CN113704429A - Semi-supervised learning-based intention identification method, device, equipment and medium - Google Patents
Semi-supervised learning-based intention identification method, device, equipment and medium Download PDFInfo
- Publication number
- CN113704429A CN113704429A CN202111009255.0A CN202111009255A CN113704429A CN 113704429 A CN113704429 A CN 113704429A CN 202111009255 A CN202111009255 A CN 202111009255A CN 113704429 A CN113704429 A CN 113704429A
- Authority
- CN
- China
- Prior art keywords
- intention
- training
- entity
- semi
- 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 70
- 238000012549 training Methods 0.000 claims abstract description 295
- 238000002372 labelling Methods 0.000 claims abstract description 77
- 238000012545 processing Methods 0.000 claims abstract description 32
- 238000012795 verification Methods 0.000 claims description 32
- 239000013598 vector Substances 0.000 claims description 28
- 238000013528 artificial neural network Methods 0.000 claims description 21
- 238000002790 cross-validation Methods 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 11
- 238000012216 screening Methods 0.000 claims description 8
- 238000013507 mapping Methods 0.000 claims description 6
- 238000005728 strengthening Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 4
- 239000000126 substance Substances 0.000 claims 1
- 238000013473 artificial intelligence Methods 0.000 abstract description 10
- 238000005516 engineering process Methods 0.000 abstract description 10
- 238000003058 natural language processing Methods 0.000 description 9
- 238000004891 communication Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 7
- 238000007726 management method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 239000004973 liquid crystal related substance Substances 0.000 description 4
- 238000010200 validation analysis Methods 0.000 description 4
- 230000006399 behavior Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 239000003795 chemical substances by application Substances 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000006467 substitution reaction 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
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- 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
- 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
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
- G06F40/295—Named entity recognition
-
- 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
Abstract
The invention relates to an artificial intelligence digital medical technology, and discloses an intention identification method based on semi-supervised learning, which comprises the following steps: extracting an intention text, carrying out entity labeling and entity enhancement processing to obtain an intention entity set, carrying out label clustering on the intention entity set, obtaining a labeled training set by using the clustered intention labels, training to obtain an original intention recognition model, extracting an unlabeled training set, carrying out iterative training and cross training on the original intention recognition model by using the unlabeled training set and the labeled training set to obtain a standard intention recognition model, and outputting an intention recognition result of the text to be recognized by using the standard intention recognition model. Furthermore, the invention also relates to blockchain techniques, the intention text being storable in a node of the blockchain. The invention also provides an intention identification method and device based on semi-supervised learning, electronic equipment and a storage medium. The invention can solve the problem of low intention identification accuracy.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to an intention identification method and device based on semi-supervised learning, electronic equipment and a computer readable storage medium.
Background
With the development of artificial intelligence, AI customer service is widely used in various fields to perform user service, and intent recognition is often used as an important module in the AI customer service, so that the problem of low accuracy of intent recognition often exists.
In the prior art, a large amount of manually labeled intention identification data is often needed to ensure the accuracy of intention model training. However, due to the limitation of the manual labeling speed, a large amount of labeling data is not provided to support the model to quickly reach high accuracy, so that the intention identification accuracy is not high.
Disclosure of Invention
The invention provides an intention identification method, an intention identification device, intention identification equipment and a storage medium based on semi-supervised learning, and mainly aims to solve the problem of low intention identification accuracy.
In order to achieve the above object, the present invention provides an intention identification method based on semi-supervised learning, including:
extracting an intention text of a user from a conversation intention corpus, and carrying out entity labeling and entity enhancement processing on the intention text to obtain an intention entity set;
performing label clustering on the intention entities in the intention entity set to obtain intention categories;
obtaining a labeling training set by using the intention labels of the intention categories, and training a pre-constructed neural network by using the labeling training set to obtain an original intention recognition model;
extracting an unlabeled training set from the conversation intention corpus, and performing iterative training on the original intention recognition model by using the unlabeled training set and the labeled training set to obtain a semi-supervised training model;
and performing cross training on the semi-supervised training model by using a cross verification method to obtain a standard intention recognition model, and outputting an intention recognition result of the text to be recognized by using the standard intention recognition model.
Optionally, the performing entity labeling and entity enhancement processing on the intention text to obtain an intention entity set includes:
carrying out entity labeling processing on the intention text by using a preset sequence labeling method;
entity prediction is carried out on a plurality of entities contained in an entity labeling result by utilizing an entity recognition layer in a preset entity recognition model, and type prediction scores of the plurality of entities are obtained;
and screening the entities according to the type prediction scores and a preset entity limiting rule, performing entity strengthening processing on the screened entities, and summarizing all entities strengthened by the entities to obtain the intention entity set.
Optionally, the tag clustering of the intention entities in the intention entity set to obtain an intention category includes:
vector mapping is carried out on entities in the intention entity set by utilizing a preset natural language model, and an intention vector set is obtained;
randomly selecting a preset number of samples from the intention vector set as a clustering center;
sequentially calculating the distance from each sample in the intention vector set to the clustering center, and classifying each sample into the category corresponding to the clustering center with the minimum distance to obtain a plurality of category clusters;
and recalculating the clustering center of each category cluster, returning and calculating the distance from each sample in the intention vector set to the clustering center in sequence until the clustering centers of the plurality of category clusters converge, and determining the converged category cluster as the intention category.
Optionally, the calculating a cluster center of each category cluster includes:
calculating the cluster center of each category cluster by the following clustering formula:
wherein E isiIs the ith cluster center, CiIs the ith class cluster and x is the sample in the class cluster.
Optionally, the training a pre-constructed neural network with the labeled training set to obtain an original intention recognition model includes:
performing intention prediction on the text in the labeling training set by using the neural network to obtain a prediction result;
and calculating prediction accuracy according to the prediction result and the intention labels in the standard training set, and obtaining the original intention recognition model when the prediction accuracy is greater than or equal to a preset first accuracy threshold.
Optionally, the performing iterative training on the original intention recognition model by using the unlabeled training set and the labeled training set to obtain a semi-supervised training model, including:
outputting the prediction labels of the corpora in the unmarked training set and the prediction probabilities corresponding to the prediction labels by using the original intention recognition model;
selecting a prediction label with the prediction probability more than or equal to a preset first prediction threshold value and a corpus corresponding to the prediction label as supplementary labeling training data to be added to the labeling training set;
and predicting the intention label of each corpus in the supplemented labeled training set again by using the original intention recognition model, calculating the accuracy, returning to the step of extracting the unlabeled training set from the conversation intention corpus when the accuracy is not converged, and stopping iteration until the accuracy is converged to obtain the semi-supervised training model.
Optionally, the cross-training the semi-supervised training model by using a cross-validation method to obtain a standard intention recognition model, including:
dividing all the label training sets into label training sets with preset number;
randomly selecting a label training set as a verification set, training the semi-supervised training model by using the unselected label training set, calculating the accuracy of the semi-supervised training model, and obtaining an identification model when the accuracy is greater than or equal to a preset second accuracy threshold;
performing label prediction on the verification set by using the identification model to obtain a prediction label of which the prediction probability in the verification set is less than or equal to a preset second prediction threshold;
and returning to the step of randomly selecting one label training set as a verification set until all the label training sets are selected as the verification set, adjusting the prediction labels and the linguistic data corresponding to the prediction labels with the prediction probability less than or equal to the second prediction threshold value in all the verification sets, and training the semi-supervised training model by using the adjusted label training set to obtain the standard intention recognition model.
In order to solve the above problems, the present invention also provides an intention recognition apparatus based on semi-supervised learning, the apparatus including:
the intention entity extraction module is used for extracting an intention text of a user from the conversation intention corpus, and carrying out entity labeling and entity enhancement processing on the intention text to obtain an intention entity set;
the label clustering module is used for carrying out label clustering on the intention entities in the intention entity set to obtain intention categories;
the original model training module is used for obtaining a labeling training set by using the intention labels of the intention categories and training a pre-constructed neural network by using the labeling training set to obtain an original intention identification model;
the iterative training module is used for extracting an unlabeled training set from the conversation intention corpus and performing iterative training on the original intention recognition model by using the unlabeled training set and the labeled training set to obtain a semi-supervised training model;
and the intention recognition module is used for performing cross training on the semi-supervised training model by using a cross verification method to obtain a standard intention recognition model and outputting an intention recognition result of the text to be recognized by using the standard intention recognition model.
In order to solve the above problem, the present invention also provides an electronic device, including:
a memory storing at least one computer program; and
a processor executing the computer program stored in the memory to implement the semi-supervised learning based intention identifying method described above.
In order to solve the above problem, the present invention also provides a computer-readable storage medium having at least one computer program stored therein, the at least one computer program being executed by a processor in an electronic device to implement the semi-supervised learning based intention identifying method described above.
According to the method and the device, the entity labeling and the entity enhancement processing are carried out on the intention text, so that the useful information in the intention text can be focused more, and the accuracy of model training is improved. And the clustering-based intention labels of the intention categories are used for labeling, so that the difficulty of labeling the training data is greatly reduced, and meanwhile, the original intention recognition model is subjected to iterative training by using the unlabeled training set and the labeled training set, so that the precision of model training is improved. And the low-probability corpora generated in the model training can be continuously adjusted through a cross validation method, and cross training is performed by using the adjusted corpora, so that the accuracy of model identification is further improved. Therefore, the semi-supervised learning based intention identification method, the semi-supervised learning based intention identification device, the electronic equipment and the computer readable storage medium can solve the problem of low intention identification accuracy.
Drawings
Fig. 1 is a schematic flowchart of an intent recognition method based on semi-supervised learning according to an embodiment of the present invention;
FIG. 2 is a functional block diagram of an intent recognition apparatus based on semi-supervised learning according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of an electronic device implementing the semi-supervised learning based intention identification method according to an embodiment of the present invention.
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The embodiment of the application provides an intention identification method based on semi-supervised learning. The execution subject of the semi-supervised learning based intention identification method includes, but is not limited to, at least one of a server, a terminal and other electronic devices which can be configured to execute the method provided by the embodiment of the application. In other words, the semi-supervised learning based intention identification method may be performed by software or hardware installed in a terminal device or a server device, and the software may be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, and the like. The server may be an independent server, or may be a cloud server that provides basic cloud computing services such as a cloud service, a cloud database, cloud computing, a cloud function, cloud storage, a Network service, cloud communication, a middleware service, a domain name service, a security service, a Content Delivery Network (CDN), a big data and artificial intelligence platform, and the like.
Fig. 1 is a schematic flow chart of an intention identification method based on semi-supervised learning according to an embodiment of the present invention. In this embodiment, the semi-supervised learning based intention identification method includes:
s1, extracting the intention text of the user from the conversation intention corpus, and carrying out entity labeling and entity enhancement processing on the intention text to obtain an intention entity set;
in the embodiment of the invention, the conversation intention corpus records conversation texts based on the recording conversion of users and seat personnel in different links. For example, in the link of a payment will, the recording 1A includes the text content of the agent: "8888 Yuan your own loan, which is currently 1 day out of term and when you ask you to go back? ", the text content of the user: "i seem to have forgotten …". Since the text corpora of the seat and the user role are already included in the conversation intention corpus, the intention text corpora of the user can be directly extracted.
In one embodiment of the present invention, the corpus of session intents may be in any blockchain node. Further, the intention text may be an intention text in the digital medical field.
Specifically, the performing entity labeling and entity enhancement processing on the intention text to obtain an intention entity set includes:
carrying out entity labeling processing on the intention text by using a preset sequence labeling method;
entity prediction is carried out on a plurality of entities contained in an entity labeling result by utilizing an entity recognition layer in a preset entity recognition model, and type prediction scores of the plurality of entities are obtained;
and screening the entities according to the type prediction scores and a preset entity limiting rule, performing entity strengthening processing on the screened entities, and summarizing all entities strengthened by the entities to obtain the intention entity set.
In the embodiment of the invention, the model training is influenced because the text content of the user's intention corpus is longer and the available information is uncertain, and more effective user texts can be focused through entity marking and entity strengthening. The preset sequence marking method can be a BIO sequence marking mode, entity marking is carried out on the entities in the user behavior data, the words mentioned by the entities are marked as B or I, and non-entities are marked as O.
Optionally, the preset entity recognition model includes a Bi-LSTM (entity recognition) layer and a CRF (entity screening) layer. The Bi-LSTM (entity recognition) layer is used for carrying out entity recognition on the input intention text and outputting type prediction scores of all entities, namely prediction scores of different annotation types. And the CRF (entity screening) layer screens the entities corresponding to the prediction scores through a preset rule, so that the accuracy of the finally identified entities is higher. The preset entity restriction rule may be: the first word in the sentence always starts with the label "B-" or "O-" instead of the constraint "I-".
In an optional embodiment of the present invention, a preset natural language processing library may be used to perform entity enhancement, where the preset natural language processing library may be a HanLP natural language processing library, and a dependency syntax parsing tool in the HanLP natural language processing library is used to analyze a prefix of a current entity to perform entity enhancement on the current entity, where for example, the current entity is "cuke", the prefix is "CEO of apple company", and the enhanced entity is "CEO cuke of apple company".
In the embodiment of the invention, the precision of model training can be improved through entity marking and entity enhancement processing.
S2, performing label clustering on the intention entities in the intention entity set to obtain intention categories;
in the embodiment of the invention, because the intended text of the user is unsupervised corpora (namely, unmarked corpora), in the model training, because the unsupervised corpora are often huge in quantity, if the unsupervised corpora are too large in quantity, noise is added to influence the accuracy of the model, and if the unsupervised corpora are too small in quantity, the semi-supervised training effect can not be achieved.
Specifically, the tag clustering of the intention entities in the intention entity set to obtain an intention category includes:
vector mapping is carried out on entities in the intention entity set by utilizing a preset natural language model, and an intention vector set is obtained;
randomly selecting a preset number of samples from the intention vector set as a clustering center;
sequentially calculating the distance from each sample in the intention vector set to the clustering center, and classifying each sample into the category corresponding to the clustering center with the minimum distance to obtain a plurality of category clusters;
and recalculating the clustering center of each category cluster, returning and calculating the distance from each sample in the intention vector set to the clustering center in sequence until the clustering centers of the plurality of category clusters converge, and determining the converged category cluster as the intention category.
In the embodiment of the invention, for example, each sentence is mapped into an embedding vector of 768 dimensions by using a bert model for the entities in the intention entity set. The distance may be an euclidean distance, a manhattan distance, a chebyshev distance, or the like. Meanwhile, after vector mapping is performed on the entity, a preset number (for example, 10) of far intention vectors are selected from the mapped vector coordinate system as a clustering center.
For example, user intent statements in intent category 1 include: "i have returned", "not have returned", etc., the user intent statements in intent category 2 include: "re-like bar", "like a minus bar", etc.
In an optional embodiment of the present invention, the calculating a cluster center of each category cluster includes:
calculating the cluster center of each category cluster by the following clustering formula:
wherein E isiIs the ith cluster center, CiIs the ith class cluster and x is the sample in the class cluster.
S3, obtaining a labeling training set by using the intention labels of the intention categories, and training a pre-constructed neural network by using the labeling training set to obtain an original intention recognition model;
in this embodiment of the present invention, the intention labels of the intention categories may be manually defined by business personnel according to user intention statements in each intention category, for example, a willingness to repay link, and the user intention statements in the intention category 1 include: "i have returned", "not have returned", etc., the intention tag is defined as "guest called returned", and the user intention statements in intention category 2 include: "researcher", "waiving", etc., the intent tag is defined as "late in.
In an optional embodiment of the invention, through the intention categories after clustering, business personnel can complete label definition only by extracting a plurality of sentences in each category, thereby greatly saving the time for label definition. Meanwhile, user texts in the conversation intention corpus can be directly labeled according to the defined intention labels to obtain a labeling training set, so that the complexity of data labeling is reduced.
In detail, the training of the pre-constructed neural network by using the label training set to obtain an original intention recognition model includes:
performing intention prediction on the text in the labeling training set by using the neural network to obtain a prediction result;
and calculating prediction accuracy according to the prediction result and the intention labels in the standard training set, and obtaining the original intention recognition model when the prediction accuracy is greater than or equal to a preset first accuracy threshold.
In the embodiment of the present invention, the prediction accuracy is calculated by the following formula, including:
and the Accuracy is the Accuracy, T is the number of correctly predicted samples, and T is the total number of samples in the labeled training set.
In the embodiment of the invention, the pre-constructed neural network can be a deep neural network such as CNN, LSTM, RCNN, C-LSTM, FastText and the like. The marked supervised corpus is used for training, so that the accuracy of model training can be improved.
And S4, extracting an unlabeled training set from the conversation intention corpus, and performing iterative training on the original intention recognition model by using the unlabeled training set and the labeled training set to obtain a semi-supervised training model.
In the embodiment of the present invention, the semi-supervised training refers to a method for performing model training by combining a supervised corpus (corpus with artificial labels) and an unsupervised corpus (corpus without artificial labels). Meanwhile, a preset number of unlabeled user corpora can be extracted from the session intention corpus as unlabeled training sets, and the number of data in the unlabeled training sets is greater than or equal to that in the labeled training sets.
Specifically, the performing iterative training on the original intention recognition model by using the unlabeled training set and the labeled training set to obtain a semi-supervised training model includes:
outputting the prediction labels of the corpora in the unmarked training set and the prediction probabilities corresponding to the prediction labels by using the original intention recognition model;
selecting a prediction label with the prediction probability more than or equal to a preset first prediction threshold value and a corpus corresponding to the prediction label as supplementary labeling training data to be added to the labeling training set;
and predicting the intention label of each corpus in the supplemented labeled training set again by using the original intention recognition model, calculating the accuracy, returning to the step of extracting the unlabeled training set from the conversation intention corpus when the accuracy is not converged, and stopping iteration until the accuracy is converged to obtain the semi-supervised training model.
In an optional embodiment of the invention, a large amount of labeled data can be quickly obtained by a semi-supervised iterative training method with only a small amount of labeled data, so that the model training precision is improved.
And S5, performing cross training on the semi-supervised training model by using a cross verification method to obtain a standard intention recognition model, and outputting an intention recognition result of the text to be recognized by using the standard intention recognition model.
In the embodiment of the invention, although the semi-supervised training model and a large amount of labeled (namely labeled) corpora are obtained by a semi-supervised iteration method, labels with low probability or corpora with wrong labels exist in the corpora labeled by the model, and therefore, the labeling training set is adjusted by a cross-validation method, and the precision of model training can be further improved.
Specifically, the cross-training of the semi-supervised training model by using a cross-validation method to obtain a standard intention recognition model includes:
dividing all the label training sets into label training sets with preset number;
randomly selecting a label training set as a verification set, training the semi-supervised training model by using the unselected label training set, calculating the accuracy of the semi-supervised training model, and obtaining an identification model when the accuracy is greater than or equal to a preset second accuracy threshold;
performing label prediction on the verification set by using the identification model to obtain a prediction label of which the prediction probability in the verification set is less than or equal to a preset second prediction threshold;
and returning to the step of randomly selecting one label training set as a verification set until all the label training sets are selected as the verification set, adjusting the prediction labels and the linguistic data corresponding to the prediction labels with the prediction probability less than or equal to the second prediction threshold value in all the verification sets, and training the semi-supervised training model by using the adjusted label training set to obtain the standard intention recognition model.
In the embodiment of the invention, a 5-fold cross validation method can be used for cross training, namely, all labeling training sets are divided into 5 label training sets, 1 label training set is selected as a validation set each time, other unselected label training sets are used for training the semi-supervised training model, the identification model obtained by training is used for carrying out label prediction on the validation set, labels with lower prediction probability can be directly deleted or manually corrected, inaccurate labeling linguistic data can be quickly found out through 5-fold cross validation and adjusted, and the semi-supervised training model is retrained by using all adjusted linguistic data, so that the accuracy of identification of an intention identification model can be further improved.
In an optional embodiment of the present invention, for example, in a repayment link, the intention text "of the user a is input and i forget, and after a while, …" can directly output the intention label of the user as: "later deposit".
According to the method and the device, the entity labeling and the entity enhancement processing are carried out on the intention text, so that the useful information in the intention text can be focused more, and the accuracy of model training is improved. And the clustering-based intention labels of the intention categories are used for labeling, so that the difficulty of labeling the training data is greatly reduced, and meanwhile, the original intention recognition model is subjected to iterative training by using the unlabeled training set and the labeled training set, so that the precision of model training is improved. And the low-probability corpora generated in the model training can be continuously adjusted through a cross validation method, and cross training is performed by using the adjusted corpora, so that the accuracy of model identification is further improved. Therefore, the intention identification method based on semi-supervised learning provided by the invention can solve the problem of low accuracy of intention identification.
Fig. 2 is a functional block diagram of an intention recognition apparatus based on semi-supervised learning according to an embodiment of the present invention.
The semi-supervised learning based intention recognition apparatus 100 of the present invention may be installed in an electronic device. According to the implemented functions, the semi-supervised learning based intention recognition apparatus 100 may include an intention entity extraction module 101, a label clustering module 102, an original model training module 103, an iterative training module 104 and an intention recognition module 105. The module of the present invention, which may also be referred to as a unit, refers to a series of computer program segments that can be executed by a processor of an electronic device and that can perform a fixed function, and that are stored in a memory of the electronic device.
In the present embodiment, the functions regarding the respective modules/units are as follows:
the intention entity extraction module 101 is configured to extract an intention text of a user from a session intention corpus, and perform entity labeling and entity enhancement processing on the intention text to obtain an intention entity set;
the label clustering module 102 is configured to perform label clustering on the intention entities in the intention entity set to obtain intention categories;
the original model training module 103 is configured to obtain a labeling training set by using the intention labels of the intention categories, and train a pre-constructed neural network by using the labeling training set to obtain an original intention recognition model;
the iterative training module 104 is configured to extract an unlabeled training set from the session intention corpus, and perform iterative training on the original intention recognition model by using the unlabeled training set and the labeled training set to obtain a semi-supervised training model;
the intention recognition module 105 is configured to perform cross training on the semi-supervised training model by using a cross validation method to obtain a standard intention recognition model, and output an intention recognition result of the text to be recognized by using the standard intention recognition model.
In detail, the semi-supervised learning based intention recognition apparatus 100 has the following specific implementation of the modules:
extracting an intention text of a user from a conversation intention corpus, and carrying out entity labeling and entity enhancement processing on the intention text to obtain an intention entity set;
in the embodiment of the invention, the conversation intention corpus records conversation texts based on the recording conversion of users and seat personnel in different links. For example, in the link of a payment will, the recording 1A includes the text content of the agent: "8888 Yuan your own loan, which is currently 1 day out of term and when you ask you to go back? ", the text content of the user: "i seem to have forgotten …". Since the text corpora of the seat and the user role are already included in the conversation intention corpus, the intention text corpora of the user can be directly extracted.
Specifically, the performing entity labeling and entity enhancement processing on the intention text to obtain an intention entity set includes:
carrying out entity labeling processing on the intention text by using a preset sequence labeling method;
entity prediction is carried out on a plurality of entities contained in an entity labeling result by utilizing an entity recognition layer in a preset entity recognition model, and type prediction scores of the plurality of entities are obtained;
and screening the entities according to the type prediction scores and a preset entity limiting rule, performing entity strengthening processing on the screened entities, and summarizing all entities strengthened by the entities to obtain the intention entity set.
In the embodiment of the invention, the model training is influenced because the text content of the user's intention corpus is longer and the available information is uncertain, and more effective user texts can be focused through entity marking and entity strengthening. The preset sequence marking method can be a BIO sequence marking mode, entity marking is carried out on the entities in the user behavior data, the words mentioned by the entities are marked as B or I, and non-entities are marked as O.
Optionally, the preset entity recognition model includes a Bi-LSTM (entity recognition) layer and a CRF (entity screening) layer. The Bi-LSTM (entity recognition) layer is used for carrying out entity recognition on the input intention text and outputting type prediction scores of all entities, namely prediction scores of different annotation types. And the CRF (entity screening) layer screens the entities corresponding to the prediction scores through a preset rule, so that the accuracy of the finally identified entities is higher. The preset entity restriction rule may be: the first word in the sentence always starts with the label "B-" or "O-" instead of the constraint "I-".
In an optional embodiment of the present invention, a preset natural language processing library may be used to perform entity enhancement, where the preset natural language processing library may be a HanLP natural language processing library, and a dependency syntax parsing tool in the HanLP natural language processing library is used to analyze a prefix of a current entity to perform entity enhancement on the current entity, where for example, the current entity is "cuke", the prefix is "CEO of apple company", and the enhanced entity is "CEO cuke of apple company".
In the embodiment of the invention, the precision of model training can be improved through entity marking and entity enhancement processing.
Secondly, performing label clustering on the intention entities in the intention entity set to obtain intention categories;
in the embodiment of the invention, because the intended text of the user is unsupervised corpora (namely, unmarked corpora), in the model training, because the unsupervised corpora are often huge in quantity, if the unsupervised corpora are too large in quantity, noise is added to influence the accuracy of the model, and if the unsupervised corpora are too small in quantity, the semi-supervised training effect can not be achieved.
Specifically, the tag clustering of the intention entity set to obtain an intention category includes:
vector mapping is carried out on entities in the intention entity set by utilizing a preset natural language model, and an intention vector set is obtained;
randomly selecting a preset number of samples from the intention vector set as a clustering center;
sequentially calculating the distance from each sample in the intention vector set to the clustering center, and classifying each sample into the category corresponding to the clustering center with the minimum distance to obtain a plurality of category clusters;
and recalculating the clustering center of each category cluster, returning and calculating the distance from each sample in the intention vector set to the clustering center in sequence until the clustering centers of the plurality of category clusters converge, and determining the converged category cluster as the intention category.
In the embodiment of the invention, for example, each sentence is mapped into an embedding vector of 768 dimensions by using a bert model for the entities in the intention entity set. The distance may be an euclidean distance, a manhattan distance, a chebyshev distance, or the like. Meanwhile, after vector mapping is performed on the entity, a preset number (for example, 10) of far intention vectors are selected from the mapped vector coordinate system as a clustering center.
For example, user intent statements in intent category 1 include: "i have returned", "not have returned", etc., the user intent statements in intent category 2 include: "re-like bar", "like a minus bar", etc.
In an optional embodiment of the present invention, the calculating a cluster center of each category cluster includes:
calculating the cluster center of each category cluster by the following clustering formula:
wherein E isiIs the ith cluster center, CiIs the ith class cluster and x is the sample in the class cluster.
Thirdly, obtaining a labeling training set by using the intention labels of the intention categories, and training a pre-constructed neural network by using the labeling training set to obtain an original intention recognition model;
in this embodiment of the present invention, the intention labels of the intention categories may be manually defined by business personnel according to user intention statements in each intention category, for example, a willingness to repay link, and the user intention statements in the intention category 1 include: "i have returned", "not have returned", etc., the intention tag is defined as "guest called returned", and the user intention statements in intention category 2 include: "researcher", "waiving", etc., the intent tag is defined as "late in.
In an optional embodiment of the invention, through the intention categories after clustering, business personnel can complete label definition only by extracting a plurality of sentences in each category, thereby greatly saving the time for label definition. Meanwhile, user texts in the conversation intention corpus can be directly labeled according to the defined intention labels to obtain a labeling training set, so that the complexity of data labeling is reduced.
In detail, the training of the pre-constructed neural network by using the label training set to obtain an original intention recognition model includes:
performing intention prediction on the text in the labeling training set by using the neural network to obtain a prediction result;
and calculating prediction accuracy according to the prediction result and the intention labels in the standard training set, and obtaining the original intention recognition model when the prediction accuracy is greater than or equal to a preset first accuracy threshold.
In the embodiment of the present invention, the prediction accuracy is calculated by the following formula, including:
and the Accuracy is the Accuracy, T is the number of correctly predicted samples, and T is the total number of samples in the labeled training set.
In the embodiment of the invention, the pre-constructed neural network can be a deep neural network such as CNN, LSTM, RCNN, C-LSTM, FastText and the like. The marked supervised corpus is used for training, so that the accuracy of model training can be improved.
And step four, extracting an unlabeled training set from the conversation intention corpus, and performing iterative training on the original intention recognition model by using the unlabeled training set and the labeled training set to obtain a semi-supervised training model.
In the embodiment of the present invention, the semi-supervised training refers to a method for performing model training by combining a supervised corpus (corpus with artificial labels) and an unsupervised corpus (corpus without artificial labels). Meanwhile, a preset number of unlabeled user corpora can be extracted from the session intention corpus as unlabeled training sets, and the number of data in the unlabeled training sets is greater than or equal to that in the labeled training sets.
Specifically, the performing iterative training on the original intention recognition model by using the unlabeled training set and the labeled training set to obtain a semi-supervised training model includes:
outputting the prediction labels of the corpora in the unmarked training set and the prediction probabilities corresponding to the prediction labels by using the original intention recognition model;
selecting a prediction label with the prediction probability more than or equal to a preset first prediction threshold value and a corpus corresponding to the prediction label as supplementary labeling training data to be added to the labeling training set;
and predicting the intention label of each corpus in the supplemented labeled training set again by using the original intention recognition model, calculating the accuracy, returning to the step of extracting the unlabeled training set from the conversation intention corpus when the accuracy is not converged, and stopping iteration until the accuracy is converged to obtain the semi-supervised training model.
In an optional embodiment of the invention, a large amount of labeled data can be quickly obtained by a semi-supervised iterative training method with only a small amount of labeled data, so that the model training precision is improved.
And fifthly, performing cross training on the semi-supervised training model by using a cross verification method to obtain a standard intention recognition model, and outputting an intention recognition result of the text to be recognized by using the standard intention recognition model.
In the embodiment of the invention, although the semi-supervised training model and a large amount of labeled (namely labeled) corpora are obtained by a semi-supervised iteration method, labels with low probability or corpora with wrong labels exist in the corpora labeled by the model, and therefore, the labeling training set is adjusted by a cross-validation method, and the precision of model training can be further improved.
Specifically, the cross-training of the semi-supervised training model by using a cross-validation method to obtain a standard intention recognition model includes:
dividing all the label training sets into label training sets with preset number;
randomly selecting a label training set as a verification set, training the semi-supervised training model by using the unselected label training set, calculating the accuracy of the semi-supervised training model, and obtaining an identification model when the accuracy is greater than or equal to a preset second accuracy threshold;
performing label prediction on the verification set by using the identification model to obtain a prediction label of which the prediction probability in the verification set is less than or equal to a preset second prediction threshold;
and returning to the step of randomly selecting one label training set as a verification set until all the label training sets are selected as the verification set, adjusting the prediction labels and the linguistic data corresponding to the prediction labels with the prediction probability less than or equal to the second prediction threshold value in all the verification sets, and training the semi-supervised training model by using the adjusted label training set to obtain the standard intention recognition model.
In the embodiment of the invention, a 5-fold cross validation method can be used for cross training, namely, all labeling training sets are divided into 5 label training sets, 1 label training set is selected as a validation set each time, other unselected label training sets are used for training the semi-supervised training model, the identification model obtained by training is used for carrying out label prediction on the validation set, labels with lower prediction probability can be directly deleted or manually corrected, inaccurate labeling linguistic data can be quickly found out through 5-fold cross validation and adjusted, and the semi-supervised training model is retrained by using all adjusted linguistic data, so that the accuracy of identification of an intention identification model can be further improved.
In an optional embodiment of the present invention, for example, in a repayment link, the intention text "of the user a is input and i forget, and after a while, …" can directly output the intention label of the user as: "later deposit".
According to the method and the device, the entity labeling and the entity enhancement processing are carried out on the intention text, so that the useful information in the intention text can be focused more, and the accuracy of model training is improved. And the clustering-based intention labels of the intention categories are used for labeling, so that the difficulty of labeling the training data is greatly reduced, and meanwhile, the original intention recognition model is subjected to iterative training by using the unlabeled training set and the labeled training set, so that the precision of model training is improved. And the low-probability corpora generated in the model training can be continuously adjusted through a cross validation method, and cross training is performed by using the adjusted corpora, so that the accuracy of model identification is further improved. Therefore, the intention recognition device based on semi-supervised learning provided by the invention can solve the problem of low intention recognition accuracy.
Fig. 3 is a schematic structural diagram of an electronic device implementing an intent recognition method based on semi-supervised learning according to an embodiment of the present invention.
The electronic device may include a processor 10, a memory 11, a communication interface 12, and a bus 13, and may further include a computer program, such as a semi-supervised learning based intention recognition program, stored in the memory 11 and executable on the processor 10.
The memory 11 includes at least one type of readable storage medium, which includes flash memory, removable hard disk, multimedia card, card-type memory (e.g., SD or DX memory, etc.), magnetic memory, magnetic disk, optical disk, etc. The memory 11 may in some embodiments be an internal storage unit of the electronic device, for example a removable hard disk of the electronic device. The memory 11 may also be an external storage device of the electronic device in other embodiments, such as a plug-in mobile hard disk, a Smart Memory Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, which are provided on the electronic device. Further, the memory 11 may also include both an internal storage unit and an external storage device of the electronic device. The memory 11 may be used not only to store application software installed in the electronic device and various types of data, such as codes of an intention recognition program based on semi-supervised learning, etc., but also to temporarily store data that has been output or is to be output.
The processor 10 may be composed of an integrated circuit in some embodiments, for example, a single packaged integrated circuit, or may be composed of a plurality of integrated circuits packaged with the same or different functions, including one or more Central Processing Units (CPUs), microprocessors, digital Processing chips, graphics processors, and combinations of various control chips. The processor 10 is a Control Unit (Control Unit) of the electronic device, connects various components of the whole electronic device by using various interfaces and lines, and executes various functions and processes data of the electronic device by running or executing programs or modules (e.g., an intention recognition program based on semi-supervised learning, etc.) stored in the memory 11 and calling data stored in the memory 11.
The communication interface 12 is used for communication between the electronic device and other devices, and includes a network interface and a user interface. Optionally, the network interface may include a wired interface and/or a wireless interface (e.g., WI-FI interface, bluetooth interface, etc.), which are typically used to establish a communication connection between the electronic device and other electronic devices. The user interface may be a Display (Display), an input unit such as a Keyboard (Keyboard), and optionally a standard wired interface, a wireless interface. Alternatively, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, or the like. The display, which may also be referred to as a display screen or display unit, is suitable, among other things, for displaying information processed in the electronic device and for displaying a visualized user interface.
The bus 13 may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The bus 13 may be divided into an address bus, a data bus, a control bus, etc. The bus 13 is arranged to enable connection communication between the memory 11 and at least one processor 10 or the like.
Fig. 3 shows only an electronic device having components, and those skilled in the art will appreciate that the structure shown in fig. 3 does not constitute a limitation of the electronic device, and may include fewer or more components than those shown, or some components may be combined, or a different arrangement of components.
For example, although not shown, the electronic device may further include a power supply (such as a battery) for supplying power to each component, and preferably, the power supply may be logically connected to the at least one processor 10 through a power management device, so that functions of charge management, discharge management, power consumption management and the like are realized through the power management device. The power supply may also include any component of one or more dc or ac power sources, recharging devices, power failure detection circuitry, power converters or inverters, power status indicators, and the like. The electronic device may further include various sensors, a bluetooth module, a Wi-Fi module, and the like, which are not described herein again.
Further, the electronic device may further include a network interface, and optionally, the network interface may include a wired interface and/or a wireless interface (such as a WI-FI interface, a bluetooth interface, etc.), which are generally used to establish a communication connection between the electronic device and other electronic devices.
Optionally, the electronic device may further comprise a user interface, which may be a Display (Display), an input unit (such as a Keyboard), and optionally a standard wired interface, a wireless interface. Alternatively, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, or the like. The display, which may also be referred to as a display screen or display unit, is suitable, among other things, for displaying information processed in the electronic device and for displaying a visualized user interface.
It is to be understood that the described embodiments are for purposes of illustration only and that the scope of the appended claims is not limited to such structures.
The semi-supervised learning based intention recognition program stored by the memory 11 in the electronic device is a combination of a plurality of instructions that, when executed in the processor 10, may implement:
extracting an intention text of a user from a conversation intention corpus, and carrying out entity labeling and entity enhancement processing on the intention text to obtain an intention entity set;
performing label clustering on the intention entities in the intention entity set to obtain intention categories;
obtaining a labeling training set by using the intention labels of the intention categories, and training a pre-constructed neural network by using the labeling training set to obtain an original intention recognition model;
extracting an unlabeled training set from the conversation intention corpus, and performing iterative training on the original intention recognition model by using the unlabeled training set and the labeled training set to obtain a semi-supervised training model;
and performing cross training on the semi-supervised training model by using a cross verification method to obtain a standard intention recognition model, and outputting an intention recognition result of the text to be recognized by using the standard intention recognition model.
Specifically, the specific implementation method of the processor 10 for the instruction may refer to the description of the relevant steps in the embodiment corresponding to fig. 1, which is not described herein again.
Further, the electronic device integrated module/unit, if implemented in the form of a software functional unit and sold or used as a separate product, may be stored in a computer readable storage medium. The computer readable storage medium may be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying said computer program code, recording medium, U-disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM).
The present invention also provides a computer-readable storage medium, storing a computer program which, when executed by a processor of an electronic device, may implement:
extracting an intention text of a user from a conversation intention corpus, and carrying out entity labeling and entity enhancement processing on the intention text to obtain an intention entity set;
performing label clustering on the intention entities in the intention entity set to obtain intention categories;
obtaining a labeling training set by using the intention labels of the intention categories, and training a pre-constructed neural network by using the labeling training set to obtain an original intention recognition model;
extracting an unlabeled training set from the conversation intention corpus, and performing iterative training on the original intention recognition model by using the unlabeled training set and the labeled training set to obtain a semi-supervised training model;
and performing cross training on the semi-supervised training model by using a cross verification method to obtain a standard intention recognition model, and outputting an intention recognition result of the text to be recognized by using the standard intention recognition model.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus, device and method can 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.
The embodiment of the application can acquire and process related data based on an artificial intelligence technology. Among them, Artificial Intelligence (AI) is a theory, method, technique and application system that simulates, extends and expands human Intelligence using a digital computer or a machine controlled by a digital computer, senses the environment, acquires knowledge and uses the knowledge to obtain the best result.
The artificial intelligence infrastructure generally includes technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a robot technology, a biological recognition technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and the like.
The block chain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, a consensus mechanism, an encryption algorithm and the like. A block chain (Blockchain), which is essentially a decentralized database, is a series of data blocks associated by using a cryptographic method, and each data block contains information of a batch of network transactions, so as to verify the validity (anti-counterfeiting) of the information and generate a next block. The blockchain may include a blockchain underlying platform, a platform product service layer, an application service layer, and the like.
Furthermore, it is obvious that the word "comprising" does not exclude other elements or steps, and the singular does not exclude the plural. A plurality of units or means recited in the system claims may also be implemented by one unit or means in software or hardware. The terms second, etc. are used to denote names, but not any particular order.
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 intention recognition method based on semi-supervised learning, characterized in that the method comprises:
extracting an intention text of a user from a conversation intention corpus, and carrying out entity labeling and entity enhancement processing on the intention text to obtain an intention entity set;
performing label clustering on the intention entities in the intention entity set to obtain intention categories;
obtaining a labeling training set by using the intention labels of the intention categories, and training a pre-constructed neural network by using the labeling training set to obtain an original intention recognition model;
extracting an unlabeled training set from the conversation intention corpus, and performing iterative training on the original intention recognition model by using the unlabeled training set and the labeled training set to obtain a semi-supervised training model;
and performing cross training on the semi-supervised training model by using a cross verification method to obtain a standard intention recognition model, and outputting an intention recognition result of the text to be recognized by using the standard intention recognition model.
2. The semi-supervised learning based intention recognition method as claimed in claim 1, wherein the entity labeling and entity enhancement processing are performed on the intention text to obtain an intention entity set, and the method comprises the following steps:
carrying out entity labeling processing on the intention text by using a preset sequence labeling method;
entity prediction is carried out on a plurality of entities contained in an entity labeling result by utilizing an entity recognition layer in a preset entity recognition model, and type prediction scores of the plurality of entities are obtained;
and screening the entities according to the type prediction scores and a preset entity limiting rule, performing entity strengthening processing on the screened entities, and summarizing all entities strengthened by the entities to obtain the intention entity set.
3. The semi-supervised learning based intention recognition method as claimed in claim 2, wherein the tag clustering of the intention entities in the intention entity set to obtain an intention category comprises:
vector mapping is carried out on entities in the intention entity set by utilizing a preset natural language model, and an intention vector set is obtained;
randomly selecting a preset number of samples from the intention vector set as a clustering center;
sequentially calculating the distance from each sample in the intention vector set to the clustering center, and classifying each sample into the category corresponding to the clustering center with the minimum distance to obtain a plurality of category clusters;
and recalculating the clustering center of each category cluster, returning and calculating the distance from each sample in the intention vector set to the clustering center in sequence until the clustering centers of the plurality of category clusters converge, and determining the converged category cluster as the intention category.
4. The semi-supervised learning based intention recognition method of claim 3, wherein the calculating the cluster center of each category cluster comprises:
calculating the cluster center of each category cluster by the following clustering formula:
wherein E isiIs the ith cluster center, CiIs the ith class cluster and x is the sample in the class cluster.
5. The semi-supervised learning based intention recognition method of claim 1, wherein training a pre-constructed neural network with the labeled training set to obtain an original intention recognition model comprises:
performing intention prediction on the text in the labeling training set by using the neural network to obtain a prediction result;
and calculating prediction accuracy according to the prediction result and the intention labels in the standard training set, and obtaining the original intention recognition model when the prediction accuracy is greater than or equal to a preset first accuracy threshold.
6. The semi-supervised learning based intention recognition method of claim 1, wherein the iterative training of the original intention recognition model by using the unlabeled training set and the labeled training set to obtain a semi-supervised training model comprises:
outputting the prediction labels of the corpora in the unmarked training set and the prediction probabilities corresponding to the prediction labels by using the original intention recognition model;
selecting a prediction label with the prediction probability more than or equal to a preset first prediction threshold value and a corpus corresponding to the prediction label as supplementary labeling training data to be added to the labeling training set;
and predicting the intention label of each corpus in the supplemented labeled training set again by using the original intention recognition model, calculating the accuracy, returning to the step of extracting the unlabeled training set from the conversation intention corpus when the accuracy is not converged, and stopping iteration until the accuracy is converged to obtain the semi-supervised training model.
7. The semi-supervised learning based intention recognition method of claim 1, wherein the cross-training of the semi-supervised training model by using a cross-validation method to obtain a standard intention recognition model comprises:
dividing all the label training sets into label training sets with preset number;
randomly selecting a label training set as a verification set, training the semi-supervised training model by using the unselected label training set, calculating the accuracy of the semi-supervised training model, and obtaining an identification model when the accuracy is greater than or equal to a preset second accuracy threshold;
performing label prediction on the verification set by using the identification model to obtain a prediction label of which the prediction probability in the verification set is less than or equal to a preset second prediction threshold;
and returning to the step of randomly selecting one label training set as a verification set until all the label training sets are selected as the verification set, adjusting the prediction labels and the linguistic data corresponding to the prediction labels with the prediction probability less than or equal to the second prediction threshold value in all the verification sets, and training the semi-supervised training model by using the adjusted label training set to obtain the standard intention recognition model.
8. An intent recognition apparatus based on semi-supervised learning, the apparatus comprising:
the intention entity extraction module is used for extracting an intention text of a user from the conversation intention corpus, and carrying out entity labeling and entity enhancement processing on the intention text to obtain an intention entity set;
the label clustering module is used for carrying out label clustering on the intention entities in the intention entity set to obtain intention categories;
the original model training module is used for obtaining a labeling training set by using the intention labels of the intention categories and training a pre-constructed neural network by using the labeling training set to obtain an original intention identification model;
the iterative training module is used for extracting an unlabeled training set from the conversation intention corpus and performing iterative training on the original intention recognition model by using the unlabeled training set and the labeled training set to obtain a semi-supervised training model;
and the intention recognition module is used for performing cross training on the semi-supervised training model by using a cross verification method to obtain a standard intention recognition model and outputting an intention recognition result of the text to be recognized by using the standard intention recognition model.
9. An electronic device, characterized in that the electronic device comprises:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the semi-supervised learning based intention identifying method as claimed in any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the semi-supervised learning based intention recognition method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111009255.0A CN113704429A (en) | 2021-08-31 | 2021-08-31 | Semi-supervised learning-based intention identification method, device, equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111009255.0A CN113704429A (en) | 2021-08-31 | 2021-08-31 | Semi-supervised learning-based intention identification method, device, equipment and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113704429A true CN113704429A (en) | 2021-11-26 |
Family
ID=78657539
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111009255.0A Pending CN113704429A (en) | 2021-08-31 | 2021-08-31 | Semi-supervised learning-based intention identification method, device, equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113704429A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114462531A (en) * | 2022-01-30 | 2022-05-10 | 支付宝(杭州)信息技术有限公司 | Model training method and device and electronic equipment |
| CN114691852A (en) * | 2022-06-01 | 2022-07-01 | 阿里巴巴达摩院(杭州)科技有限公司 | Man-machine conversation system and method |
| CN115269809A (en) * | 2022-09-19 | 2022-11-01 | 支付宝(杭州)信息技术有限公司 | Method and device for training intention recognition model and method and device for recognizing intention |
| CN115601617A (en) * | 2022-11-25 | 2023-01-13 | 安徽数智建造研究院有限公司（Cn） | Training method and device of banded void recognition model based on semi-supervised learning |
| CN116628177A (en) * | 2023-05-22 | 2023-08-22 | 福建省网络与信息安全测评中心 | Interactive data processing method and system for network security platform |
| CN116702048A (en) * | 2023-08-09 | 2023-09-05 | 恒生电子股份有限公司 | Newly added intention recognition method, training method and device of distributed external monitoring model and electronic equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108363690A (en) * | 2018-02-08 | 2018-08-03 | 北京十三科技有限公司 | Dialog semantics Intention Anticipation method based on neural network and learning training method |
| US20200242302A1 (en) * | 2019-01-29 | 2020-07-30 | Ricoh Company, Ltd. | Intention identification method, intention identification apparatus, and computer-readable recording medium |
| CN112417121A (en) * | 2020-11-20 | 2021-02-26 | 平安普惠企业管理有限公司 | Client intention recognition method and device, computer equipment and storage medium |
| CN112988963A (en) * | 2021-02-19 | 2021-06-18 | 平安科技（深圳）有限公司 | User intention prediction method, device, equipment and medium based on multi-process node |
| CN113051356A (en) * | 2021-04-21 | 2021-06-29 | 深圳壹账通智能科技有限公司 | Open relationship extraction method and device, electronic equipment and storage medium |
| CN113191148A (en) * | 2021-04-30 | 2021-07-30 | 西安理工大学 | Rail transit entity identification method based on semi-supervised learning and clustering |
-
2021
- 2021-08-31 CN CN202111009255.0A patent/CN113704429A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108363690A (en) * | 2018-02-08 | 2018-08-03 | 北京十三科技有限公司 | Dialog semantics Intention Anticipation method based on neural network and learning training method |
| US20200242302A1 (en) * | 2019-01-29 | 2020-07-30 | Ricoh Company, Ltd. | Intention identification method, intention identification apparatus, and computer-readable recording medium |
| CN112417121A (en) * | 2020-11-20 | 2021-02-26 | 平安普惠企业管理有限公司 | Client intention recognition method and device, computer equipment and storage medium |
| CN112988963A (en) * | 2021-02-19 | 2021-06-18 | 平安科技（深圳）有限公司 | User intention prediction method, device, equipment and medium based on multi-process node |
| CN113051356A (en) * | 2021-04-21 | 2021-06-29 | 深圳壹账通智能科技有限公司 | Open relationship extraction method and device, electronic equipment and storage medium |
| CN113191148A (en) * | 2021-04-30 | 2021-07-30 | 西安理工大学 | Rail transit entity identification method based on semi-supervised learning and clustering |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114462531A (en) * | 2022-01-30 | 2022-05-10 | 支付宝(杭州)信息技术有限公司 | Model training method and device and electronic equipment |
| CN114691852A (en) * | 2022-06-01 | 2022-07-01 | 阿里巴巴达摩院(杭州)科技有限公司 | Man-machine conversation system and method |
| CN114691852B (en) * | 2022-06-01 | 2022-08-12 | 阿里巴巴达摩院(杭州)科技有限公司 | Man-machine conversation system and method |
| CN115269809A (en) * | 2022-09-19 | 2022-11-01 | 支付宝(杭州)信息技术有限公司 | Method and device for training intention recognition model and method and device for recognizing intention |
| CN115601617A (en) * | 2022-11-25 | 2023-01-13 | 安徽数智建造研究院有限公司（Cn） | Training method and device of banded void recognition model based on semi-supervised learning |
| CN116628177A (en) * | 2023-05-22 | 2023-08-22 | 福建省网络与信息安全测评中心 | Interactive data processing method and system for network security platform |
| CN116628177B (en) * | 2023-05-22 | 2023-11-14 | 福建省网络与信息安全测评中心 | Interactive data processing method and system for network security platform |
| CN116702048A (en) * | 2023-08-09 | 2023-09-05 | 恒生电子股份有限公司 | Newly added intention recognition method, training method and device of distributed external monitoring model and electronic equipment |
| CN116702048B (en) * | 2023-08-09 | 2023-11-10 | 恒生电子股份有限公司 | Newly added intention recognition method, model training method, device and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113704429A (en) | Semi-supervised learning-based intention identification method, device, equipment and medium | |
| CN113051356A (en) | Open relationship extraction method and device, electronic equipment and storage medium | |
| CN113378970A (en) | Sentence similarity detection method and device, electronic equipment and storage medium | |
| CN112559687A (en) | Question identification and query method and device, electronic equipment and storage medium | |
| CN113807973A (en) | Text error correction method and device, electronic equipment and computer readable storage medium | |
| CN112667775A (en) | Keyword prompt-based retrieval method and device, electronic equipment and storage medium | |
| CN113821622A (en) | Answer retrieval method and device based on artificial intelligence, electronic equipment and medium | |
| CN113722483A (en) | Topic classification method, device, equipment and storage medium | |
| CN113360654B (en) | Text classification method, apparatus, electronic device and readable storage medium | |
| CN113204698A (en) | News subject term generation method, device, equipment and medium | |
| CN113658002A (en) | Decision tree-based transaction result generation method and device, electronic equipment and medium | |
| CN113254814A (en) | Network course video labeling method and device, electronic equipment and medium | |
| CN113205814A (en) | Voice data labeling method and device, electronic equipment and storage medium | |
| CN114757154B (en) | Job generation method, device and equipment based on deep learning and storage medium | |
| CN115146064A (en) | Intention recognition model optimization method, device, equipment and storage medium | |
| CN113420542B (en) | Dialogue generation method, device, electronic equipment and storage medium | |
| CN115510188A (en) | Text keyword association method, device, equipment and storage medium | |
| CN114943306A (en) | Intention classification method, device, equipment and storage medium | |
| CN112989043B (en) | Reference resolution method, reference resolution device, electronic equipment and readable storage medium | |
| CN115221323A (en) | Cold start processing method, device, equipment and medium based on intention recognition model | |
| CN115203364A (en) | Software fault feedback processing method, device, equipment and readable storage medium | |
| CN114595321A (en) | Question marking method and device, electronic equipment and storage medium | |
| CN113806540A (en) | Text labeling method and device, electronic equipment and storage medium | |
| CN114462411B (en) | Named entity recognition method, device, equipment and storage medium | |
| CN111680513B (en) | Feature information identification method and device 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 |