阅读说明：本技术 数据处理的方法、训练神经网络模型的方法及设备 (Data processing method, method and equipment for training neural network model ) 是由 李成 于 2019-08-07 设计创作，主要内容包括：一种数据处理的方法,包括：获取多个待处理数据(501)；使用第一神经网络模型对多个待处理数据进行处理,得到与多个待处理数据一一对应的多个第一矢量(502),其中,第一神经网络模型是基于通用数据训练获得；获取第一关联关系信息(503),第一关联关系信息用于指示至少一个第一矢量组,每个第一矢量组包括满足先验假设的两个第一矢量；将多个第一矢量以及第一关联关系信息输入第二神经网络模型,得到针对第一待处理数据的处理结果(504),第一待处理数据是多个待处理数据中的任一数据。所述数据处理的方法的目的在于削弱神经网络模型对待训练数据的依赖性。(A method of data processing, comprising: acquiring a plurality of data to be processed (501); processing the data to be processed by using a first neural network model to obtain a plurality of first vectors (502) which are in one-to-one correspondence with the data to be processed, wherein the first neural network model is obtained based on general data training; acquiring first incidence relation information (503), wherein the first incidence relation information is used for indicating at least one first vector group, and each first vector group comprises two first vectors meeting the prior hypothesis; the plurality of first vectors and the first incidence relation information are input into a second neural network model, and a processing result (504) aiming at first data to be processed is obtained, wherein the first data to be processed is any one of the plurality of data to be processed. The method for processing the data aims to weaken the dependence of the neural network model on the data to be trained.)

A method of data processing, comprising:

acquiring a plurality of data to be processed;

processing the plurality of data to be processed by using a first neural network model to obtain a plurality of first vectors which are in one-to-one correspondence with the plurality of data to be processed, wherein the first neural network model is obtained based on general data training;

acquiring first incidence relation information, wherein the first incidence relation information is used for indicating at least one first vector group, and each first vector group comprises two first vectors meeting a priori assumption;

and inputting the plurality of first vectors and the first incidence relation information into a second neural network model to obtain a processing result aiming at first data to be processed, wherein the first data to be processed is any one of the plurality of data to be processed.

The method according to claim 1, wherein the first association information is used to indicate N first vector groups, where N is an integer greater than 1, and before the inputting the plurality of first vectors and the first association information into a second neural network model to obtain a processing result for the first data to be processed, the method further comprises:

acquiring second incidence relation information, wherein the second incidence relation information is used for indicating N second vector groups, the N second vector groups belong to the N first vector groups, N is smaller than N, and N is a positive integer;

inputting the plurality of first vectors and the first incidence relation information into a second neural network model to obtain a processing result for first data to be processed, including:

and inputting the plurality of first vectors, the first incidence relation information and the second incidence relation information into the second neural network model to obtain a processing result aiming at the first to-be-processed data.

The method of claim 1 or 2, wherein the obtaining a plurality of data to be processed comprises:

acquiring target data, wherein the target data is one of the plurality of data to be processed;

acquiring association data, wherein the association data and the target data have an association relation meeting the prior assumption, and the plurality of data to be processed comprise the association data.

The method according to any one of claims 1 to 3, wherein the first association information includes an association matrix, a vector in a first dimension in the association matrix includes a plurality of elements in one-to-one correspondence with the plurality of first vectors, and a vector in a second dimension in the association matrix includes a plurality of elements in one-to-one correspondence with the plurality of first vectors, wherein any element in the association matrix is used to indicate whether there is an association between the vector corresponding to the any element in the first dimension and the vector corresponding to the any element in the second dimension, which satisfies the prior assumption.

The method of any one of claims 1 to 4, wherein the weight parameters of the second neural network model are obtained by:

acquiring a plurality of data to be trained;

processing the multiple data to be trained by using the first neural network model to obtain multiple fourth vectors which correspond to the multiple data to be trained one by one;

obtaining third association relation information, wherein the third association relation information is used for indicating at least one third vector group, and each third vector group comprises two fourth vectors meeting the prior hypothesis;

and inputting the fourth vectors and the third correlation information into the second neural network model to obtain a first processing result aiming at first data to be trained, wherein the first data to be trained is any one of the data to be trained, and the first processing result is used for correcting the weight parameters of the second neural network model.

The method of claim 5, wherein obtaining the first processing result for the first data to be trained comprises:

obtaining the first processing result and a second processing result aiming at second data to be trained, wherein the label of the first data to be trained is a first label, the label of the second data to be trained is a second label, and the first data to be trained and the second data to be trained are any two data in the plurality of data to be trained;

the method further comprises the following steps:

and matching the similarity between the first label and the second label with the similarity between the first processing result and the second processing result to obtain a matching result, wherein the matching result is used for correcting the weight parameter of the second neural network model.

The method according to claim 5 or 6, wherein the third correlation information is used to indicate M third vector groups, M being an integer greater than 1, and before the inputting the plurality of fourth vectors and the third correlation information into the second neural network model to obtain the first processing result for the first data to be trained, the method further comprises:

acquiring fourth incidence relation information, wherein the fourth incidence relation information is used for indicating M fourth vector groups, the M fourth vector groups belong to the M third vector groups, M is smaller than M, and M is a positive integer;

inputting the fourth vectors and the third correlation information into the second neural network model to obtain a first processing result for first data to be trained, including:

and inputting the fourth vectors, the third association relation information and the fourth association relation information into the second neural network model to obtain the first processing result.

The method of any of claims 5 to 7, wherein the first processing result is further used to modify a weight parameter of the first neural network model.

The method of any one of claims 5 to 8, wherein the plurality of data to be trained comprises one or more target type data, each target type data having a label for modifying the weight parameter.

A method of training a neural network model, comprising:

acquiring a plurality of data to be trained;

processing the multiple data to be trained by using a first neural network model to obtain multiple fourth vectors which correspond to the multiple data to be trained one by one;

obtaining third association relation information, wherein the third association relation information is used for indicating at least one third vector group, and each third vector group comprises two fourth vectors meeting the prior hypothesis;

and inputting the fourth vectors and the third correlation information into a second neural network model to obtain a first processing result aiming at first data to be trained, wherein the first data to be trained is any one of the data to be trained, and the first processing result is used for correcting the weight parameters of the second neural network model.

The method of claim 10, wherein obtaining the first processing result for the first data to be trained comprises:

obtaining the first processing result and a second processing result aiming at second data to be trained, wherein the label of the first data to be trained is a first label, the label of the second data to be trained is a second label, and the first data to be trained and the second data to be trained are any two data in the plurality of data to be trained;

the method further comprises the following steps:

and matching the similarity between the first label and the second label with the similarity between the first processing result and the second processing result to obtain a matching result, wherein the matching result is used for correcting the weight parameter of the second neural network model.

The method according to claim 10 or 11, wherein the third correlation information is used to indicate M third vector groups, and before the inputting the plurality of fourth vectors and the third correlation information into the second neural network model to obtain the first processing result for the first data to be trained, the method further comprises:

acquiring fourth incidence relation information, wherein the fourth incidence relation information is used for indicating M fourth vector groups, the M fourth vector groups belong to the M third vector groups, M is smaller than M, and M is a positive integer;

inputting the fourth vectors and the third correlation information into the second neural network model to obtain a first processing result for first data to be trained, including:

and inputting the fourth vectors, the third association relation information and the fourth association relation information into the second neural network model to obtain the first processing result.

The method of any one of claims 10 to 12, wherein the first processing result is further used to modify a weight parameter of the first neural network model.

The method of any one of claims 10 to 13, wherein the plurality of data to be trained comprises one or more target type data, each target type data having a label for modifying the weight parameter.

An apparatus for data processing, comprising:

the acquisition module is used for acquiring a plurality of data to be processed;

the processing module is used for processing the data to be processed by using a first neural network model to obtain a plurality of first vectors which are in one-to-one correspondence with the data to be processed, wherein the first neural network model is obtained based on general data training;

the obtaining module is further configured to obtain first association relationship information, where the first association relationship information is used to indicate at least one first vector group, and each first vector group includes two first vectors that satisfy a priori assumption;

the processing module is further configured to input the plurality of first vectors and the first incidence relation information into a second neural network model, so as to obtain a processing result for first to-be-processed data, where the first to-be-processed data is any one of the plurality of to-be-processed data.

The apparatus of claim 15, wherein the first association information is used to indicate N first vector groups, N being an integer greater than 1, before the processing module inputs the plurality of first vectors and the first association information into a second neural network model to obtain a processing result for the first data to be processed,

the obtaining module is further configured to obtain second association relationship information, where the second association relationship information is used to indicate N second vector groups, where the N second vector groups belong to the N first vector groups, N is smaller than N, and N is a positive integer;

the processing module is specifically configured to input the plurality of first vectors, the first incidence relation information, and the second incidence relation information into the second neural network model, so as to obtain a processing result for the first to-be-processed data.

The device according to claim 15 or 16, wherein the obtaining module is specifically configured to:

acquiring target data, wherein the target data is one of the plurality of data to be processed;

acquiring association data, wherein the association data and the target data have an association relation meeting the prior assumption, and the plurality of data to be processed comprise the association data.

The apparatus according to any one of claims 15 to 17, wherein the first association information comprises an association matrix, a vector in a first dimension of the association matrix comprises a plurality of elements in one-to-one correspondence with the plurality of first vectors, and a vector in a second dimension of the association matrix comprises a plurality of elements in one-to-one correspondence with the plurality of first vectors, wherein any element of the association matrix is used to indicate whether there is an association between the vector corresponding to the any element in the first dimension and the vector corresponding to the any element in the second dimension, which satisfies the prior assumption.

The apparatus according to any one of claims 15 to 18,

the acquisition module is further used for acquiring a plurality of data to be trained;

the processing module is further configured to process the multiple data to be trained by using the first neural network model to obtain multiple fourth vectors corresponding to the multiple data to be trained one by one;

the obtaining module is further configured to obtain third association relationship information, where the third association relationship information is used to indicate at least one third vector group, and each third vector group includes two fourth vectors that satisfy the prior hypothesis;

the processing module is further configured to input the plurality of fourth vectors and the third correlation information into the second neural network model to obtain a first processing result for first data to be trained, where the first data to be trained is any one of the plurality of data to be trained, and the first processing result is used to correct a weight parameter of the second neural network model.

The apparatus of claim 19,

the processing module is specifically configured to obtain the first processing result and a second processing result for second data to be trained, where a label of the first data to be trained is a first label, and a label of the second data to be trained is a second label;

the processing module is further configured to match the similarity between the first tag and the second tag with the similarity between the first processing result and the second processing result to obtain a matching result, and the matching result is used to correct the weight parameter of the second neural network model.

The apparatus of claim 19 or 20, wherein the third correlation information is used to indicate M third vector groups, M being an integer greater than 1, before the processing module inputs the plurality of fourth vectors and the third correlation information into the second neural network model to obtain the first processing result for the first data to be trained,

the obtaining module is further configured to obtain fourth association relationship information, where the fourth association relationship information is used to indicate M fourth vector groups, the M fourth vector groups belong to the M third vector groups, M is smaller than M, and M is a positive integer;

the processing module is specifically configured to input the plurality of fourth vectors, the third association relationship information, and the fourth association relationship information into the second neural network model, so as to obtain the first processing result.

The apparatus of any one of claims 19 to 21, wherein the first processing result is further used to modify a weight parameter of the first neural network model.

The apparatus of any one of claims 19 to 22, wherein the plurality of data to be trained comprises one or more target type data, each target type data having a label for modifying the weight parameter.

An apparatus for training a neural network model, comprising:

the acquisition module is used for acquiring a plurality of data to be trained;

the processing module is used for processing the data to be trained by using a first neural network model to obtain a plurality of fourth vectors which are in one-to-one correspondence with the data to be trained;

the obtaining module is further configured to obtain third association relationship information, where the third association relationship information is used to indicate at least one third vector group, and each third vector group includes two fourth vectors that satisfy the prior hypothesis;

the processing module is further configured to input the plurality of fourth vectors and the third correlation information into a second neural network model to obtain a first processing result for first data to be trained, where the first data to be trained is any one of the plurality of data to be trained, and the first processing result is used to correct a weight parameter of the second neural network model.

The apparatus according to claim 24, wherein the processing module is specifically configured to obtain the first processing result and a second processing result for second data to be trained, where a label of the first data to be trained is a first label, and a label of the second data to be trained is a second label;

the processing module is further configured to match the similarity between the first tag and the second tag with the similarity between the first processing result and the second processing result to obtain a matching result, where the matching result is used to correct the weight parameter of the second neural network model.

The apparatus according to claim 24 or 25, wherein the third correlation information is used to indicate M third vector groups, before the processing module is used to input the plurality of fourth vectors and the third correlation information into the second neural network model to obtain the first processing result for the first data to be trained,

the obtaining module is further configured to obtain fourth association relationship information, where the fourth association relationship information is used to indicate M fourth vector groups, the M fourth vector groups belong to the M third vector groups, M is smaller than M, and M is a positive integer;

the processing module is specifically configured to input the plurality of fourth vectors, the third association relationship information, and the fourth association relationship information into the second neural network model, so as to obtain the first processing result.

The apparatus of any one of claims 24 to 26, wherein the first processing result is further used to modify a weight parameter of the first neural network model.

The apparatus of any one of claims 24 to 27, wherein the plurality of data to be trained comprises one or more target type data, each target type data having a label for modifying the weight parameter.

A computer-readable storage medium, characterized in that the computer-readable medium stores program code for execution by a device, the program code comprising instructions for performing the method of any of claims 1-14.

A chip comprising a processor and a data interface, the processor reading instructions stored on a memory through the data interface to perform the method of any one of claims 1-14.