阅读说明：本技术 一种文章大纲生成方法、装置和用于生成文章大纲的装置 (Article outline generation method and device for generating article outline ) 是由 龚能 孙嘉楠 徐竟达 涂曼姝 于 2021-08-06 设计创作，主要内容包括：本发明实施例提供了一种文章大纲生成方法、装置和用于生成文章大纲的装置。所述方法包括：获取符合预设主题的至少两篇参考文章并确定每篇参考文章的段落；确定每篇参考文章的每个段落对应的候选标题；对所述至少两篇参考文章的每个段落对应的候选标题进行聚类处理,得到所述预设主题对应的全局标题连接图；根据所述全局标题连接图生成所述预设主题对应的文章大纲。本发明生成的文章大纲,能够充分涵盖符合预设主题的多篇文章中与预设主题相关的信息,提高了生成的文章大纲与预设主题的相关度；并且,本发明通过基于多篇参考文章的段落标题生成文章大纲,增强了生成的文章大纲的可信度,有利于提升写作辅助效果。(The embodiment of the invention provides an article outline generating method and device and a device for generating an article outline. The method comprises the following steps: acquiring at least two reference articles conforming to a preset theme and determining paragraphs of each reference article; determining a candidate title corresponding to each paragraph of each reference article; clustering candidate titles corresponding to each paragraph of the at least two reference articles to obtain a global title connection graph corresponding to the preset topic; and generating an article outline corresponding to the preset theme according to the global title connection diagram. The article outline generated by the invention can fully cover the information related to the preset theme in a plurality of articles conforming to the preset theme, and the relevance of the generated article outline and the preset theme is improved; in addition, the article outline is generated based on the paragraph titles of the multiple reference articles, so that the reliability of the generated article outline is enhanced, and the writing auxiliary effect is favorably improved.)

1. A method for generating an article outline, the method comprising:

acquiring at least two reference articles conforming to a preset theme and determining paragraphs of each reference article;

determining a candidate title corresponding to each paragraph of each reference article;

clustering candidate titles corresponding to each paragraph of the at least two reference articles to obtain a global title connection graph corresponding to the preset topic;

and generating an article outline corresponding to the preset theme according to the global title connection diagram.

2. The method of claim 1, wherein the clustering the candidate headlines corresponding to each paragraph of the at least two reference articles to obtain the global headline connection graph corresponding to the preset topic comprises:

merging each candidate title corresponding to each paragraph to obtain a target title corresponding to each paragraph;

sequentially connecting target headings corresponding to paragraphs belonging to the same reference article according to the sequence of the paragraphs to obtain at least two internal heading connection graphs, wherein each internal heading connection graph corresponds to one reference article;

and clustering each target title in the at least two internal title connection graphs according to the semantics of the target title to obtain a global title connection graph corresponding to the preset theme.

3. The method according to claim 2, wherein the clustering each target title in the at least two internal title connection graphs according to the semantics of the target title to obtain a global title connection graph corresponding to the preset topic comprises:

clustering all target titles belonging to different reference articles according to the semantics of the target titles to obtain a cluster to which each target title belongs;

and connecting all target titles belonging to the same cluster in the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset theme.

4. The method according to claim 3, wherein the connecting each target title belonging to the same cluster in the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset topic comprises:

judging whether at least two identical target titles exist in the same cluster of the at least two internal title connection graphs or not;

if at least two identical target titles exist in the same cluster, merging the at least two identical target titles to obtain a merged target title, and updating the target title in the cluster to which the merged target title belongs according to the merged target title;

and connecting the updated target titles in the same cluster of the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset theme.

5. The method of claim 1, wherein the generating of the article outline corresponding to the preset topic according to the global title connection diagram comprises:

extracting a target subgraph in the global title connection graph based on a community discovery algorithm;

and generating an article outline corresponding to the preset theme according to the target subgraph.

6. The method of claim 1, wherein the candidate headings comprise a first candidate heading and a second candidate heading, and wherein determining the candidate heading for each paragraph of each reference article comprises:

generating at least two paragraph titles for each paragraph of each reference article;

determining a first candidate title from the paragraph titles of each paragraph, wherein the similarity between the first candidate title and other paragraph titles of the paragraph to which the first candidate title belongs meets a first preset condition;

and determining a second candidate title of which the similarity with the first candidate title meets a second preset condition.

7. The method of claim 6, wherein determining the first candidate title from the respective paragraph titles of each paragraph comprises:

calculating the sum of the similarity of a first paragraph title of the same paragraph and other paragraph titles to obtain a correlation index of the first paragraph title, wherein the first paragraph title is any paragraph title of at least two paragraph titles of the paragraph to which the first paragraph title belongs;

and if the correlation index of the first paragraph title is larger than the correlation indexes of other paragraph titles of the paragraph to which the first paragraph title belongs, determining that the first paragraph title is a first candidate title.

8. The method according to claim 6, wherein the second preset condition comprises: the similarity between the second candidate title and the first candidate title is smaller than the similarity between a second paragraph title and the first candidate title, and the second paragraph title is any paragraph title which does not belong to the candidate title in at least two paragraph titles of the paragraph to which the first candidate title belongs.

9. The method of claim 1, wherein obtaining at least two reference articles that meet a predetermined topic and determining a passage of each reference article comprises:

receiving a preset theme and a screening condition input by a user, wherein the screening condition comprises at least one of article types, quoted times, publication time, article reading amount and article praise amount;

searching at least two reference articles meeting the preset theme according to the screening condition;

the paragraphs of each reference article are determined based on a pre-trained segmentation model.

10. The method of claim 9, wherein prior to determining the passage of each reference article based on the pre-trained segmentation model, the method further comprises:

constructing a segmentation model based on a bidirectional self-coding model and a conditional random field;

acquiring a segmented training corpus, wherein the segmented training corpus comprises a sample article and segmentation points corresponding to the sample article;

and training the segmented model according to the segmented training corpus until the segmented model meets a convergence condition to obtain the trained segmented model.

11. An article outline generation apparatus, characterized in that the apparatus comprises:

the reference article acquisition module is used for acquiring at least two reference articles conforming to a preset theme and determining paragraphs of each reference article;

the candidate title determining module is used for determining a candidate title corresponding to each paragraph of each reference article;

a headline join graph generation module, configured to perform clustering processing on candidate headlines corresponding to each paragraph of the at least two reference articles to obtain a global headline join graph corresponding to the preset topic;

and the article outline generating module is used for generating the article outline corresponding to the preset theme according to the global title connection diagram.

12. The apparatus of claim 11, wherein the header connection map generation module comprises:

the title merging submodule is used for merging all candidate titles corresponding to each paragraph to obtain a target title corresponding to each paragraph;

the internal headline connection graph generation sub-module is used for sequentially connecting target headlines corresponding to paragraphs belonging to the same reference article according to the sequence of the paragraphs to obtain at least two internal headline connection graphs, and each internal headline connection graph corresponds to one reference article;

and the global title connection graph generation sub-module is used for clustering each target title in the at least two internal title connection graphs according to the semantics of the target title to obtain a global title connection graph corresponding to the preset theme.

13. The apparatus of claim 12, wherein the global title connection map generation sub-module comprises:

the clustering unit is used for clustering all target titles belonging to different reference articles according to the semantics of the target titles to obtain a cluster to which each target title belongs;

and the global title connection graph generating unit is used for connecting all target titles belonging to the same cluster in the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset theme.

14. An apparatus for generating an article outline, the apparatus comprising a memory, and one or more programs, wherein the one or more programs are stored in the memory, and wherein the one or more programs configured to be executed by the one or more processors comprise instructions for performing the article outline generation method of any one of claims 1 to 10.

15. A machine-readable medium having stored thereon instructions, which when executed by one or more processors, cause an apparatus to perform the article schema generation method of any one of claims 1-10.

Technical Field

The invention relates to the technical field of computers, in particular to a method and a device for generating an article outline and a device for generating the article outline.

Background

With the rapid development of the mobile internet technology, massive information is gathered in various information flow platforms, and massive articles provide nearly endless choices for readers, but for producers of the articles, writing and publishing thresholds are reduced, but difficulties of looking away from massive articles are increased.

At present, a segmentation model is generally trained on an outline text based on Wikipedia and a paper, a first-level segmentation and a paragraph title of an article are determined according to the trained segmentation model, and an article outline is generated according to the paragraph title and recommended to a user to assist the user in writing.

Disclosure of Invention

The embodiment of the invention provides an article outline generating method and device and an article outline generating device, which can improve the relevance of the generated article outline and a preset theme and enhance the reliability of the article outline.

In order to solve the above problem, an embodiment of the present invention discloses a method for generating an article outline, where the method includes:

acquiring at least two reference articles conforming to a preset theme and determining paragraphs of each reference article;

determining a candidate title corresponding to each paragraph of each reference article;

clustering candidate titles corresponding to each paragraph of the at least two reference articles to obtain a global title connection graph corresponding to the preset topic;

and generating an article outline corresponding to the preset theme according to the global title connection diagram.

Optionally, the clustering the candidate titles corresponding to each paragraph of the at least two reference articles to obtain a global title connection graph corresponding to the preset topic includes:

merging each candidate title corresponding to each paragraph to obtain a target title corresponding to each paragraph;

sequentially connecting target headings corresponding to paragraphs belonging to the same reference article according to the sequence of the paragraphs to obtain at least two internal heading connection graphs, wherein each internal heading connection graph corresponds to one reference article;

and clustering each target title in the at least two internal title connection graphs according to the semantics of the target title to obtain a global title connection graph corresponding to the preset theme.

Optionally, the clustering, according to the semantics of the target title, each target title in the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset topic includes:

clustering all target titles belonging to different reference articles according to the semantics of the target titles to obtain a cluster to which each target title belongs;

and connecting all target titles belonging to the same cluster in the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset theme.

Optionally, the connecting each target title belonging to the same cluster in the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset topic includes:

judging whether at least two identical target titles exist in the same cluster of the at least two internal title connection graphs or not;

if at least two identical target titles exist in the same cluster, merging the at least two identical target titles to obtain a merged target title, and updating the target title in the cluster to which the merged target title belongs according to the merged target title;

and connecting the updated target titles in the same cluster of the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset theme.

Optionally, the generating the article outline corresponding to the preset topic according to the global title connection diagram includes:

extracting a target subgraph in the global title connection graph based on a community discovery algorithm;

and generating an article outline corresponding to the preset theme according to the target subgraph.

Optionally, the determining the candidate headline corresponding to each paragraph of each reference article includes:

generating at least two paragraph titles for each paragraph of each reference article;

determining a first candidate title from the paragraph titles of each paragraph, wherein the similarity between the first candidate title and other paragraph titles of the paragraph to which the first candidate title belongs meets a first preset condition;

and determining a second candidate title of which the similarity with the first candidate title meets a second preset condition.

Optionally, the determining a first candidate title from the respective paragraph titles of each paragraph includes:

calculating the sum of the similarity of a first paragraph title of the same paragraph and other paragraph titles to obtain a correlation index of the first paragraph title, wherein the first paragraph title is any paragraph title of at least two paragraph titles of the paragraph to which the first paragraph title belongs;

and if the correlation index of the first paragraph title is larger than the correlation indexes of other paragraph titles of the paragraph to which the first paragraph title belongs, determining that the first paragraph title is a first candidate title.

Optionally, the second preset condition includes: the similarity between the second candidate title and the first candidate title is smaller than the similarity between a second paragraph title and the first candidate title, and the second paragraph title is any paragraph title which does not belong to the candidate title in at least two paragraph titles of the paragraph to which the first candidate title belongs.

Optionally, the obtaining at least two reference articles meeting a preset topic and determining a paragraph of each reference article includes:

receiving a preset theme and a screening condition input by a user, wherein the screening condition comprises at least one of article types, quoted times, publication time, article reading amount and article praise amount;

searching at least two reference articles meeting the preset theme according to the screening condition;

the paragraphs of each reference article are determined based on a pre-trained segmentation model.

Optionally, before determining the paragraphs of each reference article based on the pre-trained segmentation model, the method further comprises:

constructing a segmentation model based on a bidirectional self-coding model and a conditional random field;

acquiring a segmented training corpus, wherein the segmented training corpus comprises a sample article and segmentation points corresponding to the sample article;

and training the segmented model according to the segmented training corpus until the segmented model meets a convergence condition to obtain the trained segmented model.

On the other hand, the embodiment of the invention discloses an article outline generating device, which comprises:

the reference article acquisition module is used for acquiring at least two reference articles conforming to a preset theme and determining paragraphs of each reference article;

the candidate title determining module is used for determining a candidate title corresponding to each paragraph of each reference article;

a headline join graph generation module, configured to perform clustering processing on candidate headlines corresponding to each paragraph of the at least two reference articles to obtain a global headline join graph corresponding to the preset topic;

and the article outline generating module is used for generating the article outline corresponding to the preset theme according to the global title connection diagram.

Optionally, the title connection graph generating module includes:

the title merging submodule is used for merging all candidate titles corresponding to each paragraph to obtain a target title corresponding to each paragraph;

the internal headline connection graph generation sub-module is used for sequentially connecting target headlines corresponding to paragraphs belonging to the same reference article according to the sequence of the paragraphs to obtain at least two internal headline connection graphs, and each internal headline connection graph corresponds to one reference article;

and the global title connection graph generation sub-module is used for clustering each target title in the at least two internal title connection graphs according to the semantics of the target title to obtain a global title connection graph corresponding to the preset theme.

Optionally, the global title connection graph generating sub-module includes:

the clustering unit is used for clustering all target titles belonging to different reference articles according to the semantics of the target titles to obtain a cluster to which each target title belongs;

and the global title connection graph generating unit is used for connecting all target titles belonging to the same cluster in the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset theme.

Optionally, the global title connection map generating unit includes:

a target title judging subunit, configured to judge whether at least two identical target titles exist in the same cluster of the at least two internal title connection graphs;

the target title merging subunit is configured to merge at least two identical target titles to obtain a merged target title if the at least two identical target titles exist in the same class cluster, and update the target title in the class cluster to which the merged target title belongs according to the merged target title;

and the target title connection subunit is used for connecting the updated target titles in the same cluster of the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset theme.

Optionally, the article outline generating module includes:

the target subgraph extraction submodule is used for extracting the target subgraph in the global title connection graph based on a community discovery algorithm;

and the article outline generation submodule is used for generating the article outline corresponding to the preset theme according to the target subgraph.

Optionally, the candidate titles include a first candidate title and a second candidate title, and the candidate title determination module includes:

the paragraph title generation submodule is used for generating at least two paragraph titles aiming at each paragraph of each reference article;

the first candidate title determining sub-module is used for determining a first candidate title from the paragraph titles of each paragraph, and the similarity between the first candidate title and other paragraph titles of the paragraph to which the first candidate title belongs meets a first preset condition;

and the second candidate title determining submodule is used for determining a second candidate title of which the similarity with the first candidate title meets a second preset condition.

Optionally, the first candidate title determining sub-module includes:

the system comprises a correlation index calculation unit, a correlation index calculation unit and a display unit, wherein the correlation index calculation unit is used for calculating the sum of the similarity of a first paragraph title of the same paragraph and other paragraph titles to obtain the correlation index of the first paragraph title, and the first paragraph title is any paragraph title of at least two paragraph titles of the paragraph to which the first paragraph title belongs;

a first candidate title determining unit, configured to determine that the first paragraph title is a first candidate title if the correlation index of the first paragraph title is greater than the correlation indexes of the other paragraph titles of the paragraph to which the first paragraph title belongs.

Optionally, the second preset condition includes: the similarity between the second candidate title and the first candidate title is smaller than the similarity between a second paragraph title and the first candidate title, and the second paragraph title is any paragraph title which does not belong to the candidate title in at least two paragraph titles of the paragraph to which the first candidate title belongs.

Optionally, the reference article obtaining module includes:

the data receiving submodule is used for receiving a preset theme and a screening condition input by a user, wherein the screening condition comprises at least one item of article types, quoted times, publication time, article reading quantity and article praise quantity;

the article screening submodule is used for searching at least two reference articles meeting the preset theme according to the screening condition;

and the paragraph determining sub-module is used for determining the paragraph of each reference article based on the pre-trained segmentation model.

Optionally, the reference article obtaining module further includes:

the segmentation module construction submodule is used for constructing a segmentation model based on the bidirectional self-coding model and the conditional random field;

the corpus acquiring submodule is used for acquiring a segmented corpus, and the segmented corpus comprises a sample article and segmentation points corresponding to the sample article;

and the segmental model training submodule is used for training the segmental model according to the segmental training corpus until the segmental model meets a convergence condition to obtain a trained segmental model.

In yet another aspect, an embodiment of the present invention discloses an apparatus for generating an article outline, the apparatus comprising a memory, and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs configured to be executed by the one or more processors comprise instructions for performing one or more of the article outline generation methods described above.

In yet another aspect, an embodiment of the present invention discloses a machine-readable medium having stored thereon instructions, which when executed by one or more processors, cause an apparatus to perform a method for generating an article outline as described in one or more of the preceding.

The embodiment of the invention has the following advantages:

the method comprises the steps of obtaining at least two reference articles according with a preset theme, determining paragraphs of each reference article, and further determining candidate titles corresponding to each paragraph of each reference article; clustering candidate titles corresponding to each paragraph of the at least two reference articles to obtain a global title connection graph corresponding to the preset topic; and finally, generating an article outline corresponding to the preset theme according to the global title connection diagram. Compared with the prior art that one article is screened out from a plurality of articles and an article outline is generated based on the screened one article, the article outline generated by the invention can fully cover information related to the preset theme in the plurality of articles conforming to the preset theme, and the correlation degree of the generated article outline and the preset theme is improved; in addition, the article outline is generated based on the paragraph titles of the multiple reference articles, so that the reliability of the generated article outline is enhanced, and the writing auxiliary effect is favorably improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments of the present invention will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without inventive labor.

FIG. 1 is a flow chart of steps of an embodiment of a method for generating an outline of an article of the present invention;

FIG. 2 is a block diagram showing the structure of an embodiment of an article outline generating apparatus according to the present invention;

FIG. 3 is a block diagram of an apparatus 800 for generating an article outline according to the present invention;

fig. 4 is a schematic diagram of a server in some embodiments of the invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Method embodiment

Referring to fig. 1, a flowchart illustrating steps of an embodiment of a speech processing method according to the present invention is shown, where the method specifically includes the following steps:

step 101, obtaining at least two reference articles conforming to a preset theme and determining paragraphs of each reference article.

And 102, determining candidate titles corresponding to each paragraph of each reference article.

Step 103, clustering the candidate titles corresponding to each paragraph of the at least two reference articles to obtain a global title connection graph corresponding to the preset topic.

And 104, generating an article outline corresponding to the preset theme according to the global title connection diagram.

The article outline generation method provided by the embodiment of the invention can be applied to terminal equipment and can also be applied to a server. Wherein, the terminal device may include but is not limited to: smart terminals, computers, Personal Digital Assistants (PDAs), tablet computers, e-book readers, laptop portable computers, in-vehicle devices, smart televisions, wearable devices, and the like.

The server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as cloud service, a cloud database, cloud computing, a cloud function, cloud storage, cloud communication, Network service, middleware service, Content Delivery Network (CDN), big data and artificial intelligence platform, and the like.

It should be noted that the preset theme may be a theme obtained by analyzing, by the terminal device and/or the server, the writing requirement submitted by the user. The writing requirement submitted by the user can be character information, and the terminal equipment and/or the server determine a preset theme by performing semantic analysis on the character information; or, the writing requirement submitted by the user may also be voice information, and the terminal device and/or the server may convert the voice information into text information through a built-in voice recognition model, and further perform semantic analysis on the converted text information to obtain the preset theme; or, the writing information submitted by the user may also be a picture, and the terminal device and/or the server determines the content contained in the picture by performing feature analysis and character recognition on the picture, and then determines a preset theme according to the content contained in the picture, and so on.

Wherein the reference article may be any type of article that conforms to the preset topic. For example, the reference article may be a paper, a newsletter, a public article, a forum post, a blog, a web document, etc. that conforms to the preset topic. The article type of the reference article is not particularly limited in the embodiment of the present invention, as long as the reference article conforms to the preset topic.

When the reference article conforming to the preset theme is obtained, the article material database can be constructed in advance, and the reference article conforming to the preset theme can be searched from the article material database. The article material database can utilize data services provided by a third-party system, such as various thesis databases, literature databases, science and technology databases, and the like. Specifically, free data can be collected from the existing media platform (including an article retrieval platform, a mainstream media platform, a microblog, a forum and the like) through a web crawler technology, or authorized data can be collected in cooperation with a third party, or a third party system is directly utilized, and technical personnel can upload article materials periodically and continuously update the article material database.

For each article in the article material database, the article content can be analyzed first, the article theme is determined, and a corresponding theme tag is generated, so that when a reference article is obtained, the reference article can be determined directly based on the similarity between the theme tag of each article and the preset theme.

After the reference articles meeting the preset theme are determined, the paragraphs of each article are further determined. In particular, the paragraphs of each article may be determined based on a pre-trained segmentation model. The segmentation model may be any neural network model that is good at segmenting in the field, and the embodiment of the present invention is not particularly limited.

Then, candidate headings corresponding to each paragraph of each reference article are determined. Specifically, the content of the paragraph may be analyzed, key entity words (such as name entities like place names and names of people, or other keywords) and subject words (such as sports, entertainment, emotion, etc.) in the paragraph may be extracted, then, a paragraph title may be generated according to the extracted key entity words and subject words, and a preset number of candidate titles may be selected from the generated multiple titles. Among them, candidate titles can be selected from the generated multiple titles based on MMR (maximum boundary correlation algorithm). Of course, other methods may be used to select candidate titles.

After the candidate titles corresponding to the paragraphs of each reference article are determined, clustering the candidate titles corresponding to the paragraphs of all the reference articles to obtain a global title connection graph corresponding to the preset topic. The global title connection graph is formed by candidate titles corresponding to all paragraphs of the reference articles, each node of the global title connection graph is generated according to the candidate titles, each candidate title corresponds to one node, and each corresponding node is connected according to the incidence relation among the candidate titles, so that the global title connection graph can be obtained.

And finally, generating an article outline corresponding to the preset theme according to the global title connection diagram. Specifically, extracting important subgraphs, namely target subgraphs, in the global title connection graph, and generating an article outline according to titles corresponding to nodes in the target subgraphs.

The target subgraph comprises paragraph titles with the highest relevance among the multiple reference articles and accords with the writing thought of the multiple reference articles, so that the article outline generated based on the target subgraph can simultaneously cover the content of the multiple reference articles, and the relevance between the generated article outline and the preset theme is improved; in addition, the article outline is generated based on the paragraph titles of the multiple reference articles, so that the reliability of the generated article outline is enhanced, and the writing auxiliary effect is favorably improved.

In an optional embodiment of the present invention, the acquiring at least two reference articles meeting a preset topic and determining a paragraph of each reference article in step 101 includes:

step S11, receiving a preset theme and a screening condition input by a user, wherein the screening condition comprises at least one of article type, quoted times, publication time, article reading amount and article praise amount;

step S12, searching at least two reference articles meeting the preset theme according to the screening condition;

step S13, determining paragraphs of each reference article based on the pre-trained segmentation model.

After the user inputs the drawn preset theme, the user can further submit the screening conditions to the terminal device and/or the server according to the actual requirements and personal preferences, and further explicitly refer to the requirements which the article needs to meet. The screening condition may include at least one of article type, number of times of being referred to, publication time, article reading amount, and article approval amount. The article types may include publication types of articles, such as articles (academic articles, journal articles, etc.), journal articles, public articles, forum posts, news reports, etc.; the article type may also be a written type of a newspaper article, such as a narrative, lyrics, treatise, narrative, application, and the like.

In the embodiment of the invention, articles in the article material database can be analyzed in advance, and the corresponding topic tags, type tags, time tags, heat tags and the like of each article are determined, so that reference articles meeting preset topics and screening conditions can be conveniently searched based on the tags.

After the reference articles meeting the preset subjects and the screening conditions are found, the reference articles need to be further segmented to determine paragraphs of the reference articles. In particular, the paragraphs of each article may be determined based on a pre-trained segmentation model. The segmentation model may be any neural network model that is good at segmenting in the field, and the embodiment of the present invention is not particularly limited.

In order to be suitable for various types of articles and improve the accuracy of the article segmentation result, a sequence labeling method can be adopted, a segmentation model is constructed based on BERT (Bidirectional Encoder reconstruction from transformations) and CRF (Conditional Random Field), the constructed segmentation model is trained, and the reference article is segmented based on the trained segmentation model.

In an alternative embodiment of the present invention, before determining the paragraphs of each reference article based on the pre-trained segmentation model in step S13, the method further comprises:

the substep S131 is that a segmented model is constructed based on the bidirectional self-coding model and the conditional random field;

substep S132, obtaining a segmented corpus, wherein the segmented corpus comprises a sample article and a segmentation point corresponding to the sample article;

and S133, training the segmented model according to the segmented training corpus until the segmented model meets a convergence condition to obtain a trained segmented model.

Wherein BERT uses an occlusion language model to achieve a pre-trained deep bi-directional representation. The pre-trained BERT representation can be used for model fine adjustment through an additional output layer, is suitable for building a neural network model of a wide range of tasks, and does not need to greatly modify a network architecture aiming at specific tasks. Conditional Random Fields (CRF) are a sequence labeling algorithm and can also be considered as discriminant models that predict an input sequence based on the input sequence. The segmentation model constructed according to BERT and CRF can be trained through a small amount of labeled data, can be quickly migrated to a new application scene, and can be suitable for segmentation processing of various articles.

Of course, in the embodiment of the present invention, other neural network models that are good at segmenting articles may also be selected, and the embodiment of the present invention is not particularly limited thereto.

In an optional embodiment of the present invention, the step 102 of determining the candidate headlines includes a first candidate headline and a second candidate headline, and the step includes:

step S21, generating at least two paragraph titles aiming at each paragraph of each reference article;

step S22, determining a first candidate title from the paragraph titles of each paragraph, wherein the similarity between the first candidate title and other paragraph titles of the paragraph to which the first candidate title belongs satisfies a first preset condition;

and step S23, determining a second candidate title whose similarity to the first candidate title satisfies a second preset condition.

When the paragraph titles are generated, if the paragraph titles are generated only according to the content of the paragraphs themselves, the obtained paragraph titles may have a deviation from the actual content of the paragraphs, and therefore, in the embodiment of the present invention, the full-text keyword of the reference article may be determined by referring to the full-text content of the article, and then the paragraph titles are generated based on the full-text keyword and the content of the paragraphs themselves, so as to improve the accuracy of the paragraph titles.

There are often multiple keywords in the full text of the reference article, and there may be multiple title generation models for generating the section titles, so that multiple section titles can be obtained for the same section. Some common points or similar points often exist among the obtained paragraph titles, so that in order to avoid data redundancy and improve processing efficiency, the obtained paragraph titles can be screened, part of the obtained paragraph titles are selected as candidate titles, and an article outline is generated according to the selected candidate titles. The number of the selected candidate titles may be set according to actual requirements, for example, 3 titles are selected as the candidate titles, and the number of the candidate titles is not specifically limited in the embodiment of the present invention.

In order to ensure the reliability of the finally generated article outline, the selected candidate titles are required to represent the paragraph contents, and therefore, in the embodiment of the invention, the candidate titles are determined based on the similarity between the paragraph titles. The first candidate title is a first selected candidate title, and after the first candidate title is determined, other candidate titles, namely a second candidate title, are further determined based on the similarity between the paragraph title and the first candidate title.

Specifically, candidate titles may be extracted based on MMR (maximum boundary correlation algorithm). Then, the first preset condition is that the sum of the similarity of the first candidate title and the similarity of other section titles is the highest, and the second preset condition is that the similarity of the second candidate title and the first candidate title is the lowest. Therefore, the diversity of the candidate titles can be improved while the selected candidate titles can represent the paragraph contents, so that data redundancy is avoided, more information is covered as far as possible, and the reliability of the finally generated article outline is improved.

Of course, the candidate titles may also be selected in other manners, and the first preset condition and the second preset condition may be determined according to a specific candidate title selection manner, which is not specifically limited in this embodiment of the present invention.

Taking the example of selecting candidate titles based on the MMR algorithm, step S22 in an alternative embodiment of the present invention, the determining the first candidate title from the paragraph titles of each paragraph includes:

step S221, calculating the sum of the similarity of a first paragraph title of the same paragraph and other paragraph titles to obtain a correlation index of the first paragraph title, wherein the first paragraph title is any paragraph title of at least two paragraph titles of the paragraph to which the first paragraph title belongs;

and a substep S222, if the correlation index of the first paragraph title is greater than the correlation indexes of the other paragraph titles of the paragraph to which the first paragraph title belongs, determining that the first paragraph title is a first candidate title.

When the first candidate title is determined, the sum of the similarity of each paragraph title of the same paragraph and other paragraph titles is calculated respectively to obtain the correlation index of each paragraph title. Then, the correlation indexes of the paragraph titles are compared, and the paragraph title with the largest correlation index is determined to be the first candidate title. Therefore, the first candidate title which is selected and the generated paragraph titles which have the highest correlation can be determined, namely, the first candidate title contains the common points among the paragraph titles, so that the paragraph contents can be represented to the maximum extent.

In an optional embodiment of the invention, the second preset condition comprises: the similarity between the second candidate title and the first candidate title is smaller than the similarity between a second paragraph title and the first candidate title, and the second paragraph title is any paragraph title which does not belong to the candidate title in at least two paragraph titles of the paragraph to which the first candidate title belongs.

After the first candidate title capable of characterizing the paragraph contents is selected, in order to further perform information supplementation on the first candidate title and improve the accuracy of the selected candidate title, the similarity between each paragraph title and the first candidate title may be continuously calculated, and the paragraph title with the lowest similarity of the first candidate title is selected as the second candidate title, thereby improving the diversity of the candidate titles.

In an optional embodiment of the present invention, the clustering, in step 103, the candidate headings corresponding to each paragraph of the at least two reference articles to obtain a global heading connection graph corresponding to the preset topic includes:

step S31, merging each candidate title corresponding to each paragraph to obtain a target title corresponding to each paragraph;

step S32, sequentially connecting target headings corresponding to paragraphs belonging to the same reference article according to the sequence of the paragraphs to obtain at least two internal heading connection graphs, wherein each internal heading connection graph corresponds to one reference article;

and step S33, clustering each target title in the at least two internal title connection graphs according to the semantics of the target title to obtain a global title connection graph corresponding to the preset theme.

When clustering is performed on candidate titles of each paragraph of each reference article, in order to reduce data processing amount and improve data processing efficiency, merging processing may be performed on each candidate title corresponding to each paragraph, and multiple candidate titles are merged into one target title, where the merged target title includes information of each candidate title of the paragraph. Specifically, the common features of the candidate titles may be extracted, and the extracted common features and the distinguishing features of the candidate titles may be combined to generate the target title according with the syntax semantics.

And then, sequentially connecting the target headings belonging to the same reference article according to the sequence of the corresponding paragraphs in the reference article to obtain internal heading connection graphs corresponding to the reference articles, wherein each reference article corresponds to one internal heading connection graph. Specifically, while generating the target titles, the paragraph position information corresponding to each target title and the reference article information to which the corresponding paragraph position information belongs may be recorded, so that the target titles corresponding to the paragraphs belonging to the same reference article are sequentially connected according to the sequence between the paragraphs based on the paragraph position information corresponding to each target title and the reference article information to which the corresponding paragraph position information belongs.

It should be noted that the internal headline connection diagram reflects the internal connection relationship between the target headlines of the reference article. The global title connection graph not only contains internal connection relations among all target titles belonging to the same reference article, but also contains external connection relations among all target titles belonging to different reference articles. Therefore, after the internal headline connection graph is generated, all the target headlines in each generated internal headline connection graph need to be further clustered, and the external connection relationship between the target headlines is determined according to the cluster to which each target headline of different reference articles belongs.

In an optional embodiment of the present invention, the step S33 of clustering each target title in the at least two internal title connection graphs according to the semantics of the target title to obtain a global title connection graph corresponding to the preset topic includes:

substep S331, clustering each target title belonging to different reference articles according to the semantics of the target title to obtain a cluster to which each target title belongs;

and a substep S332 of connecting the target titles belonging to the same cluster in the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset theme.

In the embodiment of the invention, the target titles belonging to different reference articles are clustered according to the semantics of the target titles, the class cluster to which each target title belongs is determined, and then the target titles belonging to the same class cluster in each internal title connection diagram are connected to obtain the global title connection diagram. The length of the external connecting line between the target titles included in the same cluster in the global title connection graph may be determined according to the similarity between the target titles in the cluster, for example, the higher the similarity between two target titles is, the shorter the external connecting line between them is, the more compact the representation relationship is, and the like.

In an optional embodiment of the present invention, in sub-step S332, the connecting each target title belonging to the same cluster in the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset topic includes:

step P11, judging whether there are at least two same target titles in the same cluster of the at least two internal title connection graphs;

step P12, if at least two identical target titles exist in the same cluster, merging the at least two identical target titles to obtain a merged target title, and updating the target title in the cluster to which the merged target title belongs according to the merged target title;

and P13, connecting the updated target titles in the same cluster of the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset theme.

It should be noted that, in practical applications, there are cases where target titles of different reference articles are the same, for example, for a reference article a, target titles a1, a2, a3 are corresponding to the reference article a, and for a reference article B, target titles B1, B2, B3, B4 are corresponding to the reference article B, where the target title a2 of the reference article a and the target title B3 of the reference article belong to the same cluster, and the target title a2 and the target title B3 are the same, in this embodiment of the present invention, in order to simplify the global title connection diagram and avoid data redundancy, the same target title a2 and the target title B3 may be merged to generate a new target title, for example, the newly generated target title may be denoted as a2B3, and then the target titles of the same cluster are connected to obtain the global title connection diagram.

In an optional embodiment of the present invention, the generating, in step 104, the article outline corresponding to the preset topic according to the global title connection diagram includes:

step S41, extracting a target subgraph in the global title connection graph based on a community discovery algorithm;

and step S42, generating an article outline corresponding to the preset theme according to the target subgraph.

The target subgraph comprises paragraph titles with the highest relevance among the multiple reference articles, and accords with the writing thought of the multiple reference articles, so that the article outline generated based on the target subgraph can simultaneously cover the content of the multiple reference articles, and the relevance between the generated article outline and the preset theme is improved. Target subgraphs in the global title connection graph can be extracted based on a community discovery algorithm. Specifically, by measuring the similarity between the target titles inside the class clusters and the similarity outside the class clusters in the sub-graphs included in the global title connection graph, the sub-graphs with high target title similarity inside the class clusters and low target title similarity outside the class clusters are selected as the target sub-graphs.

It should be noted that target titles with similar semantics exist in the target subgraph, so that a plurality of article synopsis can be generated according to different branches in the target subgraph. In practical application, the generated article outlines can be recommended to the user at the same time for the user to select.

To sum up, when an article outline is generated, a plurality of reference articles conforming to a preset theme are obtained, paragraph titles of each reference article are clustered, a global title connection graph including all the paragraph titles of the reference articles is obtained, and finally the article outline is generated based on the global title connection graph for a user to refer to; in addition, the article outline is generated based on the paragraph titles of the multiple reference articles, so that the reliability of the generated article outline is enhanced, and the writing auxiliary effect is favorably improved.

It should be noted that, for simplicity of description, the method embodiments are described as a series of acts or combination of acts, but those skilled in the art will recognize that the present invention is not limited by the illustrated order of acts, as some steps may occur in other orders or concurrently in accordance with the embodiments of the present invention. Further, those skilled in the art will appreciate that the embodiments described in the specification are presently preferred and that no particular act is required to implement the invention.

Device embodiment

Referring to fig. 2, a block diagram of an embodiment of an article outline generation apparatus according to the present invention is shown, where the apparatus may include:

the reference article acquisition module 201 is configured to acquire at least two reference articles that meet a preset topic and determine a paragraph of each reference article;

a candidate headline determining module 202, configured to determine a candidate headline corresponding to each paragraph of each reference article;

a headline join graph generating module 203, configured to perform clustering on candidate headlines corresponding to each paragraph of the at least two reference articles to obtain a global headline join graph corresponding to the preset topic;

and the article outline generating module 204 is configured to generate an article outline corresponding to the preset topic according to the global title connection diagram.

Optionally, the title connection graph generating module includes:

the title merging submodule is used for merging all candidate titles corresponding to each paragraph to obtain a target title corresponding to each paragraph;

the internal headline connection graph generation sub-module is used for sequentially connecting target headlines corresponding to paragraphs belonging to the same reference article according to the sequence of the paragraphs to obtain at least two internal headline connection graphs, and each internal headline connection graph corresponds to one reference article;

and the global title connection graph generation sub-module is used for clustering each target title in the at least two internal title connection graphs according to the semantics of the target title to obtain a global title connection graph corresponding to the preset theme.

Optionally, the global title connection graph generating sub-module includes:

the clustering unit is used for clustering all target titles belonging to different reference articles according to the semantics of the target titles to obtain a cluster to which each target title belongs;

and the global title connection graph generating unit is used for connecting all target titles belonging to the same cluster in the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset theme.

Optionally, the global title connection map generating unit includes:

a target title judging subunit, configured to judge whether at least two identical target titles exist in the same cluster of the at least two internal title connection graphs;

the target title merging subunit is configured to merge at least two identical target titles to obtain a merged target title if the at least two identical target titles exist in the same class cluster, and update the target title in the class cluster to which the merged target title belongs according to the merged target title;

and the target title connection subunit is used for connecting the updated target titles in the same cluster of the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset theme.

Optionally, the article outline generating module includes:

the target subgraph extraction submodule is used for extracting the target subgraph in the global title connection graph based on a community discovery algorithm;

and the article outline generation submodule is used for generating the article outline corresponding to the preset theme according to the target subgraph.

Optionally, the candidate titles include a first candidate title and a second candidate title, and the candidate title determination module includes:

the paragraph title generation submodule is used for generating at least two paragraph titles aiming at each paragraph of each reference article;

the first candidate title determining sub-module is used for determining a first candidate title from the paragraph titles of each paragraph, and the similarity between the first candidate title and other paragraph titles of the paragraph to which the first candidate title belongs meets a first preset condition;

and the second candidate title determining submodule is used for determining a second candidate title of which the similarity with the first candidate title meets a second preset condition.

Optionally, the first candidate title determining sub-module includes:

the system comprises a correlation index calculation unit, a correlation index calculation unit and a display unit, wherein the correlation index calculation unit is used for calculating the sum of the similarity of a first paragraph title of the same paragraph and other paragraph titles to obtain the correlation index of the first paragraph title, and the first paragraph title is any paragraph title of at least two paragraph titles of the paragraph to which the first paragraph title belongs;

a first candidate title determining unit, configured to determine that the first paragraph title is a first candidate title if the correlation index of the first paragraph title is greater than the correlation indexes of the other paragraph titles of the paragraph to which the first paragraph title belongs.

Optionally, the second preset condition includes: the similarity between the second candidate title and the first candidate title is smaller than the similarity between a second paragraph title and the first candidate title, and the second paragraph title is any paragraph title which does not belong to the candidate title in at least two paragraph titles of the paragraph to which the first candidate title belongs.

Optionally, the reference article obtaining module includes:

the data receiving submodule is used for receiving a preset theme and a screening condition input by a user, wherein the screening condition comprises at least one item of article types, quoted times, publication time, article reading quantity and article praise quantity;

the article screening submodule is used for searching at least two reference articles meeting the preset theme according to the screening condition;

and the paragraph determining sub-module is used for determining the paragraph of each reference article based on the pre-trained segmentation model.

Optionally, the reference article obtaining module further includes:

the segmentation module construction submodule is used for constructing a segmentation model based on the bidirectional self-coding model and the conditional random field;

the corpus acquiring submodule is used for acquiring a segmented corpus, and the segmented corpus comprises a sample article and segmentation points corresponding to the sample article;

and the segmental model training submodule is used for training the segmental model according to the segmental training corpus until the segmental model meets a convergence condition to obtain a trained segmental model.

To sum up, when an article outline is generated, a plurality of reference articles conforming to a preset theme are obtained, paragraph titles of each reference article are clustered, a global title connection graph including all the paragraph titles of the reference articles is obtained, and finally the article outline is generated based on the global title connection graph for a user to refer to; in addition, the article outline is generated based on the paragraph titles of the multiple reference articles, so that the reliability of the generated article outline is enhanced, and the writing auxiliary effect is favorably improved.

For the device embodiment, since it is basically similar to the method embodiment, the description is simple, and for the relevant points, refer to the partial description of the method embodiment.

The embodiments in the present specification are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.

With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.

An embodiment of the present invention provides an apparatus for generating an article outline, the apparatus comprising a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors comprise instructions for:

a11, acquiring at least two reference articles conforming to a preset theme and determining paragraphs of each reference article;

a12, determining candidate titles corresponding to each paragraph of each reference article;

a13, clustering candidate titles corresponding to each paragraph of the at least two reference articles to obtain a global title connection graph corresponding to the preset topic;

and A14, generating an article outline corresponding to the preset theme according to the global title connection diagram.

Optionally, the clustering the candidate titles corresponding to each paragraph of the at least two reference articles to obtain a global title connection graph corresponding to the preset topic includes:

merging each candidate title corresponding to each paragraph to obtain a target title corresponding to each paragraph;

sequentially connecting target headings corresponding to paragraphs belonging to the same reference article according to the sequence of the paragraphs to obtain at least two internal heading connection graphs, wherein each internal heading connection graph corresponds to one reference article;

and clustering each target title in the at least two internal title connection graphs according to the semantics of the target title to obtain a global title connection graph corresponding to the preset theme.

Optionally, the clustering, according to the semantics of the target title, each target title in the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset topic includes:

clustering all target titles belonging to different reference articles according to the semantics of the target titles to obtain a cluster to which each target title belongs;

and connecting all target titles belonging to the same cluster in the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset theme.

Optionally, the connecting each target title belonging to the same cluster in the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset topic includes:

judging whether at least two identical target titles exist in the same cluster of the at least two internal title connection graphs or not;

if at least two identical target titles exist in the same cluster, merging the at least two identical target titles to obtain a merged target title, and updating the target title in the cluster to which the merged target title belongs according to the merged target title;

and connecting the updated target titles in the same cluster of the at least two internal title connection graphs to obtain a global title connection graph corresponding to the preset theme.

Optionally, the generating the article outline corresponding to the preset topic according to the global title connection diagram includes:

extracting a target subgraph in the global title connection graph based on a community discovery algorithm;

and generating an article outline corresponding to the preset theme according to the target subgraph.

Optionally, the determining the candidate headline corresponding to each paragraph of each reference article includes:

generating at least two paragraph titles for each paragraph of each reference article;

determining a first candidate title from the paragraph titles of each paragraph, wherein the similarity between the first candidate title and other paragraph titles of the paragraph to which the first candidate title belongs meets a first preset condition;

and determining a second candidate title of which the similarity with the first candidate title meets a second preset condition.

Optionally, the determining a first candidate title from the respective paragraph titles of each paragraph includes:

calculating the sum of the similarity of a first paragraph title of the same paragraph and other paragraph titles to obtain a correlation index of the first paragraph title, wherein the first paragraph title is any paragraph title of at least two paragraph titles of the paragraph to which the first paragraph title belongs;

and if the correlation index of the first paragraph title is larger than the correlation indexes of other paragraph titles of the paragraph to which the first paragraph title belongs, determining that the first paragraph title is a first candidate title.

Optionally, the second preset condition includes: the similarity between the second candidate title and the first candidate title is smaller than the similarity between a second paragraph title and the first candidate title, and the second paragraph title is any paragraph title which does not belong to the candidate title in at least two paragraph titles of the paragraph to which the first candidate title belongs.

Optionally, the obtaining at least two reference articles meeting a preset topic and determining a paragraph of each reference article includes:

receiving a preset theme and a screening condition input by a user, wherein the screening condition comprises at least one of article types, quoted times, publication time, article reading amount and article praise amount;

searching at least two reference articles meeting the preset theme according to the screening condition;

the paragraphs of each reference article are determined based on a pre-trained segmentation model.

Optionally, the device is also configured to execute the one or more programs by the one or more processors including instructions for:

constructing a segmentation model based on a bidirectional self-coding model and a conditional random field;

acquiring a segmented training corpus, wherein the segmented training corpus comprises a sample article and segmentation points corresponding to the sample article;

and training the segmented model according to the segmented training corpus until the segmented model meets a convergence condition to obtain the trained segmented model.

Fig. 3 is a block diagram illustrating an apparatus 800 for generating an article outline according to an example embodiment. For example, the apparatus 800 may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, an exercise device, a personal digital assistant, and the like.

Referring to fig. 3, the apparatus 800 may include one or more of the following components: processing component 802, memory 804, power component 806, multimedia component 808, audio component 810, input/output (I/O) interface 812, sensor component 814, and communication component 816.

The processing component 802 generally controls overall operation of the device 800, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations. The processing elements 802 may include one or more processors 820 to execute instructions to perform all or a portion of the steps of the methods described above. Further, the processing component 802 can include one or more modules that facilitate interaction between the processing component 802 and other components. For example, the processing component 802 can include a multimedia module to facilitate interaction between the multimedia component 808 and the processing component 802.

The memory 804 is configured to store various types of data to support operation at the device 800. Examples of such data include instructions for any application or method operating on device 800, contact data, phonebook data, messages, pictures, videos, and so forth. The memory 804 may be implemented by any type or combination of volatile or non-volatile memory devices such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disks.

Power components 806 provide power to the various components of device 800. The power components 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the apparatus 800.

The multimedia component 808 includes a screen that provides an output interface between the device 800 and a user. In some embodiments, the screen may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive an input signal from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 808 includes a front facing camera and/or a rear facing camera. The front-facing camera and/or the rear-facing camera may receive external multimedia data when the device 800 is in an operating mode, such as a shooting mode or a video mode. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

The audio component 810 is configured to output and/or input audio signals. For example, the audio component 810 includes a Microphone (MIC) configured to receive external audio signals when the apparatus 800 is in an operational mode, such as a call mode, a recording mode, and a voice information processing mode. The received audio signals may further be stored in the memory 804 or transmitted via the communication component 816. In some embodiments, audio component 810 also includes a speaker for outputting audio signals.

The I/O interface 812 provides an interface between the processing component 802 and peripheral interface modules, which may be keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to: a home button, a volume button, a start button, and a lock button.

The sensor assembly 814 includes one or more sensors for providing various aspects of state assessment for the device 800. For example, the sensor assembly 814 may detect the open/closed status of the device 800, the relative positioning of components, such as a display and keypad of the apparatus 800, the change in position of the device 800 or a component of the device 800, the presence or absence of user contact with the device 800, the orientation or acceleration/deceleration of the device 800, and the change in temperature of the device 800. Sensor assembly 814 may include a proximity sensor configured to detect the presence of a nearby object without any physical contact. The sensor assembly 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 814 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

The communication component 816 is configured to facilitate communications between the apparatus 800 and other devices in a wired or wireless manner. The device 800 may access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 816 receives a broadcast signal or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 816 further includes a Near Field Communication (NFC) module to facilitate short-range communications. For example, the NFC module may be implemented based on radio frequency information processing (RFID) technology, infrared data association (IrDA) technology, Ultra Wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

In an exemplary embodiment, the apparatus 800 may be implemented by one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, micro-controllers, microprocessors or other electronic components for performing the above-described methods.

In an exemplary embodiment, a non-transitory computer-readable storage medium comprising instructions, such as the memory 804 comprising instructions, executable by the processor 820 of the device 800 to perform the above-described method is also provided. For example, the non-transitory computer readable storage medium may be a ROM, a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.

Fig. 4 is a schematic diagram of a server in some embodiments of the invention. The server 1900 may vary widely by configuration or performance and may include one or more Central Processing Units (CPUs) 1922 (e.g., one or more processors) and memory 1932, one or more storage media 1930 (e.g., one or more mass storage devices) storing applications 1942 or data 1944. Memory 1932 and storage medium 1930 can be, among other things, transient or persistent storage. The program stored in the storage medium 1930 may include one or more modules (not shown), each of which may include a series of instructions operating on a server. Still further, a central processor 1922 may be provided in communication with the storage medium 1930 to execute a series of instruction operations in the storage medium 1930 on the server 1900.

The server 1900 may also include one or more power supplies 1926, one or more wired or wireless network interfaces 1950, one or more input-output interfaces 1958, one or more keyboards 1956, and/or one or more operating systems 1941, such as Windows Server, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.

A non-transitory computer-readable storage medium in which instructions, when executed by a processor of an apparatus (server or terminal), enable the apparatus to perform the article outline generation method shown in fig. 1.

A non-transitory computer-readable storage medium in which instructions, when executed by a processor of an apparatus (server or terminal), enable the apparatus to perform a method of generating an article schema, the method comprising: acquiring at least two reference articles conforming to a preset theme and determining paragraphs of each reference article; determining a candidate title corresponding to each paragraph of each reference article; clustering candidate titles corresponding to each paragraph of the at least two reference articles to obtain a global title connection graph corresponding to the preset topic; and generating an article outline corresponding to the preset theme according to the global title connection diagram.

Other embodiments of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the invention and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the invention being indicated by the following claims.

It will be understood that the invention is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the invention is limited only by the appended claims.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

The article outline generating method, the article outline generating device and the device for generating the article outline provided by the invention are described in detail, specific examples are applied in the text to explain the principle and the implementation mode of the invention, and the description of the above embodiments is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.