digital content in various alternative ways. For example, the digital media clustering system 108 may also distribute electronic communications including digital media to another client device via private messages (e.g., via instant messaging through the network system 110). Similarly, the digital media clustering system 108 may distribute digital media as short-time electronic communications.

Turning now to FIG. 10, additional details are provided regarding the components and capabilities of the digital media clustering system 108 in accordance with one or more embodiments. As shown, the digital media clustering system 108 is implemented by the server 106 and the client device 102 a. In other embodiments, the components of the digital media clustering system 108 may be implemented by a single device (e.g., server 106 or client device 102a) or other devices not shown (e.g., client devices 102b, 102 n).

As shown in connection with fig. 10, the server 106 includes a network system account manager 1002, a short-time digital content manager 1004, a distribution category generator 1006, a communication manager 1008, and a storage manager 1010 (which includes a network system profile 1010a, a network distribution history 1010b, a distribution classification model 1010c, training data 1010d, a distribution category 1010e, and a social graph 1010f that includes edges 1010g and nodes 1010 h). Further, as shown, the client device 102a includes a digital media cluster generation engine 1012, a distribution category manager 1014, a user interface manager 1018, a communication manager 1020, and a storage manager 1022 (which includes a digital media repository 1022a, digital media clusters 1022b, and a digital media cluster model 1022 c). Additional details regarding each of the components 1002-1022c will be provided in turn.

Network system account manager 1002 may identify, generate, monitor, and/or manage network accounts. For example, the network system may generate one or more network accounts for the user and generate information about the network accounts (according to privacy permissions). For example, the network system account manager 1002 may identify other users (e.g., friends) that are allowed to view or share digital content from or to the network account. The network system account manager 1002 may also monitor digital content distributed by network accounts to generate a network distribution history.

The short-time digital content manager 1004 may identify, generate, create, publish, distribute, remove, hide, delete, and/or monitor short-time digital content. As described above, the short-time digital content manager 1004 may share the short-time digital content, remove the digital content from one or more client devices, record the short-time digital content, and/or transmit the short-time digital content.

The distribution category generator 1006 may create, determine, identify, and/or generate one or more distribution categories. As described above, the distribution category generator 1006 may analyze the network distribution history to generate one or more distribution categories corresponding to network accounts. The communication manager 1008 may manage digital communications to and from the server 106. For example, the communication manager 1008 may transmit the distribution category to the client device 102 a. Similarly, the communication manager 1008 may receive digital content from the client device 102a and distribute the digital content to other client devices.

The storage manager 1010 maintains data for the digital media clustering system 108 at the server 106. The storage manager 1010 may maintain any type, size, or kind of data necessary to perform the functions of the digital media clustering system 108, including network system profiles 1010a, network distribution history 1010b, distribution classification models 1010c (e.g., convolutional neural networks or other machine learning models described above), training data 1010d (e.g., training network distribution history or other training data used to train machine learning models), distribution categories 1010e, and a social graph 1010f including edges 1010g and nodes 1010h (additional details regarding the social graph 1010f, edges 1010g, and nodes 1010h are provided below in connection with fig. 14-15).

The digital media cluster generation engine 1012 can identify, create, and/or determine digital media clusters. In particular, the digital media cluster generation engine can analyze the digital media repository 1022a to generate the digital media cluster 1022 b. As described above, the digital media clusters 1022b can include groups of digital media items that share common characteristics.

The distribution category manager 1014 can identify a distribution category. In particular, the distribution category manager 1014 can receive the distribution categories from the server 106. In addition, the distribution category manager 1014 may provide the distribution categories to the digital media cluster selector 1016. In this regard, the digital media cluster selector 1016 may identify, select, determine, and provide suggested digital media clusters (e.g., select a suggested digital media cluster from among the digital media clusters generated by the digital media cluster generation engine 1012). As described above, the digital media cluster selector 1016 may analyze the distribution categories (received from the server 106) to identify suggested digital media clusters that the user is most likely to distribute in future electronic communications.

The user interface manager 1018 may provide, manage, and/or control a graphical user interface (or simply "user interface"). In particular, user interface manager 1018 may generate and display a user interface via a display screen comprised of a plurality of graphical components, objects, and/or elements that allow a user to perform functions. The user interface manager 408 may present various types of information including text, digital media items, representative digital images corresponding to digital media clusters, images, video, audio, accounts, or other information (e.g., as shown above in connection with user interfaces 704, 804, 904, 910, 914).

The communication manager 1020 may manage digital communications to and from the client device 102 a. For example, the communication manager 1020 may transmit the selected digital media item to the server 106 (e.g., for publication via a network system). Similarly, the communication manager 1008 may receive the distribution category from the server 106.

The storage manager 1022 maintains data for the digital media clustering system 108 at the client device 102 a. The storage manager 1010 may maintain any type, size, or kind of data necessary to perform the functions of the digital media clustering system 108, including a digital media repository 1022a (e.g., a plurality of digital media items, such as a camera roll), digital media clusters 1022b, and a digital media cluster model 1022 c.

The components 1002-1022c may comprise software, hardware, or both. For example, the components 1002-1022c include one or more instructions stored on a computer-readable storage medium and executable by a processor of one or more computing devices (such as a client device or a server device). When executed by one or more processors, the computer-executable instructions of computing device 1000 and/or digital media clustering system 108 may cause the computing device to perform the feature learning methods described herein. Alternatively, the components 1002-1022c may comprise hardware, such as a dedicated processing device performing a certain function or group of functions. Alternatively, the components 1002-1022c may comprise a combination of computer-executable instructions and hardware.

Further, the components 1002-1022c can be implemented, for example, as one or more operating systems, one or more stand-alone applications, one or more modules of an application, one or more plug-ins, one or more library functions or functions invoked by other applications, and/or a cloud computing model. Thus, the components 1002-1022c can be implemented as stand-alone applications, such as desktop or mobile applications. Further, components 1002-1022c can be implemented as one or more web-based applications hosted on a remote server. The components 1002-1022c can also be implemented in a suite of mobile device applications or "apps".

1-10, corresponding text and examples provide several different systems, methods, techniques, components and/or apparatus of a digital media clustering system 108, according to one or more embodiments. In addition to the above, one or more embodiments may also be described in terms of flowcharts that include acts for achieving particular results. For example, fig. 11 and 12 show flowcharts of example sequences of actions in accordance with one or more embodiments. Additionally, fig. 11 and 12 may be performed with more or fewer acts. Further, the actions may be performed in a different order. Additionally, acts described herein may be repeated or performed in parallel with each other, or with different instances of the same or similar acts.

Although fig. 11 and 12 show a series of acts 1100 and a series of acts 1200 according to particular embodiments, alternative embodiments may omit, add to, reorder, and/or modify any acts shown. The series of acts of fig. 11 and 12 may be performed as part of a method. Alternatively, a non-transitory computer-readable medium may include instructions that, when executed by one or more processors, cause a computing device (e.g., a client device and/or a server device) to perform the series of acts of fig. 11 and 12. In further embodiments, the system performs the acts of fig. 11 and 12.

By way of example, figure 11 shows a flow diagram of a series of acts 1100 for generating and providing a predicted distribution category based on a network distribution history in accordance with one or more embodiments. As shown, series of acts 1100 includes an act 1110 of identifying a network distribution history corresponding to a network account. For example, act 1110 may include identifying a network system profile that includes a network distribution history corresponding to a network account of the user. In some embodiments, act 1110 includes identifying a network distribution history by identifying a plurality of historical posts of the social networking system, wherein the plurality of historical posts includes a plurality of historical digital media items.

As shown, a series of acts 1100 includes an act 1120 of generating a predicted distribution category corresponding to the network account. For example, act 1120 may include generating, using a distribution classification model, a predicted distribution category for future electronic communications corresponding to the user's network account based on the network distribution history. In some embodiments, act 1120 includes generating a predicted distribution category from a plurality of distribution categories based on the plurality of historical digital media items.

Series of acts 1100 also includes act 1130: the predicted distribution category is provided to a client device for generating a digital media cluster from a repository of digital media items. For example, act 1130 may include providing the predicted distribution category to a client device of the user, such that the client device provides, for display, a digital media cluster corresponding to the predicted distribution category from a digital media item repository on the client device.

In one or more embodiments, the series of acts 1100 further includes generating and providing a predicted distribution category to the client device without receiving information from the client device regarding a repository of digital media items on the client device. Further, in some embodiments, series of actions 1110 includes: receiving, from a client device, a user indication of a digital media item in a digital media cluster; and in response to receiving the user indication of the digital media item, distributing the digital media item to a plurality of client devices of a plurality of users via the network system.

Further, series of acts 1100 may include: identifying a plurality of short-term digital content items provided by a client device; generating a suggested short-time digital content item from a plurality of short-time digital content items; and providing the suggested short-time digital content item to the client device for display with a digital media cluster corresponding to the predicted distribution category from a digital media item repository on the client device.

Further, series of acts 1100 may also include: identifying an additional network account of the user corresponding to the network system; determining an additional network distribution history corresponding to an additional network account of the user (which corresponds to the network system); and generating, using the distribution classification model, additional predicted distribution categories for future electronic communications corresponding to the additional network accounts of the user based on the additional network distribution history. Additionally, the series of acts 1100 can include providing, to a client device of the user, additional predicted distribution categories corresponding to additional network accounts of the user, such that the client device provides, from a digital media item repository on the client device, additional digital media clusters corresponding to the additional predicted distribution categories for display.

In some embodiments, series of acts 1100 further comprises: determining suggested digital media style modifications based on a network distribution history; and providing the suggested digital media style modification to a client device of the user, such that the client device modifies a set of digital media items with the suggested digital media style modification and provides the set of digital media items for display in relation to the digital media cluster.

Additionally, fig. 12 shows a flow diagram of a series of acts 1200 for generating a digital media cluster based on a predicted distribution category in accordance with one or more embodiments. As shown, a series of acts 1200 includes an act 1210 of generating a plurality of digital media clusters from a digital media repository. For example, act 1210 may include generating a plurality of digital media clusters from a digital media repository on a client device of a user having a network account of a network system. In some embodiments, act 1210 includes: generating a first cluster corresponding to a first set of digital media items from a digital media repository based on time; generating a second cluster from the digital media repository corresponding to a second set of digital media items based on the location; and generating a third cluster based on the one or more public individuals identified in the third set of digital media items from the digital media repository using a facial recognition algorithm.

As shown, a series of acts 1200 includes an act 1220 of identifying a predicted distribution category. For example, act 1220 may include identifying a predicted distribution category for future posts of the user's network account, the predicted distribution category based on a network distribution history of the network account. In some embodiments, act 1220 includes receiving the predicted distribution category from the one or more remote servers without transmitting information regarding a repository of digital media items on the client device to the one or more remote servers.

The series of acts 1200 also includes an act 1230 of selecting a digital media cluster using the predicted distribution category. For example, act 1230 may include selecting a digital media cluster from a plurality of digital media clusters using the predicted distribution category. In some embodiments, act 1230 comprises comparing the predicted distribution category to a plurality of digital media clusters.

Additionally, series of acts 1200 includes an act 1240 of providing the digital media cluster for display. For example, act 1240 may include providing, via a user interface of the client device, the digital media cluster for display.

Further, series of acts 1200 includes acts 1250: digital media items in a digital media cluster are provided for distribution to client devices of a network system. For example, act 1250 may include providing the digital media item to one or more remote servers for distribution to client devices of other users of the network system based on the user's interaction with the digital media item in the digital media cluster. The series of acts 1200 may also include: identifying a suggested digital media pattern modification, wherein the suggested digital media pattern is based on a network distribution history; modifying a set of digital media items by applying the suggested digital media style; and providing the set of digital media items for display in relation to the digital media cluster.

Further, series of acts 1200 may also include: identifying an additional network account of the user corresponding to the network system; identifying an additional predicted distribution category for future electronic communications corresponding to the additional network account of the user, the additional predicted distribution category based on an additional network distribution history for the additional network account; and selecting an additional digital media cluster from the plurality of digital media clusters using the additional predictive distribution category. Further, in some embodiments, series of acts 1200 includes: identifying a suggested short-term digital content item from a plurality of short-term digital content items previously distributed from a network account of a user; and providing the suggested short-time digital content item for display with the digital media cluster.

As discussed in more detail below, embodiments of the present disclosure may comprise or utilize a special purpose or general-purpose computer including computer hardware (e.g., one or more processors and system memory). Embodiments within the scope of the present disclosure also include physical and other computer-readable media for carrying or storing computer-executable instructions and/or data structures. In particular, one or more processes described herein may be implemented, at least in part, as instructions embodied in a non-transitory computer-readable medium and executable by one or more computing devices (e.g., any of the media content access devices described herein). Generally, a processor (e.g., a microprocessor) receives instructions from a non-transitory computer-readable medium (e.g., a memory, etc.) and executes the instructions, thereby performing one or more processes, including one or more of the processes described herein.

Computer-readable media can be any available media that can be accessed by a general purpose or special purpose computer system. Computer-readable media storing computer-executable instructions are non-transitory computer-readable storage media (devices). A computer-readable medium carrying computer-executable instructions is a transmission medium. Thus, by way of example, and not limitation, embodiments of the disclosure can include at least two distinct computer-readable media: non-transitory computer-readable storage media (devices) and transmission media.

Non-transitory computer-readable storage media (devices) include RAM, ROM, EEPROM, CD-ROM, solid state drives ("SSDs"), flash memory, phase change memory ("PCM"), other types of memory, other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store desired program code means in the form of computer-executable instructions or data structures and which can be accessed by a general purpose or special purpose computer.

A "network" is defined as one or more data links that enable the transfer of electronic data between computer systems and/or modules and/or other electronic devices. When information is transferred or provided over a network or another communications connection (either hardwired, wireless, or a combination of hardwired or wireless) to a computer, the computer properly views the connection as a transmission medium. Transmission media can include a network and/or data links which can be used to carry desired program code means in the form of computer-executable instructions or data structures and which can be accessed by a general purpose or special purpose computer. Combinations of the above should also be included within the scope of computer-readable media.

Further, upon reaching various computer system components, program code means in the form of computer-executable instructions or data structures can be transferred automatically from transmission media to non-transitory computer-readable storage media (devices) (or vice versa). For example, computer-executable instructions or data structures received over a network or a data link may be buffered in RAM within a network interface module (e.g., a "NIC") and then eventually transferred to computer system RAM and/or to less volatile computer storage media (devices) at a computer system. Thus, it should be understood that non-transitory computer-readable storage media (devices) can be included in computer system components that also (or even primarily) utilize transmission media.

Computer-executable instructions comprise, for example, instructions and data which, when executed at a processor, cause a general purpose computer, special purpose computer, or special purpose processing device to perform a certain function or group of functions. In some embodiments, computer-executable instructions are executed on a general-purpose computer to turn the general-purpose computer into a special-purpose computer that implements the elements of the present disclosure. For example, the computer-executable instructions may be binaries, intermediate format instructions (such as assembly language), or even source code. Although the subject matter has been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the described features or acts described above. Rather, the described features and acts are disclosed as example forms of implementing the claims.

Those skilled in the art will appreciate that the disclosure may be practiced in network computing environments with many types of computer system configurations, including, personal computers, desktop computers, laptop computers, message processors, hand-held devices, multi-processor systems, microprocessor-based or programmable consumer electronics, network PCs, minicomputers, mainframe computers, mobile telephones, PDAs, tablets, pagers, routers, switches, and the like. The disclosure may also be practiced in distributed system environments where local and remote computer systems, which are linked (either by hardwired data links, wireless data links, or by a combination of hardwired and wireless data links) through a network, both perform tasks. In a distributed system environment, program modules may be located in both local and remote memory storage devices.

Embodiments of the present disclosure may also be implemented in a cloud computing environment. In this specification, "cloud computing" is defined as a model for enabling on-demand network access to a shared pool of configurable computing resources (shared pool). For example, cloud computing may be employed in the marketplace to provide ubiquitous and convenient on-demand access to a shared pool of configurable computing resources. The shared pool of configurable computing resources may be quickly provisioned via virtualization, released with little management workload or service provider interaction, and then expanded accordingly.

The cloud computing model may be composed of various features such as on-demand self-service, extensive network access, resource pooling, fast elasticity, billing services, and the like. The cloud computing model may also expose various service models, such as software as a service ("SaaS"), platform as a service ("PaaS"), and infrastructure as a service ("IaaS"). The cloud computing model may also be deployed using different deployment models (e.g., private cloud, community cloud, public cloud, hybrid cloud, etc.). In this specification and claims, a "cloud computing environment" is an environment that uses cloud computing.

Fig. 13 illustrates a block diagram of an exemplary computing device 1300, which exemplary computing device 1300 may be configured to perform one or more of the processes described above. It will be understood that one or more computing devices, such as computing device 1300, may implement one or more components of the digital media clustering system 108. As shown in fig. 13, computing device 1300 may include a processor 1302, memory 1304, storage 1306, I/O interface 1308, and communication interface 1310, which may be communicatively coupled through a communication infrastructure 1312. While an exemplary computing device 1300 is illustrated in fig. 13, the components illustrated in fig. 13 are not intended to be limiting. Additional or alternative components may be used in other embodiments. Further, in some embodiments, computing device 1300 may include fewer components than those shown in fig. 13. The components of the computing device 1300 shown in FIG. 13 will now be described in greater detail.

In one or more embodiments, processor 1302 includes hardware for executing instructions (e.g., those making up a computer program). By way of example, and not limitation, to execute instructions, processor 1302 may retrieve (or fetch) instructions from internal registers, internal caches, memory 1304, or storage device 1306, and decode and execute them. In one or more embodiments, processor 1302 may include one or more internal caches for data, instructions, or addresses. By way of example, and not limitation, processor 1302 may include one or more instruction caches, one or more data caches, and one or more Translation Lookaside Buffers (TLBs). The instructions in the instruction cache may be copies of the instructions in memory 1304 or storage 1306.

The memory 1304 may be used to store data, metadata, and programs that are executed by the processor. The memory 1304 may include one or more of volatile and non-volatile memory, such as random access memory ("RAM"), read only memory ("ROM"), solid state disk ("SSD"), flash memory, phase change memory ("PCM"), or other types of data storage devices. The memory 1304 may be internal or distributed.

Storage 1306 includes storage for storing data or instructions. By way of example, and not limitation, storage device 1306 may comprise the aforementioned non-transitory storage media. Storage 1306 may include a Hard Disk Drive (HDD), a floppy disk drive, flash memory, an optical disk, a magneto-optical disk, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Storage devices 1306 may include removable or non-removable (or fixed) media, where appropriate. Storage device 1306 may be internal or external to computing device 1300. In one or more embodiments, storage device 1306 is non-volatile solid-state memory. In other embodiments, storage device 1306 includes Read Only Memory (ROM). Where appropriate, the ROM may be mask-programmed ROM, programmable ROM (prom), erasable prom (eprom), electrically erasable prom (eeprom), electrically variable ROM (earom), or flash memory, or a combination of two or more of these.

I/O interface 1308 allows a user to provide input to computing device 1300, receive output from computing device 1300, and otherwise transfer data to and from computing device 1300. The I/O interface 1308 may include a mouse, keypad (keypad) or keyboard, touch screen, camera, optical scanner, network interface, modem, other known I/O devices, or a combination of these I/O interfaces. I/O interface 1308 may include one or more devices for presenting output to a user, including but not limited to a graphics engine, a display (e.g., a display screen), one or more output drivers (e.g., a display driver), one or more audio speakers, and one or more audio drivers. In certain embodiments, I/O interface 1308 is configured to provide graphical data to a display for presentation to a user. The graphical data may represent one or more graphical user interfaces and/or any other graphical content that may serve a particular implementation.

Communication interface 1310 may include hardware, software, or both. Regardless, the communication interface 1310 may provide one or more interfaces for communication (e.g., packet-based communication) between the computing device 1300 and one or more other computing devices or networks. By way of example, and not limitation, communication interface 1310 may include a Network Interface Controller (NIC) or network adapter for communicating with an ethernet or other wire-based network, or a wireless NIC (wnic) or wireless adapter for communicating with a wireless network, such as WI-FI.

Additionally or alternatively, communication interface 1310 may facilitate communication with an ad hoc network, a Personal Area Network (PAN), a Local Area Network (LAN), a Wide Area Network (WAN), a Metropolitan Area Network (MAN), or one or more portions of the internet, or a combination of two or more of these. One or more portions of one or more of these networks may be wired or wireless. By way of example, communication interface 1310 may facilitate communication with a wireless PAN ("WPAN") (e.g., a bluetooth WPAN), a WI-FI network, a WI-MAX network, a cellular telephone network (e.g., a global system for mobile communications ("GSM") network), or other suitable wireless network, or a combination of two or more of these.

Additionally, communication interface 1310 may facilitate communication of various communication protocols. Examples of communication protocols that may be used include, but are not limited to, data transmission media, communication devices, transmission control protocol ("TCP"), internet protocol ("IP"), file transfer protocol ("FTP"), telnet, hypertext transfer protocol ("HTTP"), hypertext transfer protocol secure ("HTTPs"), session initiation protocol ("SIP"), simple object access protocol ("SOAP"), extensible markup language ("XML") and variants thereof, simple mail transfer protocol ("SMTP"), real-time transfer protocol ("RTP"), user datagram protocol ("UDP"), global system for mobile communications ("GSM") technology, code division multiple access ("CDMA") technology, time division multiple access ("TDMA") technology, short message service ("SMS"), multimedia message service ("MMS"), radio frequency ("RF") signaling technology, long term evolution ("LTE") technology, and the like, Wireless communication technologies, in-band and out-of-band signaling technologies, and other suitable communication networks and technologies.

The communication infrastructure 1312 may include hardware, software, or both that couple the components of the computing device 1300 to each other. By way of example, and not limitation, communication infrastructure 1312 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a front-side bus (FSB), a hyper random port (ht) interconnect, an Industry Standard Architecture (ISA) bus, an INFINIBAND interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (extended) bus, a Serial Advanced Technology Attachment (SATA) bus, a video electronics standards association local (VLB) bus, or any other suitable bus or combination thereof.

As described above, in one or more embodiments, the digital media clustering system 108 operates in conjunction with a social networking system (e.g., the networking system 110 described with reference to FIG. 1). In addition to the description given above, a social networking system may enable its users (such as individuals or organizations) to interact with the system and with each other. The social networking system may utilize input from the user to create and store a user profile (user profile) associated with the user in the social networking system. The user profile may include demographic information, communication channel information, and information about personal interests of the user. The social networking system may also create and store records of the user's relationships with other users of the social networking system with input from the user, as well as provide services (e.g., posts, photo sharing, event organization, messaging, games, or advertisements) to facilitate social interactions between or among users.

The social-networking system may store records of users and relationships between users in a social graph that includes a plurality of nodes and a plurality of edges connecting the nodes. The nodes may include a plurality of user nodes and a plurality of concept nodes. The user nodes of the social graph may correspond to users of the social-networking system. A user may be an individual (human user), an entity (e.g., an enterprise, a company, or a third party application), or a group (e.g., a group of individuals or entities). The user nodes corresponding to the users may include information provided by the users and information collected by various systems, including social-networking systems.

For example, the user may provide his or her name, profile picture, city of residence, contact information, date of birth, gender, marital status, family status, employment, educational background, preferences, interests, and other demographic information to be included in the user node. Each user node of the social graph may have a corresponding web page (often referred to as a profile page). In response to a request including a user name, the social networking system may access the user node corresponding to the user name and build a profile page including the name, profile pictures, and other information associated with the user. Based on one or more privacy settings of the first user and the relationship between the first user and the second user, the profile page of the first user may display all or part of the information of the first user to the second user.

The concept nodes may correspond to concepts of a social networking system. For example, a concept may represent a real-world entity, such as a movie, song, sports team, celebrity, group, restaurant, place, or place. An administrative user of a concept node corresponding to a concept may create or update the concept node by providing information of the concept (e.g., by filling out an online form) such that the social networking system associates the information with the concept node. For example, but not by way of limitation, information associated with a concept may include a name or title, one or more images (e.g., images of the cover of a book), a website (e.g., a URL address), or contact information (e.g., a phone number, an email address). Each concept node of the social graph may correspond to a web page. For example, in response to a request that includes a name, the social networking system may access the concept node corresponding to the name and construct a web page that includes the name and other information associated with the concept.

An edge between a pair of nodes may represent a relationship between the pair of nodes. For example, an edge between two user nodes may represent a friendship between two users. For another example, the social networking system may build a web page (or structured document) of concept nodes (e.g., restaurants, celebrities) that incorporate one or more selectable options or selectable elements (e.g., "like", "check in"). The user may access the page using a web browser hosted by the user's client device and select a selectable option or selectable element, causing the client device to send a request to the social networking system to create an edge between the user node of the user and the concept node of the concept, the edge indicating a relationship between the user and the concept (e.g., the user checked in at a restaurant, or the user "liked" a celebrity).

As an example, a user may provide (or change) his or her city of residence, enabling the social networking system to create an edge between a user node corresponding to the user and a concept node corresponding to a city that the user declares as his or her city of residence. Furthermore, the degree of separation (degree of separation) between any two nodes is defined as the minimum number of hops required to traverse the social graph from one node to another. The degree of separation between two nodes can be thought of as a measure of the correlation between users or concepts represented by the two nodes in the social graph. For example, two users having user nodes (i.e., level one nodes) directly connected by an edge may be described as "connected (connected) users" or "friends". Similarly, two users having user nodes (i.e., secondary nodes) connected only by another user node may be described as "friends of friends".

Social networking systems may support a variety of applications, such as photo sharing, online calendars and events, games, instant messaging, and advertisements. For example, the social networking system may also include media sharing capabilities. In addition, the social networking system may allow users to post photos and other multimedia content items to user profile pages (commonly referred to as "wall posts" or "timeline posts") or in albums, both of which may be accessed by other users of the social networking system according to user-configured privacy settings. The social networking system may also allow a user to configure events. For example, the first user may configure the event with attributes including the time and date of the event, the location of the event, and other users invited to attend the event. The invited user may receive the invitation to the event and respond (e.g., by accepting the invitation or rejecting it). In addition, the social networking system may allow users to maintain personal calendars. Similar to events, calendar entries may include time, date, location, and other user identities.

FIG. 14 illustrates an example network environment 1400 of a social networking system. Network environment 1400 includes client devices 1406, network systems 1402 (e.g., social networking systems and/or electronic messaging systems), and third-party systems 1408, which are interconnected via a network 1404. Although fig. 14 illustrates a particular arrangement of client device 1406, network system 1402, third-party system 1408, and network 1404, this disclosure contemplates any suitable arrangement of client device 1406, network system 1402, third-party system 1408, and network 1404. By way of example and not limitation, two or more of client device 1406, network system 1402, and third-party system 1408 may bypass network 1404 and connect directly to each other. As another example, two or more of client device 1406, network system 1402, and third-party system 1408 may all or partially be physically or logically co-located with each other. Moreover, although fig. 14 illustrates a particular number of client devices 1406, network systems 1402, third-party systems 1408, and networks 1404, this disclosure contemplates any suitable number of client devices 1406, network systems 1402, third-party systems 1408, and networks 1404. By way of example, and not limitation, network environment 1400 may include a plurality of client devices 1406, a network system 1402, a third-party system 1408, and a network 1404.

The present disclosure contemplates any suitable network 1404. By way of example and not limitation, one or more portions of network 1404 may include an ad hoc network, an intranet, an extranet, a Virtual Private Network (VPN), a Local Area Network (LAN), a wireless LAN (wlan), a Wide Area Network (WAN), a wireless WAN (wwan), a Metropolitan Area Network (MAN), a portion of the internet, a portion of the Public Switched Telephone Network (PSTN), a cellular telephone network, or a combination of two or more of these. The network 1404 may include one or more networks 1404.

The links may connect client device 1406, network system 1402, and third party system 1408 to communication network 1404 or to each other. The present disclosure contemplates any suitable links. In particular embodiments, the one or more links include one or more wired (e.g., Digital Subscriber Line (DSL) or cable-based data service interface specification (DOCSIS)) links, wireless (e.g., Wi-Fi or Worldwide Interoperability for Microwave Access (WiMAX)) links, or optical (e.g., Synchronous Optical Network (SONET) or Synchronous Digital Hierarchy (SDH)) links. In particular embodiments, the one or more links each include an ad hoc network, an intranet, an extranet, a VPN, a LAN, a WLAN, a WAN, a WWAN, a MAN, a portion of the internet, a portion of the PSTN, a cellular technology-based network, a satellite communication technology-based network, another link, or a combination of two or more such links. The links need not be the same throughout network environment 1400. The one or more first links may differ in one or more respects from the one or more second links.

In particular embodiments, client device 1406 may be an electronic device that includes hardware, software, or embedded logic components, or a combination of two or more such components, and is capable of performing the appropriate functions implemented or supported by client device 1406. By way of example and not limitation, client device 1406 may include a computer system such as an augmented reality display device, a desktop computer, a notebook or laptop computer, a netbook, a tablet computer, an e-book reader, a GPS device, a camera, a Personal Digital Assistant (PDA), a handheld electronic device, a cellular telephone, a smartphone, other suitable electronic device, or any suitable combination thereof. The present disclosure contemplates any suitable client device 1406. Client device 1406 may enable a network user at client device 1406 to access network 1404. The client device 1406 may enable its user to communicate with other users at other client devices 1406.

In particular embodiments, client device 1406 may include a web browser, such as MICROSOFT INTERNET EXPLORER, GOOGLE CHROME, or MOZILLA FIREFOX, and possibly one or more add-ons, plug-ins, or other extensions, such as TOOLBAR or YAHOO TOOLBAR. A user at client device 1406 may enter a uniform resource locator ("URL") or other address directing the web browser to a particular server, such as a server or a server associated with third-party system 1408, and the web browser may generate a hypertext transfer protocol ("HTTP") request and communicate the HTTP request to the server. The server may accept the HTTP request and communicate one or more hypertext markup language (HTML) files to the client device 1406 in response to the HTTP request. The client device 1406 may render the web page based on the HTML files from the server for presentation to the user. The present disclosure contemplates any suitable web page files. By way of example and not limitation, web pages may be rendered from HTML files, extensible hypertext markup language (XHTML) files, or extensible markup language (XML) files, according to particular needs. Such pages may also execute scripts, such as, without limitation, scripts written in JAVASCRIPT, JAVA, MICROSOFT SILVERLIGHT, combinations of markup languages and scripts (e.g., AJAX (asynchronous JAVASCRIPT and XML)), and the like. Herein, reference to a web page includes one or more corresponding web page files (which a browser may use to render the web page), and vice versa, where appropriate.

In particular embodiments, network system 1402 may be a network addressable computing system that may host an online social network. The network system 1402 may generate, store, receive, and send social network data (e.g., user profile data, concept profile data, social graph information, or other suitable data related to an online social network). Network system 1402 may be accessed by other components of network environment 1400, either directly or via network 1404. In particular embodiments, network system 1402 may include one or more servers. Each server may be a single server (unity server) or a distributed server spanning multiple computers or multiple data centers. The servers may be of various types, such as, for example and without limitation, a web server, a news server, a mail server, a messaging server, an advertising server, a file server, an application server, an exchange server, a database server, a proxy server, another server suitable for performing the functions or processes described herein, or any combination thereof. In particular embodiments, each server may include hardware, software, or embedded logic components, or a combination of two or more such components for performing the appropriate functions implemented or supported by the server. In certain embodiments, the network system 1402 may include one or more data stores. The data store may be used to store various types of information. In particular embodiments, the information stored in the data store may be organized according to a particular data structure. In particular embodiments, each data store may be a relational database, column (columnar) database, a relational database, or other suitable database. Although this disclosure describes or illustrates a particular type of database, this disclosure contemplates any suitable type of database. Particular embodiments may provide an interface that enables client device 1406, network system 1402, or third party system 1408 to manage, retrieve, modify, add, or delete information stored in the data store.

In particular embodiments, network system 1402 may store one or more social graphs in one or more data stores. In particular embodiments, the social graph may include a plurality of nodes, which may include a plurality of user nodes (each corresponding to a particular user) or a plurality of concept nodes (each corresponding to a particular concept), and a plurality of edges connecting the nodes. Network system 1402 may provide users of an online social network with the ability to communicate and interact with other users. In particular embodiments, users may join an online social network via the network system 1402, and then add connections (e.g., relationships) to a number of other users in the network system 1402 to whom they want to be related. Herein, the term "friend" may refer to any other user of the network system 1402 with whom the user forms an association, or relationship via the network system 1402.

In particular embodiments, the network system 1402 may provide users with the ability to take actions on various types of items or objects supported by the network system 1402. By way of example and not by way of limitation, items and objects may include groups or social networks to which a user of network system 1402 may belong, events or calendar entries that may be of interest to the user, computer-based applications that may be used by the user, transactions that allow the user to purchase or sell goods via a service, interactions with advertisements that the user may perform, or other suitable items or objects. The user may interact with anything that can be represented in the network system 1402 or by a system external to the third-party system 1408, the third-party system 1408 being separate from the network system 1402 and coupled to the network system 1402 via the network 1404.

In particular embodiments, network system 1402 can link various entities. By way of example and not limitation, network system 1402 enables users to interact with each other and receive content from third-party systems 1408 or other entities, or to allow users to interact with these entities through an Application Programming Interface (API) or other communication channel.

In particular embodiments, third-party system 1408 may include one or more types of servers, one or more data stores, one or more interfaces (including but not limited to APIs), one or more web services, one or more content sources, one or more networks, or any other suitable components (e.g., with which a server may communicate). Third-party system 1408 may be operated by an entity other than the entity operating network system 1402. However, in particular embodiments, network system 1402 and third-party system 1408 may operate in conjunction with each other to provide social networking services to users of network system 1402 or third-party system 1408. In this sense, the network system 1402 may provide a platform or backbone network that other systems (e.g., third-party system 1408) may use to provide social networking services and functionality to users across the internet.

In particular embodiments, third-party system 1408 may include a third-party content object provider. The third-party content object provider may include one or more sources of content objects that may be delivered to the client device 1406. By way of example and not limitation, content objects may include information about things or activities of interest to a user, such as, for example, movie show times, movie reviews, restaurant menus, product information and reviews, or other suitable information. As another example and not by way of limitation, the content object may include an incentive content object (e.g., a coupon, discount coupon, gift coupon, or other suitable incentive object).

In certain embodiments, the network system 1402 also includes user-generated content objects that may enhance user interaction with the network system 1402. The user-generated content may include any content that a user may add, upload, send, or "publish" to the network system 1402. By way of example, and not by way of limitation, a user transmits a post from a client device 1406 to a network system 1402. Posts may contain data such as status updates or other textual data, location information, photos, videos, links, music, or other similar data or media. Content may also be added to network system 1402 by third parties through "communication channels" (e.g., dynamic messages (news feed) or streams).

In particular embodiments, network system 1402 may include various servers, subsystems, programs, modules, logs, and data stores. In particular embodiments, network system 1402 may include one or more of the following: web servers, action recorders, API request servers, relevance and ranking engines, content object classifiers, notification controllers, action logs, third-party content object exposure logs, inference modules, authorization/privacy servers, search modules, ad-targeting modules, user interface modules, user profile stores, relational stores, third-party content stores, or location stores. The network system 1402 may also include suitable components, such as network interfaces, security mechanisms, load balancers, failover servers, management and network operations consoles, other suitable components, or any suitable combination thereof. In particular embodiments, network system 1402 may include one or more user profile stores for storing user profiles. The user profile may include, for example, biographical information, demographic information, behavioral information, social information, or other types of descriptive information (e.g., work experience, educational history, hobbies or preferences, interests, affinities, or locations). The interest information may include interests related to one or more categories. The categories may be general or specific. By way of example and not by way of limitation, if a user "likes" an article about a brand of shoes, that category may be the brand, or the general category of "shoes" or "clothing". The associative memory may be used to store information that is associative with the user. The relational information may indicate users who have similar or common work experiences, group memberships, hobbies, educational history, or are related or share common attributes in any manner. The relational information may also include user-defined relations between different users and the content (internal and external). A web server may be used to link network system 1402 to one or more client devices 1406 or one or more third party systems 1408 via network 1404. The web server may include a mail server or other messaging functionality for receiving and routing messages between the network system 1402 and one or more client devices 1406. The API request server may allow third party system 1408 to access information from network system 1402 by calling one or more APIs. The action recorder may be used to receive communications from the web server regarding the user's actions on or off the network system 1402. In conjunction with the action log, a third-party content object log of user exposure to third-party content objects may be maintained. The notification controller may provide information about the content object to the client device 1406. The information may be pushed to the client device 1406 as a notification, or the information may be pulled from the client device 1406 in response to a request received from the client device 1406. The authorization server may be used to enforce one or more privacy settings of the user of the network system 1402. The privacy settings of the user determine how particular information associated with the user may be shared. The authorization server may allow users to opt-in or opt-out to have their actions recorded by the network system 1402 or shared with other systems (e.g., third party system 1408), for example, by setting appropriate privacy settings. The third-party content object store may be used to store content objects received from third parties (e.g., third-party system 1408). The location store may be used to store location information received from a client device 1406 associated with a user. The advertisement pricing module may combine social information, current time, location information, or other suitable information to provide relevant advertisements to the user in the form of notifications.

FIG. 15 illustrates an example social graph 1500. In particular embodiments, network system 1402 may store one or more social graphs 1500 in one or more data stores. In particular embodiments, the social graph 1500 may include a plurality of nodes, which may include a plurality of user nodes 1502 or a plurality of concept nodes 1504, and a plurality of edges 1506 connecting the nodes.

For purposes of teaching, the example social graph 1500 shown in FIG. 15 is shown in a two-dimensional visual map representation. In particular embodiments, the network system 1402, client device 1406, or third party system 1408 may access the social graph 1500 and related social graph information for suitable applications. The nodes and edges of the social graph 1500 may be stored as data objects, for example, in a data store (e.g., a social graph database). Such data stores may include one or more searchable or queryable indexes of nodes or edges of the social graph 1500.

In particular embodiments, the user node 1502 may correspond to a user of the network system 1402. By way of example and not limitation, a user may be an individual (human user), an entity (e.g., an enterprise, company, or third-party application), or a community (e.g., of individuals or entities) that interacts or communicates with or through the network system 1402. In particular embodiments, when a user registers for an account with network system 1402, network system 1402 may create a user node 1502 corresponding to the user and store user node 1502 in one or more data stores. Users and user nodes 1502 described herein may refer to registered users and user nodes 1502 associated with registered users, where appropriate. Additionally or alternatively, the users and user nodes 1502 described herein may refer to users that are not registered with the network system 1402, where appropriate. In particular embodiments, the user node 1502 may be associated with information provided by a user or information collected by various systems, including the network system 1402. By way of example and not by way of limitation, a user may provide his or her name, profile picture, contact information, date of birth, gender, marital status, family status, occupation, educational background, preferences, interests, or other demographic information. In particular embodiments, user node 1502 may be associated with one or more data objects corresponding to information associated with a user. In particular embodiments, user nodes 1502 may correspond to one or more web pages.

In particular embodiments, concept node 1504 may correspond to a concept. By way of example and not by way of limitation, concepts may correspond to a location (e.g., a movie theater, restaurant, landmark, or city); a website (e.g., a website associated with network system 1402 or a third-party website associated with a web application server); an entity (e.g., an individual, a business, a group, a sports team, or a celebrity); a resource (e.g., an audio file, a video file, a digital photograph, a text file, a structured document, or an application) that may be located within network system 1402 or on an external server (e.g., a web application server); real estate or intellectual property (e.g., sculptures, paintings, movies, games, songs, ideas, photographs, or written works); playing a game; moving; an idea or theory; another suitable concept; or two or more such concepts. Concept nodes 1504 may be associated with information for concepts provided by users or information collected by various systems, including network system 1402. By way of example, and not by way of limitation, information for a concept may include a name or title; one or more images (e.g., of the cover of a book); location (e.g., address or geographic location); a website (which may be associated with a URL); contact information (e.g., a phone number or an email address); other suitable conceptual information; or any suitable combination of such information. In particular embodiments, concept node 1504 may be associated with one or more data objects that correspond to information associated with concept node 1504. In particular embodiments, concept node 1504 may correspond to one or more web pages.

In particular embodiments, the nodes in the social graph 1500 may represent or be represented by web pages (which may be referred to as "profile pages"). The profile page may be hosted by the network system 1402 or accessible to the network system 1402. The profile page may also be hosted on a third-party website associated with third-party system 1408. By way of example and not limitation, a profile page corresponding to a particular external web page may be the particular external web page, and the profile page may correspond to the particular concept node 1504. The profile page may be viewable by all other users or a selected subset of the other users. By way of example and not by way of limitation, user node 1502 may have a corresponding user profile page in which a corresponding user may add content, make statements, or otherwise express himself or herself. As another example and not by way of limitation, concept nodes 1504 may have respective concept profile pages in which one or more users may add content, make statements, or express themselves, particularly with respect to concepts corresponding to concept nodes 1504.

In particular embodiments, concept node 1504 may represent a third-party webpage or resource hosted by third-party system 1408. The third party webpage or resource may include content representing an action or activity, selectable icons or other interactable objects (which may be implemented, for example, in JavaScript, AJAX, or PHP code), and other elements. By way of example and not limitation, the third-party webpage may include selectable icons (e.g., "like," "check-in," "eat," "recommend"), or other suitable actions or activities. A user viewing the third-party webpage may perform an action by selecting one of the icons (e.g., "eat"), causing the client device 1406 to send a message to the network system 1402 indicating the user's action. In response to the message, the network system 1402 can create an edge (e.g., an "eat" edge) between a user node 1502 corresponding to the user and a concept node 1504 corresponding to a third-party webpage or resource, and store the edge 1506 in one or more data stores.

In particular embodiments, a pair of nodes in social graph 1500 may be connected to each other by one or more edges 1506. An edge 1506 connecting a pair of nodes may represent a relationship between the pair of nodes. In particular embodiments, edges 1506 may include or represent one or more data objects or attributes corresponding to a relationship between a pair of nodes. By way of example and not by way of limitation, the first user may indicate that the second user is a "friend" of the first user. In response to the indication, the network system 1402 may send a "friend request" to the second user. If the second user confirms the "friend request," the network system 1402 may create an edge 1506 in the social graph 1500 that connects the user node 1502 of the first user to the user node 1502 of the second user, and store the edge 1506 as social graph information in one or more data stores. In the example of fig. 15, the social graph 1500 includes an edge 1506 indicating a friendship between the user nodes 1502 of user "a" and user "B", and an edge indicating a friendship between the user nodes 1502 of user "C" and user "B". Although this disclosure describes or illustrates a particular edge 1506 having particular properties connecting a particular user node 1502, this disclosure contemplates any suitable edge 1506 having any suitable properties connecting user nodes 1502. By way of example and not limitation, the edge 1506 may represent a friendship, family relationship, business or employment relationship, fan relationship, follower (follower) relationship, visitor relationship, subscriber relationship, superior/subordinate relationship, reciprocal relationship, non-reciprocal relationship, another suitable type of relationship, or two or more such relationships. Further, while this disclosure generally describes nodes as being connected, this disclosure also describes users or concepts as being related. Herein, references to connected users or concepts may refer to nodes in the social graph 1500 that correspond to those users or concepts connected by one or more edges 1506, where appropriate.

In particular embodiments, an edge 1506 between a user node 1502 and a concept node 1504 may represent a particular action or activity performed by a user associated with user node 1502 towards a concept associated with concept node 1504. By way of example and not by way of limitation, as shown in fig. 15, a user may "like", "attend", "play", "listen", "cook", "work on", or "watch" concepts, each of which may correspond to an edge type or subtype. The concept profile page corresponding to the concept node 1504 may include, for example, an optional "check-in" icon (e.g., a clickable "check-in" icon) or an optional "add to favorites" icon. Similarly, after the user clicks on these icons, the network system 1402 may create a "favorites" edge or a "check-in" edge in response to the user action corresponding to the respective action. As another example and not by way of limitation, a user (user "C") may listen to a particular song ("imagining") using a particular application (sports (SPOTIFY), which is an online music application). In this case, the network system 1402 may create a "listen" edge 1506 and a "use" edge (as shown in FIG. 15) between the user node 1502 corresponding to the user and the concept node 1504 corresponding to the song and the application to indicate that the user listened to the song and used the application. In addition, the network system 1402 may create "play" edges 1506 (as shown in FIG. 15) between the concept nodes 1504 corresponding to songs and applications to indicate that a particular song is played by a particular application. In this case, the "play" edge 1506 corresponds to an action performed by the external application (SPOTIFY) on the external audio file (song "Imagine"). Although this disclosure describes a particular edge 1506 with particular properties connecting the user node 1502 and the concept node 1504, this disclosure contemplates any suitable edge 1506 with any suitable properties connecting the user node 1502 and the concept node 1504. Further, while this disclosure describes edges between user node 1502 and concept node 1504 representing a single relationship, this disclosure contemplates edges between user node 1502 and concept node 1504 representing one or more relationships. By way of example and not by way of limitation, the edge 1506 may indicate that the user likes and uses a particular concept. Alternatively, another edge 1506 may represent each type of relationship (or multiple single relationships) between a user node 1502 and a concept node 1504 (between the user node 1502 of user "E" and the concept node 1504 of "SPOTIFY" as shown in fig. 15).

In particular embodiments, the network system 1402 may create an edge 1506 between the user node 1502 and the concept node 1504 in the social graph 1500. By way of example and not limitation, a user viewing a concept profile page (e.g., by using a web browser or a dedicated application hosted by the user's client device 1406) may indicate that he or she likes the concept represented by the concept node 1504 by clicking or selecting a "like" icon, which may cause the user's client device 1406 to send a message to the network system 1402 indicating that the user likes the concept associated with the concept profile page. In response to the message, network system 1402 may create an edge 1506 between user node 1502 associated with the user and concept node 1504, as shown by the "like" edge 1506 between the user and concept node 1504. In certain embodiments, the network system 1402 may store the edges 1506 in one or more data stores. In particular embodiments, the edge 1506 may be automatically formed by the network system 1402 in response to a particular user action. As an example and not by way of limitation, if a first user uploads a picture, watches a movie, or listens to a song, an edge 1506 may be formed between a user node 1502 corresponding to the first user and concept nodes 1504 corresponding to those concepts. Although this disclosure describes forming a particular edge 1506 in a particular manner, this disclosure contemplates forming any suitable edge 1506 in any suitable manner.

In particular embodiments, the advertisement may be text (which may be HTML-linked), one or more images (which may be HTML-linked), one or more videos, audio, one or more ADOBE FLASH files, suitable combinations of these, or any other suitable advertisement in any suitable digital format presented on one or more web pages, in one or more emails, or in conjunction with search results requested by the user. Additionally or alternatively, the advertisement may be one or more sponsored dynamic (e.g., a message dynamic or ticker item on the network system 1402). The sponsored dynamics may be a user's social actions (e.g., "like" a page, "like" or comment on a page, reply to an event associated with a page (RSVP), vote on a question posted on a page, check in somewhere, use an application or play a game, or "like" or share a website), an advertiser, for example, promoting a social action by causing the social action to be presented within a predetermined area of a user's profile page or other page, presented with additional information associated with the advertiser, promoted in order (bump up) or otherwise highlighted in other users ' dynamic messages or instant dynamics, or otherwise promoting a social action. Advertisers may pay to promote social actions. By way of example and not limitation, advertisements may be included among search results of a search results page in which sponsored content is promoted over non-sponsored content.

In particular embodiments, an advertisement may be requested for display in a social networking system web page, a third party web page, or other page. The advertisement may be displayed in a dedicated portion of the page, such as in a banner (banner) area at the top of the page, in a column at the side of the page, in a GUI of the page, in a pop-up window, in a drop-down menu, in an input field of the page, on top of the content of the page, or elsewhere with respect to the page. Additionally or alternatively, the advertisement may be displayed in the application. The advertisements may be displayed within a dedicated page, requiring the user to interact with or view the advertisements before the user can access the page or utilize the application. The user may view the advertisement, for example, through a web browser.

The user may interact with the advertisement in any suitable manner. The user may click on or otherwise select an advertisement. By selecting the advertisement, the user may be directed to (or by a browser or other application being used by the user) a page associated with the advertisement. At the page associated with the advertisement, the user may take additional actions, such as purchasing a product or service associated with the advertisement, receiving information associated with the advertisement, or subscribing to a newsletter associated with the advertisement. By selecting an advertising component (e.g., a "play button"), an advertisement with audio or video may be played. Alternatively, by selecting an advertisement, the network system 1402 may perform or modify a particular action of the user.

The advertisement may also include social networking system functionality with which the user may interact. By way of example and not by way of limitation, an advertisement may enable a user to "endorse" or otherwise approve the advertisement by selecting an icon or link associated with an endorsement. As another example and not by way of limitation, an advertisement may enable a user to search for content related to an advertiser (e.g., by executing a query). Similarly, a user may share an advertisement with another user (e.g., via network system 1402) or Reply (RSVP) to an event associated with the advertisement (e.g., via network system 1402). Additionally or alternatively, the advertisement may include social networking system context that is directed to the user. By way of example and not by way of limitation, the advertisement may display information about the user's friends who have taken an action within the network system 1402 that is associated with the subject matter of the advertisement.

In particular embodiments, network system 1402 may determine social-graph affinity (which may be referred to herein as "affinity") of various social-graph entities with respect to each other. The affinity may represent a strength of relationship or a level of interest between particular objects associated with the online social network (such as users, concepts, content, actions, advertisements), other objects associated with the online social network, or any suitable combination thereof. Affinity may also be determined with respect to objects associated with third-party system 1408 or other suitable system. An overall affinity to the social graph entity may be established for each user, topic, or content type. The overall affinity may change based on continuous monitoring of actions or relationships associated with the social graph entity. Although this disclosure describes determining a particular affinity in a particular manner, this disclosure contemplates determining any suitable affinity in any suitable manner.

In particular embodiments, network system 1402 may use affinity coefficients (which may be referred to herein as "coefficients") to measure or quantify social graph affinity. The coefficient may represent or quantify a strength of a relationship between particular objects associated with the online social network. The coefficient may also represent a probability or function that measures a predicted probability that the user will perform a particular action based on the user's interest in the action. In this way, future actions of the user may be predicted based on previous actions of the user, where the coefficients may be calculated based at least in part on a history of user actions. The coefficients may be used to predict any number of actions that may be within or outside of the online social network. By way of example, and not by way of limitation, such actions may include various types of communications, such as sending messages, posting content, or commenting on content; various types of observation actions, such as accessing or viewing a profile page, media, or other suitable content; various types of consistent information about two or more social graph entities, such as being in the same group, being tagged in the same photograph, checking in at the same location, or attending the same event; or other suitable action. Although the present disclosure describes measuring affinity in a particular manner, the present disclosure contemplates measuring affinity in any suitable manner.

In particular embodiments, network system 1402 may use various factors to calculate the coefficients. These factors may include, for example, user actions, types of relationships between objects, location information, other suitable factors, or any combination thereof. In particular embodiments, different factors may be weighted differently when calculating the coefficients. The weight of each factor may be static, or the weight may change depending on, for example, the user, the type of relationship, the type of action, the location of the user, and so forth. The ranking of these factors may be combined according to their weights to determine the overall coefficient of the user. By way of example and not by way of limitation, a particular user action may be assigned a level and weight, while a relationship associated with the particular user action is assigned a level and associated weight (e.g., so the weights total 100%). To calculate the coefficient of the user for a particular object, the level of action assigned to the user may comprise, for example, 60% of the total coefficient, while the relationship between the user and the object may comprise 40% of the total coefficient. In particular embodiments, network system 1402 may consider various variables, such as time since information was accessed, attenuation factors, frequency of access, relationship to information or to objects to which information was accessed, relationship to social graph entities connected to objects, short-term or long-term averages of user actions, user feedback, other suitable variables, or any combination thereof, when determining weights for various factors used to calculate coefficients. By way of example and not by way of limitation, the coefficients may include an attenuation factor that attenuates the strength of the signal provided by a particular action over time such that more recent actions are more correlated in calculating the coefficients. The levels and weights may be continuously updated based on continuous tracking of the actions on which the coefficients are based. Any type of process or algorithm may be employed to assign, combine, average, etc. the rank of each factor and the weight assigned to those factors. In particular embodiments, network system 1402 may determine the coefficients using a machine learning algorithm trained from historical actions and past user responses, or data obtained from the user by exposing the user to various options and measuring responses. Although this disclosure describes calculating coefficients in a particular manner, this disclosure contemplates calculating coefficients in any suitable manner.

In particular embodiments, network system 1402 may calculate coefficients based on the user's actions. Network system 1402 can monitor such actions on an online social network, third-party system 1408, other suitable system, or any combination thereof. Any suitable type of user action may be tracked or monitored. Typical user actions include viewing a profile page, creating or publishing content, interacting with content, joining a group, listing and confirming attendance at an activity, checking in at a location, approving a particular page, creating a page, and performing other tasks that facilitate social actions. In particular embodiments, network system 1402 may calculate coefficients based on user actions with particular types of content. The content may be associated with an online social network, a third-party system 1408, or another suitable system. Content may include users, profile pages, posts, news stories, headlines, instant messages, chat room conversations, emails, advertisements, pictures, videos, music, other suitable objects, or any combination thereof. The network system 1402 may analyze the actions of the user to determine whether one or more actions indicate an affinity for the topic, content, other users, and/or the like. By way of example and not by way of limitation, if a user may frequently publish content related to "coffee" or variants thereof, the network system 1402 may determine that the user has a high coefficient with respect to the concept "coffee". A particular action or type of action may be assigned a higher weight and/or level than other actions, which may affect the overall calculated coefficients. By way of example and not by way of limitation, if a first user sends an email to a second user, the weight or level of the action may be higher than if the first user simply viewed the user profile page of the second user.

In particular embodiments, network system 1402 may calculate coefficients based on the type of relationship between particular objects. Referring to the social graph 1500, when calculating coefficients, the network system 1402 may analyze the number and/or types of edges 1506 connecting a particular user node 1502 and concept node 1504. By way of example and not by way of limitation, user nodes 1502 connected by spouse-type edges (indicating that two users have married) may be assigned a higher coefficient than user nodes 1502 connected by friend-type edges. In other words, based on the weights assigned to the actions and relationships of a particular user, it may be determined that the overall affinity for content about the user's spouse is higher than the overall affinity for content about the user's friends. In particular embodiments, a user's relationship to another object may affect the weight and/or level of user action with respect to computing coefficients for that object. By way of example and not by way of limitation, if a user is tagged in a first photo, but only likes a second photo, network system 1402 may determine that the user has a higher coefficient relative to the first photo than the second photo because having a tagged type relationship with content may be assigned a higher weight and/or ranking than having a like type relationship with content. In particular embodiments, network system 1402 may calculate the coefficient for the first user based on the relationship that one or more second users have with a particular object. In other words, the association and coefficients of other users with the object may affect the coefficients of the first user of the object. By way of example and not by way of limitation, if a first user is associated with or has a high coefficient for one or more second users, and those second users are associated with or have a high coefficient for a particular object, then the network system 1402 may determine that the first user should likewise have a relatively high coefficient for the particular object. In particular embodiments, the coefficients may be based on a degree of separation between particular objects. A lower coefficient may represent a reduced likelihood that the first user will share the interest of the content object with users indirectly associated with the first user in the social graph 1500. By way of example and not by way of limitation, social-graph entities that are closer (i.e., less separated) in the social graph 1500 may have a higher coefficient than entities that are further away in the social graph 1500.

In certain embodiments, the network system 1402 may calculate the coefficients based on the location information. Objects that are geographically closer to each other may be considered more relevant or interesting than objects that are further away from each other. In particular embodiments, the coefficient for a user for a particular object may be based on the proximity of the location of the object to the current location associated with the user (or the location of the user's client device). The first user may be more interested in other users or concepts that are closer to the first user. By way of example and not by way of limitation, if a user is one mile from an airport and two miles from a gas station, the network system 1402 can determine that the user has a higher coefficient for the airport than for the gas station based on the proximity of the airport to the user.

In particular embodiments, network system 1402 may perform particular actions with respect to the user based on the coefficient information. The coefficients may be used to predict whether a user will perform a particular action based on the user's interest in the action. The coefficients may be used when generating or presenting any type of object to a user, such as advertisements, search results, news feeds, media, messages, notifications, or other suitable objects. The coefficients may also be used to rank and order such objects appropriately. In this manner, the network system 1402 may provide information related to the user's interests and current environment, increasing the likelihood that they will find such information of interest. In particular embodiments, network system 1402 may generate content based on coefficient information. The content objects may be provided or selected based on user-specific coefficients. By way of example and not by way of limitation, the coefficients may be used to generate media for a user, where the user may be presented with media having a high overall coefficient for the media object. As another example and not by way of limitation, the coefficient may be used to generate advertisements for users, where the users may be presented with advertisements whose overall coefficient with respect to the advertisement object is high. In particular embodiments, network system 1402 may generate search results based on the coefficient information. Search results for a particular user may be scored or ranked based on coefficients associated with the search results for the querying user. By way of example and not by way of limitation, search results corresponding to objects with higher coefficients may be ranked higher on a search results page than results corresponding to objects with lower coefficients.

In particular embodiments, network system 1402 may calculate coefficients in response to a coefficient request from a particular system or process. Any process may request the calculated coefficients from the user in order to predict the actions the user may take (or may be the subject of) in a given situation. The request may also include a set of weights for various factors used to calculate the coefficients. The request may come from a process running on an online social network, from a third-party system 1408 (e.g., via an API or other communication channel), or from another suitable system. In response to the request, the network system 1402 may calculate the coefficients (or access the coefficient information if it has been previously calculated and stored). In particular embodiments, network system 1402 may measure affinity with respect to a particular process. Different processes (both internal and external to the online social network) may request coefficients for a particular object or set of objects. Network system 1402 can provide a measure of affinity related to a particular process that requested the measure of affinity. In this way, each process receives an affinity metric that is customized for a different context, where the process will use the affinity metric.

In conjunction with social graph affinity and affinity coefficients, particular embodiments may utilize one or more systems, components, elements, functions, methods, operations, or steps disclosed in U.S. patent application No. 11/503093, filed 2006, 8, 11, 2010, U.S. patent application No. 12/977027, filed 2010, 12, 22, U.S. patent application No. 12/978265, filed 2010, 12, 23, and U.S. patent application No. 13/632869, filed 2012, 10, 01, each of which is incorporated by reference.

In particular embodiments, one or more content objects of an online social network may be associated with a privacy setting. The privacy settings (or "access settings") of the object may be stored in any suitable manner, such as, for example, associated with the object, stored in an index on an authorization server, stored in another suitable manner, or any combination thereof. The privacy settings of the object may specify how the object (or particular information associated with the object) may be accessed (e.g., viewed or shared) using the online social network. Where the privacy settings of an object allow a particular user to access the object, the object may be described as "visible" to the user. By way of example and not by way of limitation, a user of an online social network may specify a privacy setting for a user profile page that identifies a group of users that may access work experience information on the user profile page, thereby excluding other users from accessing the information. In particular embodiments, the privacy settings may specify a "blacklist" of users that should not be allowed to access certain information associated with the object. In other words, the blacklist may specify one or more users or entities for which the object is not visible. By way of example and not by way of limitation, a user may specify a group of users who may not have access to an album associated with the user, thereby precluding those users from accessing the album (while some users who may not be in the group of users may also be allowed access to the album). In particular embodiments, privacy settings may be associated with particular social graph elements. Privacy settings of a social graph element (such as a node or edge) may specify how the social graph element, information associated with the social graph element, or content objects associated with the social graph element may be accessed using an online social network. By way of example and not by way of limitation, a particular concept node 1504 corresponding to a particular photo may have a privacy setting that specifies that the photo is only accessible by the user and his friends who are tagged in the photo. In particular embodiments, privacy settings may allow users to opt-in or opt-out to have their actions logged in through network system 1402 or shared with other systems (e.g., third party system 1408). In particular embodiments, the privacy settings associated with the object may specify any suitable granularity at which access is allowed or denied. As an example and not by way of limitation, access or denial of access may be specified for particular users (e.g., only me, my roommates, and my boss), users within a particular degree of separation (e.g., friends or friends of friends), groups of users (e.g., gaming clubs, my family), networks of users (e.g., employees of a particular employer, students of a particular university, or alumni), all users ("public"), no users ("private"), users of third-party system 1408, particular applications (e.g., third-party applications, external websites), other suitable users or entities, or any combination thereof. Although this disclosure describes using particular privacy settings in a particular manner, this disclosure contemplates using any suitable privacy settings in any suitable manner.

In particular embodiments, one or more of the servers may be an authorization/privacy server for enforcing privacy settings. In response to a request from a user (or other entity) for a particular object stored in the data store, the network system 1402 can send a request for the object to the data store. The request may identify a user associated with the request and may be sent to the user (or the user's client device 1406) only if the authorization server determines that the user is authorized to access the object based on the privacy settings associated with the object. If the requesting user is not authorized to access the object, the authorization server may prevent the requested object from being retrieved from the data store, or may prevent the requested object from being sent to the user. In the context of a search query, an object may be generated as a search result only if the querying user is authorized to access the object. In other words, the object must have visibility that is visible to the querying user. An object may be excluded from search results if the object has visibility that is not visible to the user. Although this disclosure describes implementing privacy settings in a particular manner, this disclosure contemplates implementing privacy settings in any suitable manner.

The foregoing description has been described with reference to specific exemplary embodiments thereof. Various embodiments and aspects of the disclosure are described with reference to details discussed herein, and the accompanying drawings illustrate the various embodiments. The foregoing description and drawings are illustrative and are not to be construed as limiting. Numerous specific details are described to provide a thorough understanding of various embodiments.

Additional or alternative embodiments may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. The described embodiments are to be considered in all respects only as illustrative and not restrictive. The scope of the invention is, therefore, indicated by the appended claims rather than by the foregoing description. All changes which come within the meaning and range of equivalency of the claims are to be embraced within their scope.