CN116881450A - Information classification method, apparatus, computer device, storage medium, and program product - Google Patents

Information classification method, apparatus, computer device, storage medium, and program product Download PDF

Info

Publication number
CN116881450A
CN116881450A CN202310686902.4A CN202310686902A CN116881450A CN 116881450 A CN116881450 A CN 116881450A CN 202310686902 A CN202310686902 A CN 202310686902A CN 116881450 A CN116881450 A CN 116881450A
Authority
CN
China
Prior art keywords
classification model
information classification
information
module
generate
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202310686902.4A
Other languages
Chinese (zh)
Inventor
徐晓健
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Bank of China Ltd
Original Assignee
Bank of China Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Bank of China Ltd filed Critical Bank of China Ltd
Priority to CN202310686902.4A priority Critical patent/CN116881450A/en
Publication of CN116881450A publication Critical patent/CN116881450A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification
    • G06F16/353—Clustering; Classification into predefined classes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Databases & Information Systems (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The application relates to the technical field of artificial intelligence, in particular to an information classification method, an information classification device, computer equipment, a storage medium and a program product. The method comprises the following steps: acquiring information data to be classified; inputting the information data to be classified into a target information classification model for information classification, and generating an information classification result of the information data to be classified; the target information classification model is generated by training and simplifying the initial information classification model based on the historical information data; outputting the information classification result of the information data to be classified. By adopting the method, the deployment performance of the target information classification model can be improved.

Description

Information classification method, apparatus, computer device, storage medium, and program product
Technical Field
The present application relates to the field of artificial intelligence technology, and in particular, to an information classification method, apparatus, computer device, storage medium, and program product.
Background
With the development of banking application software, a function of viewing news information may be added to the banking application software. Since various news information is generated every day, in order to better manage the various news information, it is necessary to classify the various news information according to the title and content of the news information.
In the conventional method, a neural network model is generally used to classify news information. However, the conventional method of classifying information by using the neural network model has the problem of poor deployment performance of the model.
Disclosure of Invention
In view of the foregoing, it is desirable to provide an information classification method, apparatus, computer device, storage medium, and program product that can improve model deployment performance.
In a first aspect, the present application provides a method for classifying information. The method comprises the following steps:
acquiring information data to be classified;
inputting the information data to be classified into a target information classification model to perform information classification, and generating an information classification result of the information data to be classified; the target information classification model is generated by training and simplifying an initial information classification model based on historical information data;
and outputting the information classification result of the information data to be classified.
In one embodiment, the method further comprises:
acquiring a plurality of historical information data; each history information data comprises history information content and a labeling classification result corresponding to the history information content;
training the initial information classification model according to a plurality of the historical information data to generate an intermediate information classification model;
Simplifying and training the intermediate information classification model according to a plurality of the historical information data to generate the target information classification model.
In one embodiment, the simplifying and training the intermediate information classification model according to a plurality of the historical information data to generate the target information classification model includes:
dividing each layer of neural network in the intermediate information classification model according to a preset classification standard to generate a plurality of modules; the module includes at least one layer of neural network;
simplifying a first module in the plurality of modules to generate a simplified intermediate information classification model, and training the simplified intermediate information classification model according to a plurality of historical information data to generate a first intermediate information classification model;
taking the next module of the first modules in the first intermediate information classification model as a new first module, circularly executing the simplification of the new first modules in the intermediate information classification model, generating a new simplified intermediate information classification model, training the new simplified intermediate information classification model according to a plurality of historical information data, and generating a new first intermediate information classification model; until iterating to the last module in the intermediate information classification model, generating the target information classification model.
In one embodiment, the dividing each layer of network in the intermediate information classification model according to a preset classification standard, to generate a plurality of modules includes:
determining the target layer number of the neural network in each module based on the total layer number of the neural network in the intermediate information classification model;
and dividing the neural network of each layer in the intermediate information classification model according to the target layer number of the neural network in each module to generate a plurality of modules.
In one embodiment, the simplifying the first module of the plurality of modules to generate the simplified intermediate information classification model includes:
selecting a target neural network from the neural networks contained in a first module in the plurality of modules;
replacing the first module by the target neural network to generate a simplified first module;
and generating the simplified intermediate information classification model based on the first simplified module and other modules in the intermediate information classification model.
In one embodiment, the training the initial information classification model according to a plurality of the historical information data to generate the intermediate information classification model includes:
Inputting the historical information content into the initial information classification model for processing aiming at each historical information data to generate a prediction classification result corresponding to the historical information content;
training the initial information classification model according to the prediction classification result corresponding to each historical information content and the labeling classification result corresponding to each historical information content to generate the intermediate information classification model.
In a second aspect, the application also provides an information classification device. The device comprises:
the acquisition module is used for acquiring information data to be classified;
the information classification module is used for inputting the information data to be classified into a target information classification model to perform information classification and generating an information classification result of the information data to be classified; the target information classification model is generated by training and simplifying an initial information classification model based on historical information data;
and the output module is used for outputting the information classification result of the information data to be classified.
In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor implementing the steps of the method in any of the embodiments of the first aspect described above when the processor executes the computer program.
In a fourth aspect, the present application also provides a computer-readable storage medium. The computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method in any of the embodiments of the first aspect described above.
In a fifth aspect, the present application also provides a computer program product. The computer program product comprising a computer program which, when executed by a processor, implements the steps of the method in any of the embodiments of the first aspect described above.
The information classification method, the information classification device, the computer equipment, the storage medium and the program product acquire information data to be classified; inputting the information data to be classified into a target information classification model for information classification, and generating an information classification result of the information data to be classified; the target information classification model is generated by training and simplifying the initial information classification model based on the historical information data; outputting the information classification result of the information data to be classified. Because the target information classification model is generated by training and simplifying the initial information classification model based on the historical information data, the target information classification model can reduce model parameters contained in the model on the premise of ensuring the accuracy of the model. Therefore, the target information classification model has better deployment performance. Therefore, the target information classification model with good deployment performance is used for classifying the information data to be classified, the information classification result of the information data to be classified can be accurately generated, and the deployment performance of the target information classification model can be improved.
Drawings
FIG. 1 is a diagram of an application environment of a method for classifying information in one embodiment;
FIG. 2 is a flow chart of a method for classifying information according to one embodiment;
FIG. 3 is a flow chart of a model simplification and training step in another embodiment;
FIG. 4 is a flowchart illustrating a target information classification model generating step according to an embodiment;
FIG. 5 is a flow diagram of a module partitioning step in one embodiment;
FIG. 6 is a simplified flowchart illustrating a procedure for generating an intermediate information classification model according to an embodiment;
FIG. 7 is a flow chart of an alternative embodiment of a method for classifying information;
FIG. 8 is a block diagram of an information classification apparatus according to an embodiment;
fig. 9 is an internal structural diagram of a computer device in one embodiment.
Detailed Description
The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.
With the development of banking application software, a function of viewing news information may be added to the banking application software. Since various news information is generated every day, in order to better manage the various news information, it is necessary to classify the various news information according to the title and content of the news information.
In the conventional method, a neural network model is generally used to classify news information. However, since the conventional neural network model includes many model parameters, the conventional information classification method using the neural network model has a problem of poor model deployment performance.
The information classification method provided by the embodiment of the application can be applied to an application environment shown in fig. 1. Wherein the terminal 102 communicates with the server 104 via a network. The data storage system may store data that the server 104 needs to process. The data storage system may be integrated on the server 104 or may be located on a cloud or other network server. The server 104 acquires information data to be classified; the server 104 inputs the information data to be classified into the target information classification model to perform information classification, and generates an information classification result of the information data to be classified; the target information classification model is generated by training and simplifying the initial information classification model based on the historical information data; the server 104 outputs the information classification result of the information data to be classified to the terminal 102. The terminal 102 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, internet of things devices, and portable wearable devices, where the internet of things devices may be smart speakers, smart televisions, smart air conditioners, smart vehicle devices, and the like. The portable wearable device may be a smart watch, smart bracelet, headset, or the like. The server 104 may be implemented as a stand-alone server or as a server cluster of multiple servers.
In one embodiment, as shown in fig. 2, a method for classifying information is provided, which is illustrated by using the method applied to the server 104 in fig. 1 as an example, and includes the following steps:
s220, obtaining information data to be classified.
The information data to be classified refers to news information data which needs information classification. The information data to be classified may include contents of news information and titles of the news information. Alternatively, the server 104 may directly obtain the information data to be classified uploaded by the user from the terminal 102. Alternatively, the server 104 may determine the news information to be classified, so that the server 104 may obtain the data (i.e., the information data to be classified) included in the news information. Of course, the method for acquiring the information data to be classified is not limited in the embodiment of the application.
S240, inputting the information data to be classified into a target information classification model for information classification, and generating an information classification result of the information data to be classified; the target information classification model is generated by training and simplifying the initial information classification model based on the history information data.
Alternatively, the server 104 may directly input the information data to be classified into the target information classification model for information classification, and generate an information classification result of the information data to be classified. Alternatively, the server 104 may perform data preprocessing on the information data to be classified to generate preprocessed information data to be classified; and inputting the preprocessed information data to be classified into the target information classification model for information classification, and generating an information classification result of the information data to be classified. By way of example, the process of data preprocessing may include, but is not limited to, data screening, data deduplication, and the like. Wherein the target information classification model is generated by training and simplifying the initial information classification model based on the history information data. The target information classification model may be any one of neural network models for which model training has been completed. By way of example, the target information classification model may include, but is not limited to, a dense neural network model with completed model training, a convolutional neural network model with completed model training, a recurrent neural network model with completed model training, and the like.
S260, outputting the information classification result of the information data to be classified.
Alternatively, the server 104 may output the information classification result of the information data to be classified to the terminal 102, so that the user may acquire the information classification result of the information data to be classified from the terminal 102. Alternatively, the server 104 may store the information classification result of the information data to be classified in a predetermined database, so that the user may acquire the information classification result of the information data to be classified from the predetermined database. The information classification result of the information data to be classified refers to an information classification result obtained after the information data to be classified is classified. Illustratively, the information classification results may include, but are not limited to, entertainment information, policy news, financial news, and the like.
In the information classification method, information data to be classified is obtained; inputting the information data to be classified into a target information classification model for information classification, and generating an information classification result of the information data to be classified; the target information classification model is generated by training and simplifying the initial information classification model based on the historical information data; outputting the information classification result of the information data to be classified. Because the target information classification model is generated by training and simplifying the initial information classification model based on the historical information data, the target information classification model can reduce model parameters contained in the model on the premise of ensuring the accuracy of the model. Therefore, the target information classification model has better deployment performance. Therefore, the target information classification model with good deployment performance is used for classifying the information data to be classified, the information classification result of the information data to be classified can be accurately generated, and the deployment performance of the target information classification model can be improved.
In the above embodiments, the target information classification model is generated by training and simplifying the initial information classification model based on the history information data, and the specific method for training and simplifying the model is described below. In one embodiment, as shown in fig. 3, the above information classification method further includes:
s320, acquiring a plurality of historical information data; each history information data includes the label classification result corresponding to the history information content.
Alternatively, the server 104 may obtain a plurality of history information data directly from the terminal 102 or a preset database. Alternatively, the server 104 may obtain a plurality of historical information contents from the public news website; then, classifying the plurality of historical information contents to obtain labeling classification results corresponding to the historical information contents; thus, a plurality of history information data can be generated based on the history information content and the label classification result corresponding to the history information content. Wherein each history information data includes each history information content and the label classification result corresponding to each history information content. The history information content refers to content of history news information or title information. The label classification result corresponding to the historical information content is the real information classification result corresponding to the historical information content.
S340, training the initial information classification model according to the plurality of historical information data to generate an intermediate information classification model.
Alternatively, the server 104 may first construct an initial information classification model. The initial information classification model can be any neural network model without model training. By way of example, the initial information classification model may include, but is not limited to, a dense neural network model without model training, a convolutional neural network model without model training, a recurrent neural network model without model training, and the like. Thus, the server 104 can train the initial information classification model according to the plurality of historical information contents and the labeling classification results corresponding to the plurality of historical information contents to generate the intermediate information classification model. The intermediate information classification model is a model obtained by performing one round of training on the intermediate information classification model.
In one alternative embodiment, S340 includes:
s342, for each history information data, input the history information content into the initial information classification model for processing, and generate the prediction classification result corresponding to the history information content.
Alternatively, for each historical information data, the server 104 may input the historical information content into the initial information classification model for information classification, and generate a prediction classification result corresponding to the historical information content. The predicted classification result corresponding to the historical information content is the predicted information classification result corresponding to the historical information content. Alternatively, the server 104 may perform data preprocessing on the historical information content to generate preprocessed historical information content; and inputting the preprocessed historical information content into the initial information classification model to classify the information, and generating a prediction classification result corresponding to the historical information content. The predicted classification result corresponding to the historical information content is the predicted information classification result corresponding to the historical information content.
S344, training the initial information classification model according to the prediction classification result corresponding to each historical information content and the labeling classification result corresponding to each historical information content to generate an intermediate information classification model.
Optionally, for each historical information data, the server 104 may calculate a value of the loss function according to the prediction classification result corresponding to the historical information content and the labeling classification result corresponding to the historical information content. Thus, according to the calculated value of the loss function, the server 104 may adjust the initial model parameters of the initial information classification model, thereby obtaining the preset model parameters and the preset information classification model corresponding to the preset model parameters. Then, the history information content is inputted into the preset information classification model to obtain the new prediction classification result. Then, the server 104 may calculate the new value of the loss function again according to the labeling classification result and the new prediction classification result, until the new value of the loss function reaches the minimum value (i.e. the model converges), and take the preset model parameter corresponding to the value of the loss function at this time as the target model parameter. And updating the initial model parameters of the initial information classification model based on the target model parameters to generate an intermediate information classification model.
S360, simplifying and training the intermediate information classification model according to the plurality of historical information data to generate the target information classification model.
Alternatively, the server 104 may simplify the intermediate information classification model first, and generate a simplified intermediate information classification model; then, the server 104 can train the simplified intermediate information classification model according to the plurality of historical information data to generate the target information classification model. Alternatively, the server 104 may first simplify a part of the intermediate information classification model to generate a partially simplified intermediate information classification model; training the intermediate information classification model after partial simplification according to a plurality of historical information data to generate a round of intermediate information classification model after simplification and training; then, the server 104 may simplify other parts of the intermediate information classification model after a round of simplification and training, and generate a intermediate information classification model after simplification again; training the intermediate information classification model after the simplification again according to the plurality of historical information data to generate a target information classification model.
In this embodiment, a plurality of history information data are obtained, so that more accurate history information content and label classification results corresponding to the history information content can be provided for the initial information classification model. Then, training the initial information classification model according to more accurate historical information data, so as to generate more accurate intermediate information classification model. Therefore, the intermediate information classification model is simplified and trained according to more accurate historical information data, and a target information classification model with simplified model and more accurate model can be generated.
In the above embodiments, the simplification and training of the intermediate information classification model based on a plurality of history information data are involved to generate the target information classification model, and the specific method thereof will be described below. In one embodiment, as shown in fig. 4, S360 includes:
s420, dividing each layer of neural network in the intermediate information classification model according to a preset classification standard to generate a plurality of modules; the module includes at least one layer of neural network.
Optionally, the server 104 may divide the neural network of each layer in the intermediate information classification model according to a preset classification standard, so as to generate a plurality of modules in the intermediate information classification model. Wherein the intermediate information classification model comprises at least one layer of neural network. Each module in the intermediate information classification model includes at least one layer of neural network. The preset classification standard refers to a classification standard which needs to be adopted when the module is divided. By way of example, the preset classification criteria may include, but are not limited to, dividing the plurality of modules according to an average number of layers of the neural network, dividing the plurality of modules according to a number of model parameters in each layer of the neural network, and the like. Of course, the present embodiment is not limited to the preset classification standard.
For example, the server 104 may obtain the number of model parameters included in each layer of neural network in the intermediate information classification model, so as to determine each layer of neural network with the same number of model parameters. Then, the server 104 may divide the neural network of each layer with the same number of model parameters according to the number of model parameters, so as to generate a plurality of modules in the intermediate information classification model.
S440, simplifying a first module in the plurality of modules to generate a simplified intermediate information classification model, and training the simplified intermediate information classification model according to the plurality of historical information data to generate the first intermediate information classification model.
Alternatively, for a first module of the plurality of modules, the server 104 may simplify the first module based on the neural networks of each layer in the first module, and generate a simplified intermediate information classification model. Then, the server 104 can input the historical information content into the simplified intermediate information classification model for processing according to each historical information data, and generate an intermediate prediction classification result corresponding to the historical information content. Thus, the server 104 may train the simplified intermediate information classification model according to the intermediate prediction classification result corresponding to each historical information content and the labeling classification result corresponding to each historical information content, so as to generate the first intermediate information classification model. The intermediate prediction classification result corresponding to the historical information content is a predicted information classification result corresponding to the historical information content, which is obtained by adopting the simplified intermediate information classification model. The first intermediate information classification model is a model obtained by performing one round of training on the simplified intermediate information classification model.
S460, the next module of the first modules in the first intermediate information classification model is used as a new first module, the new first module in the intermediate information classification model is circularly executed to simplify, a new simplified intermediate information classification model is generated, the new simplified intermediate information classification model is trained according to a plurality of historical information data, and the new first intermediate information classification model is generated; until iterating to the last module in the intermediate information classification model, a target information classification model is generated.
Alternatively, the server 104 may use a next module of the first modules in the first intermediate information classification model as a new first module, and the loop execution is performed on the new first module, and the server 104 may simplify the new first module based on the neural networks of each layer in the new first module, so as to generate a new simplified intermediate information classification model. Then, the server 104 can input the history information content into the new simplified intermediate information classification model for processing according to each history information data, and generate a new intermediate prediction classification result corresponding to the history information content. Thus, the server 104 may train the new simplified intermediate information classification model according to the new intermediate prediction classification result corresponding to each historical information content and the labeling classification result corresponding to each historical information content, and generate a new first intermediate information classification model. Until iterating to the last module in the intermediate information classification model, the server 104 may take the final new first intermediate information classification model as the target information classification model. The target information classification model is an information classification model which is trained after the number of layers of each module in the model is reduced.
In this embodiment, each layer of neural network in the intermediate information classification model is divided according to a preset classification standard, so that a plurality of modules in the intermediate information classification model can be generated. Simplifying a first module of the plurality of modules to generate a simplified intermediate information classification model, and training the simplified intermediate information classification model according to the plurality of historical information data to generate a first round of simplified and trained intermediate information classification model. The next module of the first modules in the first intermediate information classification model is used as a new first module, the steps of simplifying and training the new first modules in the intermediate information classification model are circularly executed, and a new first intermediate information classification model after second round simplification and training can be generated; until the last module in the intermediate information classification model is iterated, multiple rounds of simplified and trained target information classification models can be generated. Therefore, the target information classification model in the application can reduce model parameters contained in the model on the premise of ensuring the accuracy of the model.
In the above embodiment, the division of each layer of network in the intermediate information classification model according to the preset classification standard is involved to generate a plurality of modules, and a specific method thereof is described below. In one embodiment, as shown in fig. 5, S420 includes:
S520, determining the target layer number of the neural network in each module based on the total layer number of the neural network in the intermediate information classification model.
Alternatively, the server 104 may determine the total number of layers of the neural network in the intermediate information classification model. Thus, the server 104 may determine the target number of layers of the neural network in each module based on the total number of layers of the neural network in the intermediate information classification model. The target layer number of the neural network in each module refers to the layer number of the neural network contained in each module. For example, assuming that the intermediate information classification model includes K layers of neural networks, that is, the total number of layers of the neural networks in the intermediate information classification model is K layers, the server 104 may determine that the target number of layers of the neural networks in each module is M layers. Note that K may be divided by M.
S540, dividing the neural network of each layer in the intermediate information classification model according to the target layer number of the neural network of each module to generate a plurality of modules.
Alternatively, the server 104 may divide the neural network of each layer in the intermediate information classification model according to the target layer number of the neural network of each module, so as to generate a plurality of modules. Illustratively, it is assumed that the total number of layers of the neural network in the intermediate information classification model is K layers, and the target number of layers of the neural network in each module is M layers. Then, the server 104 may divide the K-layer neural network in the intermediate information classification model into L modules. Wherein the number of modules L is equal to K divided by M, i.e. M layers of neural network are included in each module. Note that K may be divided by L.
In the embodiment, the target layer number of the neural network in each module is determined based on the total layer number of the neural network in the intermediate information classification model; and dividing the neural network of each layer in the intermediate information classification model according to the target layer number of the neural network in each module to generate a plurality of modules. The intermediate information classification model can be more accurately divided into a plurality of modules based on the target layer number of the neural network contained in each module.
In the above embodiments, the simplification of the first module of the plurality of modules is involved, and the simplified intermediate information classification model is generated, and the specific method thereof is described below. In one embodiment, as shown in fig. 6, S440 includes:
s620, selecting a target neural network from the neural networks contained in the first module aiming at the first module in the plurality of modules.
Alternatively, for a first module of the plurality of modules, the server 104 may select the target neural network from the neural networks included in the first module of the plurality of modules. The target neural network may be any layer of neural network contained in the first module. Alternatively, the target neural network may be randomly selected; alternatively, the target neural network may be selected according to a preset sequence. For example, a first layer of neural network may be selected from a first module as the target neural network for the first module, a second layer of neural network may be selected from a second module as the target neural network for the second module, a third layer of neural network may be selected from a third module as the target neural network for the third module, and so on. The preset sequence may be set according to a neural network. Of course, the embodiment of the application does not limit the preset sequence and the selection mode of the target neural network.
And S640, replacing the first module by using the target neural network, and generating a simplified first module.
Alternatively, the server 104 may replace the first module with the target neural network, generating a simplified first module. The simplified first module is a first module obtained by replacing the neural network. For example, it is assumed that before simplification, the first module includes M-layer neural networks, and an i-th layer neural network of the M-layer neural networks is selected as a target neural network. Then, the i-th layer neural network may be used to replace the entire first module, so as to generate a simplified first module, where the simplified first module only includes the i-th layer neural network. The simplified formula of the model is shown in the following formula.
f i (x i )＝f M (f M-1 (…(f 2 (f 1 (x i ))))) (1)
Wherein x is i Input data representing a corresponding module of the ith neural network, f i () Output data representing an i-th layer neural network. Based on this, it is assumed that the target layer number of the neural network in each module is M layers, and each layer of the neural network includes N model parameters. Then the number of model parameters contained in each module is equal to N times M before simplification. After simplification, the number of model parameters contained in the simplified first module is equal to N.
S660, generating a simplified intermediate information classification model based on the first simplified module and other modules in the intermediate information classification model.
Alternatively, the server 104 may generate the simplified intermediate information classification model based on the simplified first module and other modules in the intermediate information classification model. For example, it is assumed that the simplified first module includes only the i-th layer neural network, and other modules in the intermediate information classification model include M-th layer neural networks, and the intermediate information classification model includes L modules. Then the other modules in the intermediate information classification model contain together (L-1) times the M-layer neural network. Based on this, the server 104 may multiply (L-1) of the i-th layer neural network and other modules by the M-layer neural network to connect, thereby generating a simplified intermediate information classification model.
In this embodiment, for a first module of the plurality of modules, a target neural network is selected from the neural networks included in the first module; replacing the first module by using a target neural network to generate a simplified first module; based on the simplified first module and other modules in the intermediate information classification model, a simplified intermediate information classification model is generated. The intermediate information classification model with fewer layers of the neural network can be obtained by performing model simplification in a mode of replacing the neural network, namely the intermediate information classification model with fewer model parameters can be obtained. Therefore, the deployment performance of the intermediate information classification model can be improved.
In an alternative embodiment, as shown in fig. 7, there is provided an information classification method applied to a server 104, including:
s702, acquiring a plurality of historical information data; each history information data comprises a label classification result corresponding to the history information content;
s704, inputting the historical information content into an initial information classification model for processing according to each historical information data, and generating a prediction classification result corresponding to the historical information content;
s706, training the initial information classification model according to the prediction classification result corresponding to each historical information content and the labeling classification result corresponding to each historical information content to generate an intermediate information classification model;
s708, determining the target layer number of the neural network in each module based on the total layer number of the neural network in the intermediate information classification model;
s710, dividing each layer of neural network in the intermediate information classification model according to the target layer number of the neural network in each module to generate a plurality of modules;
s712, selecting a target neural network from the neural networks contained in the first module aiming at the first module in the plurality of modules;
s714, replacing the first module by using a target neural network, and generating a simplified first module;
S716, generating a simplified intermediate information classification model based on the first simplified module and other modules in the intermediate information classification model;
s718, training the simplified intermediate information classification model according to a plurality of historical information data to generate a first intermediate information classification model;
s720, using the next module of the first modules in the first intermediate information classification model as a new first module, circularly executing the new first modules in the intermediate information classification model to simplify the new first modules, generating a new simplified intermediate information classification model, training the new simplified intermediate information classification model according to a plurality of historical information data, and generating a new first intermediate information classification model; until iterating to the last module in the intermediate information classification model, generating a target information classification model;
s722, obtaining information data to be classified;
s724, inputting the information data to be classified into the target information classification model for information classification, and generating an information classification result of the information data to be classified; the target information classification model is generated by training and simplifying the initial information classification model based on the historical information data;
s726, outputting the information classification result of the information data to be classified.
In the above information classification method, the target information classification model is generated by training and simplifying the initial information classification model based on the historical information data, so that the target information classification model in the application can reduce model parameters contained in the model on the premise of ensuring the accuracy of the model. Therefore, the target information classification model has better deployment performance. Therefore, the target information classification model with good deployment performance is used for classifying the information data to be classified, the information classification result of the information data to be classified can be accurately generated, and the deployment performance of the target information classification model can be improved. Therefore, the application provides an information classification method, which can train a target information classification model with fewer model parameters and higher precision by adopting a mode of replacing the whole module by a random neural network in each module and a multi-stage training mode in the model training process, thereby improving the deployment performance of the target information classification model.
It should be understood that, although the steps in the flowcharts related to the above embodiments are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.
Based on the same inventive concept, the embodiment of the application also provides an information classification device for realizing the above related information classification method. The implementation of the solution provided by the device is similar to that described in the above method, so the specific limitation of one or more embodiments of the information classification device provided below may be referred to the limitation of the information classification method hereinabove, and will not be repeated here.
In one embodiment, as shown in FIG. 8, an information classification apparatus 800 is provided, comprising: an acquisition module 820, an information classification module 840 and an output module 860, wherein:
the obtaining module 820 is configured to obtain information data to be classified.
The information classification module 840 is configured to input information data to be classified into the target information classification model for information classification, and generate an information classification result of the information data to be classified; the target information classification model is generated by training and simplifying the initial information classification model based on the history information data.
The output module 860 is configured to output an information classification result of the information data to be classified.
In one embodiment, the information classification apparatus 800 further comprises:
the historical information data acquisition module is used for acquiring a plurality of historical information data; each history information data comprises a label classification result corresponding to the history information content;
The intermediate information classification model generation module is used for training the initial information classification model according to a plurality of historical information data to generate an intermediate information classification model;
and the target information classification model generation module is used for simplifying and training the intermediate information classification model according to the plurality of historical information data to generate a target information classification model.
In one embodiment, the target information classification model generation module includes:
the module dividing unit is used for dividing each layer of neural network in the intermediate information classification model according to a preset classification standard to generate a plurality of modules; the module comprises at least one layer of neural network;
the first intermediate information classification model generation unit is used for simplifying a first module in the plurality of modules to generate a simplified intermediate information classification model, training the simplified intermediate information classification model according to the plurality of historical information data to generate a first intermediate information classification model;
the target information classification model generating unit is used for taking the next module of the first modules in the first intermediate information classification model as a new first module, circularly executing the simplification of the new first modules in the intermediate information classification model, generating a new simplified intermediate information classification model, training the new simplified intermediate information classification model according to a plurality of historical information data, and generating a new first intermediate information classification model; until iterating to the last module in the intermediate information classification model, a target information classification model is generated.
In one embodiment, the module dividing unit includes:
the target layer number determining subunit is used for determining the target layer number of the neural network in each module based on the total layer number of the neural network in the intermediate information classification model;
the module dividing subunit is used for dividing the neural network of each layer in the intermediate information classification model according to the target layer number of the neural network in each module to generate a plurality of modules.
In one embodiment, the first intermediate information classification model generation unit includes:
a target neural network selecting subunit, configured to select, for a first module of the plurality of modules, a target neural network from the neural networks included in the first module;
a simplified first module generating subunit, configured to replace the first module with a target neural network, and generate a simplified first module;
and the simplified intermediate information classification model generation subunit is used for generating a simplified intermediate information classification model based on the first simplified module and other modules in the intermediate information classification model.
In one embodiment, the intermediate information classification model generation module comprises:
the prediction classification result generation unit is used for inputting the historical information content into the initial information classification model for processing aiming at each historical information data to generate a prediction classification result corresponding to the historical information content;
The intermediate information classification model generating unit is used for training the initial information classification model according to the prediction classification result corresponding to each historical information content and the labeling classification result corresponding to each historical information content to generate an intermediate information classification model.
The above-mentioned individual modules in the information classification device may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, a computer device is provided, which may be a server, and the internal structure of which may be as shown in fig. 9. The computer device includes a processor, a memory, an Input/Output interface (I/O) and a communication interface. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface is connected to the system bus through the input/output interface. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is used for storing information classification data. The input/output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method of classifying information.
It will be appreciated by persons skilled in the art that the architecture shown in fig. 9 is merely a block diagram of some of the architecture relevant to the present inventive arrangements and is not limiting as to the computer device to which the present inventive arrangements are applicable, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.
In one embodiment, a computer device is provided comprising a memory and a processor, the memory having stored therein a computer program, the processor when executing the computer program performing the steps of:
acquiring information data to be classified;
inputting the information data to be classified into a target information classification model for information classification, and generating an information classification result of the information data to be classified; the target information classification model is generated by training and simplifying the initial information classification model based on the historical information data;
outputting the information classification result of the information data to be classified.
In one embodiment, the processor when executing the computer program further performs the steps of:
acquiring a plurality of historical information data; each history information data comprises a label classification result corresponding to the history information content;
Training the initial information classification model according to the plurality of historical information data to generate an intermediate information classification model;
simplifying and training the intermediate information classification model according to the plurality of historical information data to generate the target information classification model.
In one embodiment, the intermediate information classification model is simplified and trained based on a plurality of historical information data to generate a target information classification model, and the processor when executing the computer program further performs the steps of:
dividing each layer of neural network in the intermediate information classification model according to a preset classification standard to generate a plurality of modules; the module comprises at least one layer of neural network;
simplifying a first module in the plurality of modules to generate a simplified intermediate information classification model, training the simplified intermediate information classification model according to a plurality of historical information data to generate a first intermediate information classification model;
taking the next module of the first modules in the first intermediate information classification model as a new first module, circularly executing the new first modules in the intermediate information classification model to simplify the new intermediate information classification model, generating a new simplified intermediate information classification model, training the new simplified intermediate information classification model according to a plurality of historical information data, and generating a new first intermediate information classification model; until iterating to the last module in the intermediate information classification model, a target information classification model is generated.
In one embodiment, the method includes dividing each layer of network in the intermediate information classification model according to a preset classification standard to generate a plurality of modules, and when the processor executes the computer program, the following steps are further implemented:
determining the target layer number of the neural network in each module based on the total layer number of the neural network in the intermediate information classification model;
and dividing the neural network of each layer in the intermediate information classification model according to the target layer number of the neural network in each module to generate a plurality of modules.
In one embodiment, the first module of the plurality of modules is simplified to generate a simplified intermediate information classification model, and the processor when executing the computer program further performs the steps of:
selecting a target neural network from the neural networks contained in the first module aiming at the first module in the plurality of modules;
replacing the first module by using a target neural network to generate a simplified first module;
based on the simplified first module and other modules in the intermediate information classification model, a simplified intermediate information classification model is generated.
In one embodiment, training the initial information classification model based on a plurality of historical information data to generate an intermediate information classification model, the processor when executing the computer program further performs the steps of:
Inputting the historical information content into an initial information classification model for processing aiming at each historical information data to generate a prediction classification result corresponding to the historical information content;
training the initial information classification model according to the prediction classification result corresponding to each historical information content and the labeling classification result corresponding to each historical information content to generate an intermediate information classification model.
In one embodiment, a computer readable storage medium is provided having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring information data to be classified;
inputting the information data to be classified into a target information classification model for information classification, and generating an information classification result of the information data to be classified; the target information classification model is generated by training and simplifying the initial information classification model based on the historical information data;
outputting the information classification result of the information data to be classified.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring a plurality of historical information data; each history information data comprises a label classification result corresponding to the history information content;
Training the initial information classification model according to the plurality of historical information data to generate an intermediate information classification model;
simplifying and training the intermediate information classification model according to the plurality of historical information data to generate the target information classification model.
In one embodiment, the intermediate information classification model is simplified and trained based on a plurality of historical information data to generate a target information classification model, and the computer program when executed by the processor further performs the steps of:
dividing each layer of neural network in the intermediate information classification model according to a preset classification standard to generate a plurality of modules; the module comprises at least one layer of neural network;
simplifying a first module in the plurality of modules to generate a simplified intermediate information classification model, training the simplified intermediate information classification model according to a plurality of historical information data to generate a first intermediate information classification model;
taking the next module of the first modules in the first intermediate information classification model as a new first module, circularly executing the new first modules in the intermediate information classification model to simplify the new intermediate information classification model, generating a new simplified intermediate information classification model, training the new simplified intermediate information classification model according to a plurality of historical information data, and generating a new first intermediate information classification model; until iterating to the last module in the intermediate information classification model, a target information classification model is generated.
In one embodiment, the method comprises the steps of dividing each layer of network in the intermediate information classification model according to a preset classification standard to generate a plurality of modules, and executing the computer program by the processor to further realize the following steps:
determining the target layer number of the neural network in each module based on the total layer number of the neural network in the intermediate information classification model;
and dividing the neural network of each layer in the intermediate information classification model according to the target layer number of the neural network in each module to generate a plurality of modules.
In one embodiment, the first module of the plurality of modules is simplified to generate a simplified intermediate information classification model, and the computer program when executed by the processor further performs the steps of:
selecting a target neural network from the neural networks contained in the first module aiming at the first module in the plurality of modules;
replacing the first module by using a target neural network to generate a simplified first module;
based on the simplified first module and other modules in the intermediate information classification model, a simplified intermediate information classification model is generated.
In one embodiment, training the initial information classification model based on a plurality of historical information data to generate an intermediate information classification model, the computer program when executed by the processor further performs the steps of:
Inputting the historical information content into an initial information classification model for processing aiming at each historical information data to generate a prediction classification result corresponding to the historical information content;
training the initial information classification model according to the prediction classification result corresponding to each historical information content and the labeling classification result corresponding to each historical information content to generate an intermediate information classification model.
In one embodiment, a computer program product is provided comprising a computer program which, when executed by a processor, performs the steps of:
acquiring information data to be classified;
inputting the information data to be classified into a target information classification model for information classification, and generating an information classification result of the information data to be classified; the target information classification model is generated by training and simplifying the initial information classification model based on the historical information data;
outputting the information classification result of the information data to be classified.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring a plurality of historical information data; each history information data comprises a label classification result corresponding to the history information content;
Training the initial information classification model according to the plurality of historical information data to generate an intermediate information classification model;
simplifying and training the intermediate information classification model according to the plurality of historical information data to generate the target information classification model.
In one embodiment, the intermediate information classification model is simplified and trained based on a plurality of historical information data to generate a target information classification model, and the computer program when executed by the processor further performs the steps of:
dividing each layer of neural network in the intermediate information classification model according to a preset classification standard to generate a plurality of modules; the module comprises at least one layer of neural network;
simplifying a first module in the plurality of modules to generate a simplified intermediate information classification model, training the simplified intermediate information classification model according to a plurality of historical information data to generate a first intermediate information classification model;
taking the next module of the first modules in the first intermediate information classification model as a new first module, circularly executing the new first modules in the intermediate information classification model to simplify the new intermediate information classification model, generating a new simplified intermediate information classification model, training the new simplified intermediate information classification model according to a plurality of historical information data, and generating a new first intermediate information classification model; until iterating to the last module in the intermediate information classification model, a target information classification model is generated.
In one embodiment, the method comprises the steps of dividing each layer of network in the intermediate information classification model according to a preset classification standard to generate a plurality of modules, and executing the computer program by the processor to further realize the following steps:
determining the target layer number of the neural network in each module based on the total layer number of the neural network in the intermediate information classification model;
and dividing the neural network of each layer in the intermediate information classification model according to the target layer number of the neural network in each module to generate a plurality of modules.
In one embodiment, the first module of the plurality of modules is simplified to generate a simplified intermediate information classification model, and the computer program when executed by the processor further performs the steps of:
selecting a target neural network from the neural networks contained in the first module aiming at the first module in the plurality of modules;
replacing the first module by using a target neural network to generate a simplified first module;
based on the simplified first module and other modules in the intermediate information classification model, a simplified intermediate information classification model is generated.
In one embodiment, training the initial information classification model based on a plurality of historical information data to generate an intermediate information classification model, the computer program when executed by the processor further performs the steps of:
Inputting the historical information content into an initial information classification model for processing aiming at each historical information data to generate a prediction classification result corresponding to the historical information content;
training the initial information classification model according to the prediction classification result corresponding to each historical information content and the labeling classification result corresponding to each historical information content to generate an intermediate information classification model.
It should be noted that, the user information (including but not limited to user equipment information, user personal information, etc.) and the data (including but not limited to data for analysis, stored data, presented data, etc.) related to the present application are information and data authorized by the user or sufficiently authorized by each party, and the collection, use and processing of the related data need to comply with the related laws and regulations and standards of the related country and region.
Those skilled in the art will appreciate that implementing all or part of the above-described methods in accordance with the embodiments may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magnetic random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (Phase Change Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like. The databases referred to in the embodiments provided herein may include at least one of a relational database and a non-relational database. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processor referred to in the embodiments provided in the present application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, or the like, but is not limited thereto.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The foregoing examples illustrate only a few embodiments of the application, which are described in detail and are not to be construed as limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of the application should be assessed as that of the appended claims.

Claims (10)

1. An information classification method, the method comprising:
acquiring information data to be classified;
inputting the information data to be classified into a target information classification model to perform information classification, and generating an information classification result of the information data to be classified; the target information classification model is generated by training and simplifying an initial information classification model based on historical information data;
And outputting the information classification result of the information data to be classified.
2. The method according to claim 1, wherein the method further comprises:
acquiring a plurality of historical information data; each history information data comprises history information content and a labeling classification result corresponding to the history information content;
training the initial information classification model according to a plurality of the historical information data to generate an intermediate information classification model;
simplifying and training the intermediate information classification model according to a plurality of the historical information data to generate the target information classification model.
3. The method of claim 2, wherein said simplifying and training said intermediate information classification model based on a plurality of said historical information data to generate said target information classification model comprises:
dividing each layer of neural network in the intermediate information classification model according to a preset classification standard to generate a plurality of modules; the module includes at least one layer of neural network;
simplifying a first module in the plurality of modules to generate a simplified intermediate information classification model, and training the simplified intermediate information classification model according to a plurality of historical information data to generate a first intermediate information classification model;
Taking the next module of the first modules in the first intermediate information classification model as a new first module, circularly executing the simplification of the new first modules in the intermediate information classification model, generating a new simplified intermediate information classification model, training the new simplified intermediate information classification model according to a plurality of historical information data, and generating a new first intermediate information classification model; until iterating to the last module in the intermediate information classification model, generating the target information classification model.
4. The method of claim 3, wherein the dividing each layer of network in the intermediate information classification model according to a preset classification criterion generates a plurality of modules, including:
determining the target layer number of the neural network in each module based on the total layer number of the neural network in the intermediate information classification model;
and dividing the neural network of each layer in the intermediate information classification model according to the target layer number of the neural network in each module to generate a plurality of modules.
5. The method of claim 3 or 4, wherein the simplifying a first module of the plurality of modules to generate a simplified intermediate information classification model comprises:
Selecting a target neural network from the neural networks contained in a first module in the plurality of modules;
replacing the first module by the target neural network to generate a simplified first module;
and generating the simplified intermediate information classification model based on the first simplified module and other modules in the intermediate information classification model.
6. The method of any of claims 2-4, wherein training an initial information classification model based on a plurality of the historical information data to generate the intermediate information classification model comprises:
inputting the historical information content into the initial information classification model for processing aiming at each historical information data to generate a prediction classification result corresponding to the historical information content;
training the initial information classification model according to the prediction classification result corresponding to each historical information content and the labeling classification result corresponding to each historical information content to generate the intermediate information classification model.
7. An information classification apparatus, the apparatus comprising:
the acquisition module is used for acquiring information data to be classified;
The information classification module is used for inputting the information data to be classified into a target information classification model to perform information classification and generating an information classification result of the information data to be classified; the target information classification model is generated by training and simplifying an initial information classification model based on historical information data;
and the output module is used for outputting the information classification result of the information data to be classified.
8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 6 when the computer program is executed.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.
10. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.
CN202310686902.4A 2023-06-09 2023-06-09 Information classification method, apparatus, computer device, storage medium, and program product Pending CN116881450A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310686902.4A CN116881450A (en) 2023-06-09 2023-06-09 Information classification method, apparatus, computer device, storage medium, and program product

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310686902.4A CN116881450A (en) 2023-06-09 2023-06-09 Information classification method, apparatus, computer device, storage medium, and program product

Publications (1)

Publication Number Publication Date
CN116881450A true CN116881450A (en) 2023-10-13

Family

ID=88265118

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310686902.4A Pending CN116881450A (en) 2023-06-09 2023-06-09 Information classification method, apparatus, computer device, storage medium, and program product

Country Status (1)

Country Link
CN (1) CN116881450A (en)

Similar Documents

Publication Publication Date Title
US11922308B2 (en) Generating neighborhood convolutions within a large network
CN107545889B (en) Model optimization method and device suitable for pattern recognition and terminal equipment
CN114297509A (en) User interest vector generation method, product recommendation method, device, equipment and medium
US20200074277A1 (en) Fuzzy input for autoencoders
CN113901395A (en) Data processing method, data processing device, computer equipment and storage medium
CN117131222A (en) Semi-automatic labeling method and device based on open world large model
CN116738983A (en) Word embedding method, device and equipment for performing financial field task processing by model
CN116108697A (en) Acceleration test data processing method, device and equipment based on multiple performance degradation
CN116881450A (en) Information classification method, apparatus, computer device, storage medium, and program product
CN112085040A (en) Object tag determination method and device and computer equipment
CN116881543A (en) Financial resource object recommendation method, device, equipment, storage medium and product
CN116611506B (en) User analysis model training method, user label determining method and device
CN117235584B (en) Picture data classification method, device, electronic device and storage medium
CN118133044A (en) Problem extension method, device, computer equipment, storage medium and product
CN117435279A (en) Personal resource management method, device, computer equipment and storage medium
CN117033837A (en) Historical browsing data storage method, device, computer equipment and storage medium
WO2022037756A1 (en) Data processing apparatus and method for operating multi-output neural networks
CN116485096A (en) Resource processing method, device, computer equipment and storage medium
CN116881122A (en) Test case generation method, device, equipment, storage medium and program product
CN117665972A (en) Weather information prediction method, device, server and storage medium
CN116980471A (en) Content pushing method, device, apparatus, storage medium and computer program product
CN117376410A (en) Service pushing method, device, computer equipment and storage medium
CN116306985A (en) Predictive model training method, apparatus, computer device and storage medium
CN117114127A (en) Model training method, device, computer equipment and storage medium
CN117454956A (en) Training method of project recommendation model, project recommendation method and related equipment

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination