CN112507095A - Information identification method based on weak supervised learning and related equipment - Google Patents
Information identification method based on weak supervised learning and related equipment Download PDFInfo
- Publication number
- CN112507095A CN112507095A CN202011481937.7A CN202011481937A CN112507095A CN 112507095 A CN112507095 A CN 112507095A CN 202011481937 A CN202011481937 A CN 202011481937A CN 112507095 A CN112507095 A CN 112507095A
- Authority
- CN
- China
- Prior art keywords
- label
- government
- cnn
- entities
- model
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 124
- 230000006870 function Effects 0.000 claims description 61
- 239000013598 vector Substances 0.000 claims description 37
- 230000004913 activation Effects 0.000 claims description 19
- 238000012549 training Methods 0.000 claims description 18
- 238000004590 computer program Methods 0.000 claims description 12
- 238000003860 storage Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 5
- 238000011161 development Methods 0.000 abstract description 3
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 238000013528 artificial neural network Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 238000006243 chemical reaction Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000002372 labelling Methods 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 239000000575 pesticide Substances 0.000 description 3
- 238000009395 breeding Methods 0.000 description 2
- 230000001488 breeding effect Effects 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000011176 pooling Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- RWSOTUBLDIXVET-UHFFFAOYSA-N Dihydrogen sulfide Chemical compound S RWSOTUBLDIXVET-UHFFFAOYSA-N 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000008094 contradictory effect Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
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
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
-
- 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
- G06N3/045—Combinations of 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
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Tourism & Hospitality (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Educational Administration (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Development Economics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of artificial intelligence, and provides an information identification method based on weak supervised learning, which comprises the following steps: acquiring a plurality of frequently asked question answering (FAQ) data in a government scene, and constructing a first label data set based on the plurality of FAQ data; performing multi-label text classification on the first label data set by using a first Convolutional Neural Network (CNN) to obtain a multi-label CNN model; obtaining a plurality of second label data sets, and adjusting the multi-label CNN model by using the plurality of second label data sets to obtain a final multi-label classifier; receiving an input user question; inputting the user questions into a final multi-label classifier to obtain a plurality of government affair entities matched with the user questions; outputting a plurality of government entities. The invention also relates to a block chain technology, which can upload a plurality of government affair entities to the block chain. The method can be applied to the fields of intelligent government affairs, intelligent communities and the like which need information identification, and therefore the development of intelligent cities is promoted.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to an information identification method based on weak supervised learning and related equipment.
Background
At present, in a government scene, various consultation questions are often encountered, answers need to be provided, and the government question-answering robot plays a key role in question-answering. The government affair question and answer robot can assist government personnel to answer consultation questions of users, and the consultation questions generally correspond to relevant government affair entities, such as: registering individual industrial and commercial enterprises information, handling residence certificates, examining and approving pesticide advertisements, subsiding intellectual property rights, verifying and issuing production and operation licenses of breeding birds and the like.
However, in practice, it is found that government entities are generally specialized, and different users have different spoken names; in addition, the government entity information is hidden in the consultation problem expressed by the user, and it is difficult to accurately extract the government entity, so that the user experience is poor.
Therefore, how to effectively identify the government affair entity in the government affair scene is a technical problem to be solved urgently.
Disclosure of Invention
In view of the foregoing, there is a need for an information identification method and related apparatus based on weak supervised learning, which can effectively identify government affair entities in a government affair scene.
The first aspect of the present invention provides an information identification method based on weak supervised learning, including:
acquiring a plurality of frequently asked question answering (FAQ) data in a government scene, and constructing a first label data set based on the FAQ data;
performing multi-label text classification on the first label data set by using a first Convolutional Neural Network (CNN) to obtain a multi-label CNN model, wherein a full link layer in the first CNN corresponds to N activation functions, the N is the same as the dimension of an output label sequence, and the multi-label CNN model is a weakly supervised model;
obtaining a plurality of second label data sets, and adjusting the multi-label CNN model by using the plurality of second label data sets to obtain a final multi-label classifier;
receiving an input user question;
inputting the user questions into the final multi-label classifier to obtain a plurality of government affair entities matched with the user questions;
outputting the plurality of government entities.
In one possible implementation, the constructing the first tag data set based on the plurality of FAQ data includes:
scanning answer data in a plurality of FAQ data through a Chinese language model Ngram;
converting the answer data into a first vector according to a word vector model, and converting the government affair entities in a government affair entity library into a second vector according to the word vector model;
cosine similarity calculation is carried out on the first vector and the second vector to obtain a similarity score;
determining the government entity with the similarity score larger than a preset threshold value as a label matched with the FAQ data;
and constructing a first label data set according to each FAQ data and the matched label of the FAQ data.
In one possible implementation manner, in the first tag data set, each question corresponds to one tag list, each tag in the tag list corresponds to one government entity, and in the tag list, a tag of a government entity matching the question is set as a first identifier, and a tag of a government entity not matching the question is set as a second identifier.
In a possible implementation manner, the performing, by using a first convolutional neural network CNN, multi-label text classification on the first label data set to obtain a multi-label CNN model includes:
performing two-class training on each label in the first label data set by using a first Convolutional Neural Network (CNN) and adopting N sigmoid activation functions;
calculating a loss function of the label of each dimension, and calculating an average loss function of the labels in all dimensions;
determining the average loss function as a first loss function for the first tag data set;
and adjusting the model parameters of the first CNN to minimize the first loss function to obtain a multi-label CNN model.
In a possible implementation manner, the adjusting the multi-label CNN model by using the plurality of second label data sets to obtain a final multi-label classifier includes:
acquiring historical model parameters of the multi-label CNN model;
performing iterative training on the historical model parameters of the multi-label CNN model by adopting the plurality of second label data sets, and calculating a second loss function, wherein the second loss function is an average loss function of labels in all dimensions;
and adjusting the historical model parameters to minimize the second loss function to obtain the final multi-label classifier.
In one possible implementation, the outputting the plurality of government entities comprises:
acquiring a historical consultation field of a current user;
calculating a first matching degree of each government entity and the historical consultation field;
acquiring a second matching degree of each government entity and the user questions;
weighting the first matching degree and the second matching degree to obtain a weighted value of each government entity;
and outputting the plurality of government entities according to the order of the weighted values from high to low.
In a possible implementation manner, the information identification method based on weak supervised learning further includes:
judging whether a selection instruction of a current user for the plurality of government affair entities is received within preset time;
if the selection instruction of the current user for the plurality of government affair entities is not received within the preset time, obtaining the historical question-answer record of the current user;
determining the consultation field of the current user from the historical question-answer records;
screening out target government affair entities matched with the consultation field from the plurality of government affair entities;
and outputting the detailed content of the target government entity.
A second aspect of the present invention provides an information identifying apparatus comprising:
the system comprises an acquisition module, a data processing module and a data processing module, wherein the acquisition module is used for acquiring a plurality of frequently asked question answers (FAQ) data in a government scene and constructing a first label data set based on the FAQ data;
a classification module, configured to perform multi-label text classification on the first label data set by using a first Convolutional Neural Network (CNN) to obtain a multi-label CNN model, where a full link layer in the first CNN corresponds to N activation functions, N is the same as a dimension of an output label sequence, and the multi-label CNN model is a weakly supervised model;
the obtaining module is further configured to obtain a plurality of second tag data sets;
an adjusting module, configured to adjust the multi-label CNN model using the plurality of second label data sets, to obtain a final multi-label classifier;
the receiving module is used for receiving input user questions;
an input module, configured to input the user question into the final multi-label classifier, and obtain a plurality of government affair entities matching the user question;
and the output module is used for outputting the plurality of government affair entities.
A third aspect of the present invention provides an electronic device comprising a processor and a memory, wherein the processor is configured to implement the weak supervised learning based information identification method when executing a computer program stored in the memory.
A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the weak supervised learning based information identification method.
According to the technical scheme, the method can be applied to the fields of intelligent government affairs, intelligent communities and the like which need information identification, and therefore the development of intelligent cities is promoted. In the invention, a first label data set is constructed, a government entity recognition task is converted into a multi-label classification task, a first Convolutional Neural Network (CNN) is utilized to perform multi-label text classification on the first label data set to obtain a multi-label CNN model, and a plurality of second label data sets are used to adjust the multi-label CNN model to obtain a final multi-label classifier, namely, a model training optimization process from weak supervision learning to accurate adjustment is realized, the workload of manual labeling is greatly reduced, and meanwhile, the efficiency of model training and the accuracy of the model for government entity recognition are improved.
Drawings
Fig. 1 is a flowchart of a preferred embodiment of an information identification method based on weak supervised learning according to the present invention.
FIG. 2 is a functional block diagram of an information recognition apparatus according to a preferred embodiment of the present disclosure.
Fig. 3 is a schematic structural diagram of an electronic device implementing an information identification method based on weakly supervised learning according to a preferred embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The terms "first" and "second" in the description and claims of the present application and the above-described drawings are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It will be appreciated that the data so used may be interchanged under appropriate circumstances such that the embodiments described herein may be practiced otherwise than as specifically illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
It should be noted that the description relating to "first", "second", etc. in the present invention is for descriptive purposes only and is not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In addition, technical solutions between various embodiments may be combined with each other, but must be realized by a person skilled in the art, and when the technical solutions are contradictory or cannot be realized, such a combination should not be considered to exist, and is not within the protection scope of the present invention.
The electronic device is a device capable of automatically performing numerical calculation and/or information processing according to preset or stored instructions, and the hardware thereof includes, but is not limited to, a microprocessor, an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), a Digital Signal Processor (DSP), an embedded device, and the like. The electronic device may also include a network device and/or a user device. The network device includes, but is not limited to, a single network server, a server group consisting of a plurality of network servers, or a Cloud Computing (Cloud Computing) based Cloud consisting of a large number of hosts or network servers. The user device includes, but is not limited to, any electronic product that can interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device, for example, a personal computer, a tablet computer, a smart phone, a Personal Digital Assistant (PDA), or the like.
Referring to fig. 1, fig. 1 is a flowchart illustrating an information identification method based on weakly supervised learning according to a preferred embodiment of the present invention. The order of the steps in the flowchart may be changed, and some steps may be omitted.
S11, obtaining a plurality of frequently asked question answering (FAQ) data in a government scene, and constructing a first label data set based on the FAQ data.
Where, FAQ is an abbreviation for english freqently ask Questions, chinese means "Frequently Asked Questions", or more colloquially "Frequently Asked Questions answers".
In the government affair scene, a large amount of FAQ data are usually available, and the data mainly have several sources, 1, a common question bank of government affair customer service personnel, and 2, real questions of users. Such data have the following characteristics: 1. no manual marking is performed; 2. the question expression is diversified and spoken, while the answer is generally given by professional personnel, and the expression is relatively standardized and specialized; 3. typically for a particular government entity.
Wherein the first label data set is the weakly supervised label data set.
The government entities such as individual industrial and commercial enterprises register information, transact residence certificates, examine and approve pesticide advertisements, fund intellectual property rights, verify and issue operation licenses for production of breeders and the like.
Specifically, the constructing a first tag data set based on the plurality of FAQ data includes:
scanning answer data in a plurality of FAQ data through a Chinese language model Ngram;
converting the answer data into a first vector according to a word vector model, and converting the government affair entities in a government affair entity library into a second vector according to the word vector model;
cosine similarity calculation is carried out on the first vector and the second vector to obtain a similarity score;
determining the government entity with the similarity score larger than a preset threshold value as a label matched with the FAQ data;
and constructing a first label data set according to each FAQ data and the matched label of the FAQ data.
Among them, N-Gram is a common Language Model in large vocabulary continuous speech recognition, called Chinese Language Model (CLM). The Chinese language model can realize automatic conversion to Chinese characters by using collocation information between adjacent words in the context.
Where a Word vector model such as Word2vec, Word2vec is a group of related models used to generate a Word vector. The word2vec model may be used to map each word to a vector, which may be used to represent word-to-word relationships, the vector being a hidden layer of a neural network.
In the first tag data set, each question corresponds to a tag list, each tag in the tag list corresponds to a government entity, and in the tag list, the tag of the government entity matched with the question is set as a first identifier, and the tag of the government entity unmatched with the question is set as a second identifier.
Specifically, for example, the tag list may be composed of numbers having a value of 1 (i.e., the first identifier) or 0 (i.e., the second identifier), and the length of the tag list is the number of government entities. The government affair entity corresponding to the label with the value of 1 is a matched government affair entity, and the government affair entity corresponding to the label with the value of 0 is an unmatched government affair entity.
S12, using a first Convolutional Neural Network (CNN) to perform multi-label text classification on the first label data set to obtain a multi-label CNN model, wherein a full link layer in the first CNN corresponds to N activation functions, N is the same as the dimension of an output label sequence, and the multi-label CNN model is a weakly supervised model.
Among them, Convolutional Neural Networks (CNN) are a kind of feed forward Neural Networks (fed forward Neural Networks) that contain convolution computation and have a deep structure. Structure of conventional CNN: (convolutional layer + nonlinear activation function (Relu or tanh) + pooling layer) x n + multiple fully connected layers.
In the invention, the first CNN sets the last full link layer in the traditional CNN as the corresponding label length, adds one sigmoid activation function to each dimension, and totally records N sigmoid activation functions, wherein each sigmoid activation function corresponds to one dimension in a label list.
Specifically, the using the first convolutional neural network CNN to perform multi-label text classification on the first label data set to obtain a multi-label CNN model includes:
performing two-class training on each label in the first label data set by using a first Convolutional Neural Network (CNN) and adopting N sigmoid activation functions;
calculating a loss function of the label of each dimension, and calculating an average loss function of the labels in all dimensions;
determining the average loss function as a first loss function for the first tag data set;
and adjusting the model parameters of the first CNN to minimize the first loss function to obtain a multi-label CNN model.
The first label data set is trained by using the first convolutional neural network CNN, so that a weakly supervised model, namely a multi-label CNN model, can be obtained, and the multi-label CNN model belongs to a weakly supervised multi-label classifier and can roughly identify a government entity (namely the government entity).
And S13, obtaining a plurality of second label data sets, and adjusting the multi-label CNN model by using the plurality of second label data sets to obtain the final multi-label classifier.
Wherein each of the second tag data sets is an accurate set of question and answer data labeled manually, i.e., each question is configured with an accurate government entity and a tag.
Specifically, the adjusting the multi-label CNN model using the plurality of second label data sets to obtain the final multi-label classifier includes:
acquiring historical model parameters of the multi-label CNN model;
performing iterative training on the historical model parameters of the multi-label CNN model by adopting the plurality of second label data sets, and calculating a second loss function, wherein the second loss function is an average loss function of labels in all dimensions;
and adjusting the historical model parameters to minimize the second loss function to obtain the final multi-label classifier.
The multi-label CNN model is adjusted through a plurality of second label data sets which are labeled manually, so that the historical model parameters of the multi-label CNN model can be optimized, more accurate model parameters can be obtained, and the final multi-label classifier after iterative training is higher in identification precision and better in identification effect.
And S14, receiving the input user question.
Wherein the user question is a government affairs question.
And S15, inputting the user questions into the final multi-label classifier, and obtaining a plurality of government affair entities matched with the user questions.
Wherein the plurality of government entities are possible government entities with which the current user wants to consult.
And S16, outputting the plurality of government affair entities.
Specifically, the outputting the plurality of government affair entities includes:
acquiring a historical consultation field of a current user;
calculating a first matching degree of each government entity and the historical consultation field;
acquiring a second matching degree of each government entity and the user questions;
weighting the first matching degree and the second matching degree to obtain a weighted value of each government entity;
and outputting the plurality of government entities according to the order of the weighted values from high to low.
In this optional implementation, the consulting field of the same user is usually fixed, and each government entity and the historical consulting field may be subjected to conversion operation of word segmentation and sentence vectors, and after the vectors are obtained, cosine similarity calculation is performed, so that the first matching degree may be obtained. Among them, the conversion of the government affairs entity and the history consultation field into a vector belongs to the prior art and is not described in detail.
The first matching degree and the second matching degree can be preset with weights, and the weighted value of each government entity is calculated according to the weights, and the weighted value can be used for measuring the degree of conformity of the government entity with the user questions. A plurality of government entities are output according to the weighted values, so that the current user can more intuitively and preferentially see the government entities meeting the intention of the user without searching one by one, and the consultation satisfaction of the user can be improved.
Optionally, the method further includes:
judging whether a selection instruction of a current user for the plurality of government affair entities is received within preset time;
if the selection instruction of the current user for the plurality of government affair entities is not received within the preset time, obtaining the historical question-answer record of the current user;
determining the consultation field of the current user from the historical question-answer records;
screening out target government affair entities matched with the consultation field from the plurality of government affair entities;
and outputting the detailed content of the target government entity.
In the optional implementation mode, when a plurality of possible government affair entities are output, the current user is difficult to select due to personal professional level limitation, in this case, the consultation field to which the questions frequently asked by the user belong can be determined based on historical question and answer records, the range can be further reduced, and the target government affair entity matched with the consultation field can be screened out, so that more accurate service can be provided for the current user, and the consultation satisfaction degree of the user can be improved.
Optionally, the method further includes:
uploading the plurality of government entities to a blockchain.
Wherein, in order to ensure the privacy and the security of the data, the plurality of government affair entities can be uploaded to the block chain for storage.
In the method flow described in fig. 1, a first label data set is constructed, a government entity recognition task is converted into a multi-label classification task, an improved convolutional neural network CNN is used to perform multi-label text classification on the first label data set to obtain a multi-label CNN model, and a plurality of second label data sets labeled manually are used to adjust the multi-label CNN model to obtain a final multi-label classifier, so that a model training optimization process from weak supervision learning to accurate adjustment is realized, the workload of manual labeling is greatly reduced, and meanwhile, the efficiency of model training and the accuracy of the model for government entity recognition are improved.
According to the embodiment, the method and the system can be applied to the fields of intelligent government affairs, intelligent communities and the like which need information identification, and therefore development of intelligent cities is promoted. The above description is only a specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and it will be apparent to those skilled in the art that modifications may be made without departing from the inventive concept of the present invention, and these modifications are within the scope of the present invention.
Referring to fig. 2, fig. 2 is a functional block diagram of a preferred embodiment of an information recognition apparatus according to the present invention.
In some embodiments, the information identification device is run in an electronic device. The information identifying means may comprise a plurality of functional modules consisting of program code segments. Program codes of various program segments in the information identification device may be stored in the memory and executed by the at least one processor to perform part or all of the steps of the weak supervised learning based information identification method described in fig. 1.
In this embodiment, the information recognition apparatus may be divided into a plurality of functional modules according to the functions performed by the information recognition apparatus. The functional module may include: the system comprises an acquisition module 201, a classification module 202, an adjustment module 203, a receiving module 204, an input module 205 and an output module 206. The module referred to herein is a series of computer program segments capable of being executed by at least one processor and capable of performing a fixed function and is stored in memory.
The obtaining module 201 is configured to obtain multiple frequently asked question answers (FAQ) data in a government scene, and construct a first tag data set based on the multiple FAQ data.
Where, FAQ is an abbreviation for english freqently ask Questions, chinese means "Frequently Asked Questions", or more colloquially "Frequently Asked Questions answers".
In the government affair scene, a large amount of FAQ data are usually available, and the data mainly have several sources, 1, a common question bank of government affair customer service personnel, and 2, real questions of users. Such data have the following characteristics: 1. no manual marking is performed; 2. the question expression is diversified and spoken, while the answer is generally given by professional personnel, and the expression is relatively standardized and specialized; 3. typically for a particular government entity.
The government entity is a government entity, such as information registration of individual industrial and commercial enterprises, residence permit handling, pesticide advertisement approval, intellectual property right maintenance, verification and issue of operation licenses for breeding birds.
Specifically, the constructing a first tag data set based on the plurality of FAQ data includes:
scanning answer data in a plurality of FAQ data through a Chinese language model Ngram;
converting the answer data into a first vector according to a word vector model, and converting the government affair entities in a government affair entity library into a second vector according to the word vector model;
cosine similarity calculation is carried out on the first vector and the second vector to obtain a similarity score;
determining the government entity with the similarity score larger than a preset threshold value as a label matched with the FAQ data;
and constructing a first label data set according to each FAQ data and the matched label of the FAQ data.
Among them, N-Gram is a common Language Model in large vocabulary continuous speech recognition, called Chinese Language Model (CLM). The Chinese language model can realize automatic conversion to Chinese characters by using collocation information between adjacent words in the context.
Where a Word vector model such as Word2vec, Word2vec is a group of related models used to generate a Word vector. The word2vec model may be used to map each word to a vector, which may be used to represent word-to-word relationships, the vector being a hidden layer of a neural network.
In the first tag data set, each question corresponds to a tag list, each tag in the tag list corresponds to a government entity, and in the tag list, the tag of the government entity matched with the question is set as a first identifier, and the tag of the government entity unmatched with the question is set as a second identifier.
Specifically, for example, the tag list may be composed of numbers having a value of 1 (i.e., the first identifier) or 0 (i.e., the second identifier), and the length of the tag list is the number of government entities. The government affair entity corresponding to the label with the value of 1 is a matched government affair entity, and the government affair entity corresponding to the label with the value of 0 is an unmatched government affair entity.
A classification module 202, configured to perform multi-label text classification on the first label data set by using a first convolutional neural network CNN to obtain a multi-label CNN model, where a full link layer in the first CNN corresponds to N activation functions, N is the same as the dimension of an output label sequence, and the multi-label CNN model is a weakly supervised model.
Among them, Convolutional Neural Networks (CNN) are a kind of feed forward Neural Networks (fed forward Neural Networks) that contain convolution computation and have a deep structure. Structure of conventional CNN: (convolutional layer + nonlinear activation function (Relu or tanh) + pooling layer) x n + multiple fully connected layers.
In the invention, the first CNN sets the last full link layer in the traditional CNN as the corresponding label length, adds one sigmoid activation function to each dimension, and totally records N sigmoid activation functions, wherein each sigmoid activation function corresponds to one dimension in a label list.
Specifically, the using the first convolutional neural network CNN to perform multi-label text classification on the first label data set to obtain a multi-label CNN model includes:
performing two-class training on each label in the first label data set by using a first Convolutional Neural Network (CNN) and adopting N sigmoid activation functions;
calculating a loss function of the label of each dimension, and calculating an average loss function of the labels in all dimensions;
determining the average loss function as a first loss function for the first tag data set;
and adjusting the model parameters of the first CNN to minimize the first loss function to obtain a multi-label CNN model.
The first label data set is trained by using the first convolutional neural network CNN, so that a weakly supervised model, namely a multi-label CNN model, can be obtained, and the multi-label CNN model belongs to a weakly supervised multi-label classifier and can roughly identify a government entity (namely the government entity).
The obtaining module 201 is further configured to obtain a plurality of second tag data sets.
An adjusting module 203, configured to adjust the multi-label CNN model using the multiple second label data sets, so as to obtain a final multi-label classifier.
Wherein each of the second tag data sets is an accurate set of question and answer data labeled manually, i.e., each question is configured with an accurate government entity and a tag.
Specifically, the adjusting the multi-label CNN model using the plurality of second label data sets to obtain the final multi-label classifier includes:
acquiring historical model parameters of the multi-label CNN model;
performing iterative training on the historical model parameters of the multi-label CNN model by adopting the plurality of second label data sets, and calculating a second loss function, wherein the second loss function is an average loss function of labels in all dimensions;
and adjusting the historical model parameters to minimize the second loss function to obtain the final multi-label classifier.
The multi-label CNN model is adjusted through a plurality of second label data sets which are labeled manually, so that the historical model parameters of the multi-label CNN model can be optimized, more accurate model parameters can be obtained, and the final multi-label classifier after iterative training is higher in identification precision and better in identification effect.
A receiving module 204, configured to receive an input user question.
Wherein the user question is a government affairs question.
An input module 205, configured to input the user question into the final multi-label classifier, and obtain a plurality of government entities matching the user question.
Wherein the plurality of government entities are possible government entities with which the current user wants to consult.
An output module 206, configured to output the plurality of government entities.
Specifically, the outputting the plurality of government affair entities includes:
acquiring a historical consultation field of a current user;
calculating a first matching degree of each government entity and the historical consultation field;
acquiring a second matching degree of each government entity and the user questions;
weighting the first matching degree and the second matching degree to obtain a weighted value of each government entity;
and outputting the plurality of government entities according to the order of the weighted values from high to low.
In this optional implementation, the consulting field of the same user is usually fixed, and each government entity and the historical consulting field may be subjected to conversion operation of word segmentation and sentence vectors, and after the vectors are obtained, cosine similarity calculation is performed, so that the first matching degree may be obtained. The first matching degree and the second matching degree can be preset with weights, and the weighted value of each government entity is calculated according to the weights, and the weighted value can be used for measuring the degree of conformity of the government entity with the user questions. A plurality of government entities are output according to the weighted values, so that the current user can more intuitively and preferentially see the government entities meeting the intention of the user without searching one by one, and the consultation satisfaction of the user can be improved.
Optionally, the information identification apparatus further includes:
the judging module is used for judging whether a selection instruction of the current user for the plurality of government affair entities is received within preset time;
the obtaining module 201 is further configured to obtain a historical question and answer record of the current user if a selection instruction of the current user for the multiple government affair entities is not received within a preset time;
the determining module is used for determining the consultation field of the current user from the historical question-answer records;
the screening module is used for screening out target government affair entities matched with the consultation field from the plurality of government affair entities;
the output module 206 is configured to output details of the target government entity.
In the optional implementation mode, when a plurality of possible government affair entities are output, the current user is difficult to select due to personal professional level limitation, in this case, the consultation field to which the questions frequently asked by the user belong can be determined based on historical question and answer records, the range can be further reduced, and the target government affair entity matched with the consultation field can be screened out, so that more accurate service can be provided for the current user, and the consultation satisfaction degree of the user can be improved.
In the information recognition device described in fig. 2, a first label data set is constructed, a government entity recognition task is converted into a multi-label classification task, an improved convolutional neural network CNN is utilized to perform multi-label text classification on the first label data set, a multi-label CNN model is obtained, and a plurality of second label data sets labeled manually are used to adjust the multi-label CNN model, so that a final multi-label classifier is obtained, that is, a model training optimization process from weak supervision learning to accurate adjustment is realized, the workload of manual labeling is greatly reduced, and meanwhile, the efficiency of model training and the accuracy of the model for government entity recognition are improved.
Fig. 3 is a schematic structural diagram of an electronic device implementing an information identification method based on weakly supervised learning according to a preferred embodiment of the present invention. The electronic device 3 comprises a memory 31, at least one processor 32, a computer program 33 stored in the memory 31 and executable on the at least one processor 32, and at least one communication bus 34.
Those skilled in the art will appreciate that the schematic diagram shown in fig. 3 is merely an example of the electronic device 3, and does not constitute a limitation of the electronic device 3, and may include more or less components than those shown, or combine some components, or different components, for example, the electronic device 3 may further include an input/output device, a network access device, and the like.
The at least one Processor 32 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. The processor 32 may be a microprocessor or the processor 32 may be any conventional processor or the like, and the processor 32 is a control center of the electronic device 3 and connects various parts of the whole electronic device 3 by various interfaces and lines.
The memory 31 may be used to store the computer program 33 and/or the module/unit, and the processor 32 may implement various functions of the electronic device 3 by running or executing the computer program and/or the module/unit stored in the memory 31 and calling data stored in the memory 31. The memory 31 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required by at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area may store data (such as audio data) created according to the use of the electronic device 3, and the like. In addition, the memory 31 may include non-volatile and volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), at least one magnetic disk storage device, a Flash memory device, or other storage devices.
With reference to fig. 1, the memory 31 in the electronic device 3 stores a plurality of instructions to implement an information identification method based on weak supervised learning, and the processor 32 can execute the plurality of instructions to implement:
acquiring a plurality of frequently asked question answering (FAQ) data in a government scene, and constructing a first label data set based on the FAQ data;
performing multi-label text classification on the first label data set by using a first Convolutional Neural Network (CNN) to obtain a multi-label CNN model, wherein a full link layer in the first CNN corresponds to N activation functions, the N is the same as the dimension of an output label sequence, and the multi-label CNN model is a weakly supervised model;
obtaining a plurality of second label data sets, and adjusting the multi-label CNN model by using the plurality of second label data sets to obtain a final multi-label classifier;
receiving an input user question;
inputting the user questions into the final multi-label classifier to obtain a plurality of government affair entities matched with the user questions;
outputting the plurality of government entities.
Specifically, the processor 32 may refer to the description of the relevant steps in the embodiment corresponding to fig. 1 for a specific implementation method of the instruction, which is not described herein again.
In the electronic device 3 described in fig. 3, the government entity recognition task is converted into a multi-label classification task by constructing a first label data set, the first label data set is subjected to multi-label text classification by using an improved convolutional neural network CNN to obtain a multi-label CNN model, and the multi-label CNN model is adjusted by using a plurality of second label data sets labeled manually to obtain a final multi-label classifier, so that a model training optimization process from weak supervision learning to accurate adjustment is realized, the workload of manual labeling is greatly reduced, and meanwhile, the efficiency of model training and the accuracy of the model for government entity recognition are improved.
The integrated modules/units of the electronic device 3 may be stored in a computer-readable storage medium if they are implemented in the form of software functional units and sold or used as separate products. Based on such understanding, all or part of the flow of the method according to the embodiments of the present invention may also be implemented by a computer program, which may be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the method embodiments may be implemented. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying said computer program code, a recording medium, a usb-disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory.
In the embodiments provided in the present invention, it should be understood that the disclosed system, apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is only one logical functional division, and other divisions may be realized in practice.
The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.
In addition, functional modules in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional module.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference signs in the claims shall not be construed as limiting the claim concerned. A plurality of units or means recited in the present invention may also be implemented by software or hardware.
Finally, it should be noted that the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.
Claims (10)
1. An information identification method based on weak supervised learning is characterized in that the information identification method based on weak supervised learning comprises the following steps:
acquiring a plurality of frequently asked question answering (FAQ) data in a government scene, and constructing a first label data set based on the FAQ data;
performing multi-label text classification on the first label data set by using a first Convolutional Neural Network (CNN) to obtain a multi-label CNN model, wherein a full link layer in the first CNN corresponds to N activation functions, the N is the same as the dimension of an output label sequence, and the multi-label CNN model is a weakly supervised model;
obtaining a plurality of second label data sets, and adjusting the multi-label CNN model by using the plurality of second label data sets to obtain a final multi-label classifier;
receiving an input user question;
inputting the user questions into the final multi-label classifier to obtain a plurality of government affair entities matched with the user questions;
outputting the plurality of government entities.
2. The weak supervised learning based information identification method of claim 1, wherein the constructing a first label data set based on a plurality of FAQ data comprises:
scanning answer data in a plurality of FAQ data through a Chinese language model Ngram;
converting the answer data into a first vector according to a word vector model, and converting the government affair entities in a government affair entity library into a second vector according to the word vector model;
cosine similarity calculation is carried out on the first vector and the second vector to obtain a similarity score;
determining the government entity with the similarity score larger than a preset threshold value as a label matched with the FAQ data;
and constructing a first label data set according to each FAQ data and the matched label of the FAQ data.
3. The weak supervised learning based information identification method according to claim 2, wherein in the first label data set, each question corresponds to one label list, each label in the label list corresponds to one government entity, and in the label list, a label of the government entity matching the question is set as a first identifier, and a label of the government entity not matching the question is set as a second identifier.
4. The information identification method based on weak supervised learning as claimed in claim 1, wherein the performing multi-label text classification on the first label data set by using a first Convolutional Neural Network (CNN) to obtain a multi-label CNN model comprises:
performing two-class training on each label in the first label data set by using a first Convolutional Neural Network (CNN) and adopting N sigmoid activation functions;
calculating a loss function of the label of each dimension, and calculating an average loss function of the labels in all dimensions;
determining the average loss function as a first loss function for the first tag data set;
and adjusting the model parameters of the first CNN to minimize the first loss function to obtain a multi-label CNN model.
5. The weak supervised learning based information recognition method of claim 1, wherein the adjusting the multi-label CNN model by using the plurality of second label data sets to obtain a final multi-label classifier comprises:
acquiring historical model parameters of the multi-label CNN model;
performing iterative training on the historical model parameters of the multi-label CNN model by adopting the plurality of second label data sets, and calculating a second loss function, wherein the second loss function is an average loss function of labels in all dimensions;
and adjusting the historical model parameters to minimize the second loss function to obtain the final multi-label classifier.
6. The weak supervised learning based information identification method of claim 1, wherein the outputting the plurality of government entities comprises:
acquiring a historical consultation field of a current user;
calculating a first matching degree of each government entity and the historical consultation field;
acquiring a second matching degree of each government entity and the user questions;
weighting the first matching degree and the second matching degree to obtain a weighted value of each government entity;
and outputting the plurality of government entities according to the order of the weighted values from high to low.
7. The weak supervised learning based information identification method according to any one of claims 1 to 5, wherein the weak supervised learning based information identification method further comprises:
judging whether a selection instruction of a current user for the plurality of government affair entities is received within preset time;
if the selection instruction of the current user for the plurality of government affair entities is not received within the preset time, obtaining the historical question-answer record of the current user;
determining the consultation field of the current user from the historical question-answer records;
screening out target government affair entities matched with the consultation field from the plurality of government affair entities;
and outputting the detailed content of the target government entity.
8. An information recognition apparatus, characterized in that the information recognition apparatus comprises:
the system comprises an acquisition module, a data processing module and a data processing module, wherein the acquisition module is used for acquiring a plurality of frequently asked question answers (FAQ) data in a government scene and constructing a first label data set based on the FAQ data;
a classification module, configured to perform multi-label text classification on the first label data set by using a first Convolutional Neural Network (CNN) to obtain a multi-label CNN model, where a full link layer in the first CNN corresponds to N activation functions, N is the same as a dimension of an output label sequence, and the multi-label CNN model is a weakly supervised model;
the obtaining module is further configured to obtain a plurality of second tag data sets;
an adjusting module, configured to adjust the multi-label CNN model using the plurality of second label data sets, to obtain a final multi-label classifier;
the receiving module is used for receiving input user questions;
an input module, configured to input the user question into the final multi-label classifier, and obtain a plurality of government affair entities matching the user question;
and the output module is used for outputting the plurality of government affair entities.
9. An electronic device, characterized in that the electronic device comprises a processor and a memory, the processor is used for executing a computer program stored in the memory to realize the information identification method based on weak supervised learning as claimed in any one of claims 1 to 7.
10. A computer-readable storage medium storing at least one instruction, which when executed by a processor, implements a weak supervised learning based information identification method as recited in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011481937.7A CN112507095A (en) | 2020-12-15 | 2020-12-15 | Information identification method based on weak supervised learning and related equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011481937.7A CN112507095A (en) | 2020-12-15 | 2020-12-15 | Information identification method based on weak supervised learning and related equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112507095A true CN112507095A (en) | 2021-03-16 |
Family
ID=74972297
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011481937.7A Pending CN112507095A (en) | 2020-12-15 | 2020-12-15 | Information identification method based on weak supervised learning and related equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112507095A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113297360A (en) * | 2021-04-29 | 2021-08-24 | 天津汇智星源信息技术有限公司 | Law question-answering method and device based on weak supervised learning and joint learning mechanism |
| CN113360657A (en) * | 2021-06-30 | 2021-09-07 | 安徽商信政通信息技术股份有限公司 | Intelligent document distribution and handling method and device and computer equipment |
| CN115982208A (en) * | 2022-08-24 | 2023-04-18 | 淮阴工学院 | Cold chain product relevance query method and device based on block chain cross-chain cooperation |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080288292A1 (en) * | 2007-05-15 | 2008-11-20 | Siemens Medical Solutions Usa, Inc. | System and Method for Large Scale Code Classification for Medical Patient Records |
| CN101529373A (en) * | 2006-09-06 | 2009-09-09 | 奥多比公司 | System and method of determining and recommending a document control policy for a document |
| US20100185569A1 (en) * | 2009-01-19 | 2010-07-22 | Microsoft Corporation | Smart Attribute Classification (SAC) for Online Reviews |
| CN109308319A (en) * | 2018-08-21 | 2019-02-05 | 深圳中兴网信科技有限公司 | File classification method, document sorting apparatus and computer readable storage medium |
| WO2019099226A1 (en) * | 2017-11-14 | 2019-05-23 | Google Llc | Weakly-supervised action localization by sparse temporal pooling network |
| US20190325259A1 (en) * | 2018-04-12 | 2019-10-24 | Discovery Communications, Llc | Feature extraction and machine learning for automated metadata analysis |
| CN111275107A (en) * | 2020-01-20 | 2020-06-12 | 西安奥卡云数据科技有限公司 | Multi-label scene image classification method and device based on transfer learning |
| CN111737499A (en) * | 2020-07-27 | 2020-10-02 | 平安国际智慧城市科技股份有限公司 | Data searching method based on natural language processing and related equipment |
| CN112000801A (en) * | 2020-07-09 | 2020-11-27 | 山东师范大学 | Government affair text classification and hot spot problem mining method and system based on machine learning |
-
2020
- 2020-12-15 CN CN202011481937.7A patent/CN112507095A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101529373A (en) * | 2006-09-06 | 2009-09-09 | 奥多比公司 | System and method of determining and recommending a document control policy for a document |
| US20080288292A1 (en) * | 2007-05-15 | 2008-11-20 | Siemens Medical Solutions Usa, Inc. | System and Method for Large Scale Code Classification for Medical Patient Records |
| US20100185569A1 (en) * | 2009-01-19 | 2010-07-22 | Microsoft Corporation | Smart Attribute Classification (SAC) for Online Reviews |
| WO2019099226A1 (en) * | 2017-11-14 | 2019-05-23 | Google Llc | Weakly-supervised action localization by sparse temporal pooling network |
| CN110832499A (en) * | 2017-11-14 | 2020-02-21 | 谷歌有限责任公司 | Weak supervision action localization over sparse time pooling networks |
| US20190325259A1 (en) * | 2018-04-12 | 2019-10-24 | Discovery Communications, Llc | Feature extraction and machine learning for automated metadata analysis |
| CN109308319A (en) * | 2018-08-21 | 2019-02-05 | 深圳中兴网信科技有限公司 | File classification method, document sorting apparatus and computer readable storage medium |
| CN111275107A (en) * | 2020-01-20 | 2020-06-12 | 西安奥卡云数据科技有限公司 | Multi-label scene image classification method and device based on transfer learning |
| CN112000801A (en) * | 2020-07-09 | 2020-11-27 | 山东师范大学 | Government affair text classification and hot spot problem mining method and system based on machine learning |
| CN111737499A (en) * | 2020-07-27 | 2020-10-02 | 平安国际智慧城市科技股份有限公司 | Data searching method based on natural language processing and related equipment |
Non-Patent Citations (3)
| Title |
|---|
| 刘振宇;李钦富;杨硕;邓应强;刘芬;赖新明;白雪珂;: "一种基于主动学习和多种监督学习的情感分析模型", 中国电子科学研究院学报, no. 02 * |
| 孙振华;***;: "基于卷积神经网络的多标签飞机识别算法", 计算机应用与软件, no. 09 * |
| 孙松涛;何炎祥;: "基于CNN特征空间的微博多标签情感分类", 工程科学与技术, no. 03 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113297360A (en) * | 2021-04-29 | 2021-08-24 | 天津汇智星源信息技术有限公司 | Law question-answering method and device based on weak supervised learning and joint learning mechanism |
| CN113360657A (en) * | 2021-06-30 | 2021-09-07 | 安徽商信政通信息技术股份有限公司 | Intelligent document distribution and handling method and device and computer equipment |
| CN113360657B (en) * | 2021-06-30 | 2023-10-24 | 安徽商信政通信息技术股份有限公司 | Intelligent document distribution handling method and device and computer equipment |
| CN115982208A (en) * | 2022-08-24 | 2023-04-18 | 淮阴工学院 | Cold chain product relevance query method and device based on block chain cross-chain cooperation |
| CN115982208B (en) * | 2022-08-24 | 2023-09-29 | 淮阴工学院 | Cold chain product relevance query method and device based on block chain cross-chain collaboration |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111695033B (en) | Enterprise public opinion analysis method, enterprise public opinion analysis device, electronic equipment and medium | |
| CN110163478B (en) | Risk examination method and device for contract clauses | |
| CN109522553B (en) | Named entity identification method and device | |
| CN112346567B (en) | Virtual interaction model generation method and device based on AI (Artificial Intelligence) and computer equipment | |
| CN110222167B (en) | Method and system for acquiring target standard information | |
| CN112507095A (en) | Information identification method based on weak supervised learning and related equipment | |
| CN107679234A (en) | Customer service information providing method, device, electronic equipment, storage medium | |
| US20230169100A1 (en) | Method and apparatus for information acquisition, electronic device, and computer-readable storage medium | |
| CN109086265B (en) | Semantic training method and multi-semantic word disambiguation method in short text | |
| CN110929524A (en) | Data screening method, device, equipment and computer readable storage medium | |
| CN110704586A (en) | Information processing method and system | |
| CN112036184A (en) | Entity identification method, device, computer device and storage medium based on BilSTM network model and CRF model | |
| CN114037545A (en) | Client recommendation method, device, equipment and storage medium | |
| CN113112282A (en) | Method, device, equipment and medium for processing consult problem based on client portrait | |
| CN113919437A (en) | Method, device, equipment and storage medium for generating client portrait | |
| CN114626380A (en) | Entity identification method and device, electronic equipment and storage medium | |
| CN111242710A (en) | Business classification processing method and device, service platform and storage medium | |
| CN116204714A (en) | Recommendation method, recommendation device, electronic equipment and storage medium | |
| CN114511393B (en) | Financial data processing method and system | |
| CN113627194B (en) | Information extraction method and device, and communication message classification method and device | |
| CN116402166B (en) | Training method and device of prediction model, electronic equipment and storage medium | |
| CN117149969A (en) | Answer acquisition method and device based on intelligent question-answer model and electronic equipment | |
| CN112199958A (en) | Concept word sequence generation method and device, computer equipment and storage medium | |
| CN115358817A (en) | Intelligent product recommendation method, device, equipment and medium based on social data | |
| CN115905472A (en) | Business opportunity service processing method, business opportunity service processing device, business opportunity service processing server and computer readable storage medium |
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 |