CN110489517B - Automatic learning method and system of virtual assistant - Google Patents
Automatic learning method and system of virtual assistant Download PDFInfo
- Publication number
- CN110489517B CN110489517B CN201810436639.2A CN201810436639A CN110489517B CN 110489517 B CN110489517 B CN 110489517B CN 201810436639 A CN201810436639 A CN 201810436639A CN 110489517 B CN110489517 B CN 110489517B
- Authority
- CN
- China
- Prior art keywords
- vocabulary
- model
- data
- corpus
- corpus data
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 29
- 230000006870 function Effects 0.000 claims abstract description 60
- 230000011218 segmentation Effects 0.000 claims abstract description 44
- 238000004458 analytical method Methods 0.000 claims abstract description 41
- 238000003058 natural language processing Methods 0.000 claims abstract description 15
- 230000006399 behavior Effects 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 239000013598 vector Substances 0.000 claims description 4
- 238000012384 transportation and delivery Methods 0.000 description 10
- 238000013499 data model Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 238000007726 management method Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000013439 planning Methods 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 235000013402 health food Nutrition 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000015654 memory Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Machine Translation (AREA)
Abstract
An automatic learning method and system for virtual assistant. The automatic learning method of the virtual assistant comprises the following steps: receiving audio input and identifying audio to form corpus data; analyzing the corpus data by using a natural language processing model to generate language characteristic information corresponding to the corpus data; performing function context analysis on the language characteristic information according to the function context information, and judging operation corresponding to one of the intentions; if the function situation analysis cannot judge the operation corresponding to one of the intentions, word segmentation is carried out on the corpus data; judging whether new vocabulary or new corpus data exists according to the word segmentation result; if new vocabulary exists, the natural language processing model is updated according to the meaning of the new vocabulary, and if new corpus data exists, the function context analysis is updated according to the intention of the new corpus data. Therefore, the ERP system can be used by a user more quickly and conveniently.
Description
Technical Field
The present disclosure relates to an automatic learning method and system, and more particularly to an automatic learning method and system for a virtual assistant.
Background
An enterprise resource planning system (Enterprise Resource Planning, ERP), abbreviated as ERP system, refers to a management platform that provides decisions for an enterprise decision-making layer based on information technology. The method mainly manages the people stream, the logistics, the information stream and the fund stream of the enterprise uniformly so as to utilize the resources of the enterprise to the maximum extent. The ERP system has three functions of production control, logistics management and financial management, so that the ERP system is quite large in scale.
The virtual assistant is applied to the ERP system, so that a user can be helped to communicate with the huge ERP system more quickly, the time spent by the user in using the ERP system can be saved, but because of different habits of each user in using the ERP system, the virtual assistant can not understand the problem of the user, and the user is difficult to use the ERP system.
Disclosure of Invention
The invention mainly aims at providing an automatic learning method and an automatic learning system for a virtual assistant, which mainly enable the virtual assistant to have an automatic learning function, enable the virtual assistant to automatically learn speaking habits of users or words of special terms in industry in the process of communicating with the users, and achieve the effect that the users can use an ERP system more quickly and conveniently.
To achieve the above object, in a first aspect, an automatic learning method of a virtual assistant is provided, the method includes: receiving audio input and identifying audio to form corpus data; analyzing the corpus data by using a natural language processing model to generate language characteristic information corresponding to the corpus data, wherein the language characteristic information comprises a plurality of intentions, probabilities corresponding to the intentions and a plurality of words; performing function context analysis on the language characteristic information according to the function context information, and judging an operation corresponding to one of the plurality of intentions; if the job situation analysis cannot judge the operation corresponding to one of the plurality of intentions, word segmentation is carried out on the corpus data; judging whether new vocabulary or new corpus data exists according to the word segmentation result; if new vocabulary exists, updating a natural language processing model according to the meaning of the new vocabulary, and if new corpus data exists, updating functional context analysis according to the intention of the new corpus data; wherein the operation includes one of a query data operation and an execute instruction operation.
According to one embodiment of the present application, further comprising: generating a system domain vocabulary set according to an application knowledge database and a domain knowledge database; the system domain vocabulary set and the plurality of service application parameters form a key entity set, wherein the key entity set comprises a plurality of system domain vocabularies; classifying a plurality of training corpora into one of the query data operations and the execution instruction operations; differentiating the intentions of the plurality of training corpuses corresponding to the query data operation according to the categories in the enterprise database to form a plurality of query data operation intentions, and differentiating the intentions of the plurality of training corpuses corresponding to the execution instruction operation according to the service behaviors provided by the enterprise resource system to form a plurality of execution instruction operation intentions; establishing a template of the operation intents of the plurality of query data and a template of the operation intents of the plurality of execution instructions; establishing the overall database according to the key entity set, the templates of the operation intents of the plurality of query data and the templates of the operation intents of the plurality of execution instructions; identifying a plurality of first probabilities of the plurality of system domain words in the key entity set in the plurality of training corpuses, analyzing a plurality of sentence pattern structures of the plurality of training corpuses through the identified plurality of system domain words, and a plurality of correlations among the plurality of system domain words, and establishing a common word model according to the plurality of first probabilities and the plurality of correlations; and analyzing a plurality of second probabilities of the plurality of system domain vocabularies in the plurality of query data operation intents and the plurality of execution instruction operation intents, and establishing a common semantic model according to the plurality of sentence pattern structures and the plurality of second probabilities.
According to one embodiment of the present application, further comprising: classifying the relationship strength of the data in a history database by using a classifier to generate a functional situation model; and breaking and analyzing the plurality of training corpuses, and generating a functional vocabulary model according to the data in the historical database.
According to an embodiment of the present invention, the function context analysis further includes: comparing the functional context information with the functional context model by utilizing the corpus data and the functional context information, and generating a functional context identification result; and judging that one of the plurality of intentions corresponds to one of the query data operation and the execution instruction operation according to the function situation recognition result.
According to an embodiment of the present disclosure, the word segmentation process further includes: word segmentation is carried out on the corpus data according to the functional vocabulary model so as to generate a plurality of word segmentation; and calculating frequencies of the plurality of tokens.
According to one embodiment of the present application, further comprising: judging whether the frequency of the plurality of word segmentation calculated by the word segmentation processing is lower than a threshold value or not; if one of the plurality of word segments is lower than the threshold value, one of the plurality of word segments is the new vocabulary and receives the definition of the new vocabulary so as to update the common vocabulary model and the common semantic model; and if the plurality of word segments are higher than the threshold value, the corpus data is the new corpus data, and the intention of the new corpus data is received so as to update the functional situation model.
According to one embodiment of the present application, further comprising: judging whether the new corpus data is a common corpus, if so, updating the vocabulary set in the system field according to the new corpus data; and updating the vocabulary set of the system domain according to the new vocabulary.
According to an embodiment of the present invention, the analyzing the corpus data by the natural language processing model further includes: identifying whether the corpus data has the plurality of system domain vocabularies conforming to the key entity set by utilizing the common vocabulary model, setting an identification result as the plurality of vocabularies, and analyzing the occurrence probability of the plurality of vocabularies; analyzing sentence pattern structures of the corpus data according to the plurality of vocabularies; and identifying the plurality of intentions of the corpus data and the probabilities corresponding to the plurality of intentions by utilizing the common semantic model according to the probabilities of occurrence of the plurality of words and the sentence pattern structure of the corpus data.
In a second aspect, an automatic learning system for a virtual assistant is provided, which is respectively connected to an enterprise database and an enterprise resource system, and includes: a processor, a storage device and an input/output device. The storage device is electrically connected to the processor and used for storing the overall database, the application knowledge database, the domain knowledge database and the history database. The input/output device is electrically connected to the processor and is used for providing an interface for inputting audio. Wherein the processor comprises: the system comprises a voice recognition module, a corpus analysis module, a situation recognition module, an unknown corpus judgment module and an information updating module. The voice recognition module is used for recognizing the audio to form corpus data. The corpus analysis module is electrically connected with the voice recognition module and is used for analyzing corpus data by utilizing a natural language processing model so as to generate language characteristic information corresponding to the corpus data, wherein the language characteristic information comprises a plurality of intentions, probabilities corresponding to the intentions and a plurality of vocabularies. The context identification module is electrically connected with the corpus analysis module and is used for carrying out function context analysis on the language characteristic information according to the function context information and judging the operation corresponding to one of the plurality of intents. The unknown corpus judging module is electrically connected with the situation identifying module and is used for carrying out word segmentation processing on the corpus data when the situation identifying module cannot identify the operation corresponding to one of the plurality of intentions, and judging whether new vocabulary or new corpus data exists according to the word segmentation processed result. The updating information module is electrically connected with the unknown corpus judging module and is used for updating the natural language processing model according to the meaning of the new vocabulary when the new vocabulary is generated and updating the function situation analysis according to the intention of the new corpus data when the new corpus data is generated; wherein the operation includes one of a query data operation and an execute instruction operation.
According to an embodiment of the present disclosure, the processor further includes: the training module is electrically connected with the corpus analysis module and is used for generating a system domain vocabulary set according to the application knowledge database and the domain knowledge database, the system domain vocabulary set and a plurality of service application parameters form a key entity set, the key entity set comprises a plurality of system domain vocabularies, a plurality of training corpuses are classified into one of the query data operation and the execution instruction operation, a plurality of query data operation intents are formed by differentiating intents of the plurality of training corpuses corresponding to the query data operation according to categories in the enterprise database, and a plurality of execution instruction operation intents are formed by differentiating intents of the plurality of training corpuses corresponding to the execution instruction operation according to service behaviors provided by the enterprise resource system; the model establishing module is electrically connected with the training module, establishes a model of the operation intents of the plurality of query data and a model of the operation intents of the plurality of execution instructions, and establishes the overall database according to the key entity set, the model of the operation intents of the plurality of query data and the model of the operation intents of the plurality of execution instructions; the vocabulary model building module is electrically connected with the model building module, and is used for identifying a plurality of first probabilities of the plurality of system domain vocabularies in the key entity set in the plurality of training corpuses, analyzing a plurality of sentence pattern structures of the plurality of training corpuses through the identified plurality of system domain vocabularies, and a plurality of correlations among the plurality of system domain vocabularies, and building a common vocabulary model according to the plurality of first probabilities and the plurality of correlations; and a semantic model building module electrically connected with the template building module, for analyzing a plurality of second probabilities of the plurality of system domain vocabularies in the plurality of query data operation intentions and the plurality of execution instruction operation intentions, and building a common semantic model according to the plurality of sentence structures and the plurality of second probabilities.
According to an embodiment of the present disclosure, the processor further includes: the situation training module is electrically connected with the situation analysis module and is used for classifying the relation strength of the data in the historical database by utilizing a classifier to generate a functional situation model; and the vocabulary training module is electrically connected with the unknown corpus judging module and is used for carrying out word breaking and analysis on the plurality of training corpuses and generating a functional vocabulary model according to the data in the historical database.
According to an embodiment of the present invention, the context analysis module is further configured to compare the functional context information with the functional context model by using the corpus data, and generate a functional context recognition result, and determine, according to the functional context recognition result, that one of the plurality of intents corresponds to one of the query data operation and the execution instruction operation.
According to an embodiment of the present disclosure, the unknown corpus judging module is further configured to break the corpus data according to the functional vocabulary model, so as to generate a plurality of word segments, so as to calculate frequencies of the plurality of word segments.
According to an embodiment of the present disclosure, the update information module is further configured to determine whether the frequency of the plurality of word segments calculated by the word segmentation process is lower than a threshold value; if one of the plurality of word segments is lower than the threshold value, one of the plurality of word segments is the new vocabulary and receives the definition of the new vocabulary so as to update the common vocabulary model and the common semantic model; if the plurality of word segments are higher than the threshold value, the corpus data is the new corpus data, and the intention of the new corpus data is received so as to update the functional situation model.
According to an embodiment of the present disclosure, the update information module is further configured to determine whether the new corpus data is a common corpus, and if so, update the vocabulary set in the system domain according to the new corpus data; and updating the vocabulary set of the system domain according to the new vocabulary.
According to an embodiment of the present disclosure, the corpus analysis module is further configured to identify whether the corpus data has the plurality of system domain vocabularies conforming to the set of key entities according to the common vocabulary model, set the identification result as the plurality of vocabularies, analyze the probability of occurrence of the plurality of vocabularies, analyze the sentence pattern structure of the corpus data according to the plurality of vocabularies, and identify the plurality of intentions of the corpus data and the probability corresponding to the plurality of intentions according to the probability of occurrence of the plurality of vocabularies and the sentence pattern structure of the corpus data according to the common vocabulary model.
The automatic learning method of the virtual assistant and the automatic learning system of the virtual assistant mainly enable the virtual assistant to have an automatic learning function, enable the virtual assistant to automatically learn speaking habits of users or words of special words in industries in the process of communicating with the users, and achieve the effect that the users can use the ERP system more quickly and conveniently.
Drawings
The foregoing and other objects, features, advantages and embodiments of the invention will be apparent from the following description taken in conjunction with the accompanying drawings in which:
FIG. 1 is a schematic diagram of an automatic learning system of a virtual assistant according to some embodiments of the present disclosure;
FIG. 2 is a schematic diagram of a processor according to some embodiments of the present disclosure;
FIG. 3 is a flow chart of an automatic learning method of a virtual assistant according to some embodiments of the present disclosure;
FIG. 4 is a flow chart of a training data model depicted in accordance with some embodiments of the present disclosure;
FIG. 5 is a flowchart of step S320 according to some embodiments of the present disclosure;
FIG. 6 is a flowchart of step S330 according to some embodiments of the present disclosure;
FIG. 7 is a flowchart of step S340 according to some embodiments of the present disclosure; and
fig. 8 is a flowchart of step S360 according to some embodiments of the present disclosure.
Detailed Description
The following disclosure provides many different embodiments, or examples, for implementing different features of the invention. Elements and configurations in specific examples are used in the following discussion to simplify the present disclosure. Any exemplifications set out herein are for illustrative purposes only, and are not intended to limit the scope and meaning of the invention or its exemplifications in any manner. Moreover, the present disclosure may repeat reference numerals and/or letters in the various examples, which are for the purpose of simplicity and illustration, and does not in itself dictate a relationship between the various embodiments and/or configurations discussed below.
The term "about" as used throughout the specification and claims, unless otherwise indicated, shall generally have the meaning of each term used in this field, in the context of the disclosure and in the special context. Certain terms used to describe the disclosure are discussed below, or elsewhere in this specification, to provide additional guidance to those skilled in the art in describing the disclosure.
As used herein, "coupled" or "connected" may mean that two or more elements are in direct physical or electrical contact with each other, or in indirect physical or electrical contact with each other, and "coupled" or "connected" may also mean that two or more elements are in operation or action with each other.
The terms first, second, third, etc. are used herein to describe various elements, components, regions, layers and/or blocks. These elements, components, regions, layers and/or blocks should not be limited by these terms. These terms are only used to distinguish one element, component, region, layer or section from another. Accordingly, a first element, component, region, layer and/or section discussed below could be termed a second element, component, region, layer and/or section without departing from the spirit of the present invention. As used herein, the term "and/or" includes any combination of one or more of the listed associated items. Reference in the present document to "and/or" means any, all, or any combination of at least one of the listed elements.
Please refer to fig. 1. Fig. 1 is a schematic diagram of an automatic learning system 100 of a virtual assistant according to some embodiments of the present disclosure. As shown in fig. 1, the automatic learning system 100 of the virtual assistant includes a processor 110, a storage device 130, and an input/output device 150. The storage device 130 is used for storing an overall database 131, an application knowledge database 132, a domain knowledge database 133 and a history database 134, and the overall database 131, the application knowledge database 132, the domain knowledge database 133 and the history database 134 are electrically connected to the processor 110. The input/output device 150 is electrically connected to the processor 110 for providing an interface for inputting audio. In one embodiment, the input/output device 150 may be a keyboard, a touch screen, a microphone, a speaker, or other suitable input/output devices. The user can input audio through the interface provided by the input/output device.
In various embodiments of the invention, the processor 110 may be implemented as an integrated circuit such as a microcontroller (microcontroller), microprocessor (microprocessor), digital signal processor (digital signal processor), application specific integrated circuit (application specific integrated circuit, ASIC), logic circuit, or other similar element, or a combination of the above elements. The storage device 150 may be implemented as a memory, hard disk, portable disk, memory card, etc.
Referring to fig. 2, fig. 2 is a schematic diagram of a processor 110 according to some embodiments of the present disclosure. The processor 110 includes a speech recognition module 111, a corpus analysis module 112, a context recognition module 113, an unknown corpus judgment module 114, an update information module 115, a training module 121, a template creation module 122, a semantic model creation module 123, a vocabulary model creation module 124, a context training module 125, and a vocabulary training module 126. The corpus analysis module 112 is electrically connected with the voice recognition module 111, the context recognition module 113 is electrically connected with the corpus analysis module 112, the unknown corpus judgment module 114 is electrically connected with the context judgment module 113, and the update information module 115 is electrically connected with the unknown corpus judgment module 114. The training module 121 is electrically connected with the corpus analysis module 112, the template building module 122 is electrically connected with the training module 121, the semantic model building module 123 and the vocabulary model building module 124 are electrically connected with the template building module 122, the situation training module 125 is electrically connected with the situation identification module 113, and the unknown corpus judging module 114 is electrically connected with the vocabulary training module 126.
Please refer to fig. 1-3 together. Fig. 3 is a flow chart of an automatic learning method 300 of a virtual assistant according to some embodiments of the present disclosure. As shown in fig. 3, the automatic learning method 300 of the virtual assistant includes the following steps:
Step S310: receiving audio input and identifying audio to form corpus data;
step S320: analyzing the corpus data by using a natural language processing model to generate language characteristic information corresponding to the corpus data;
step S330: performing function context analysis on the language characteristic information according to the function context information, and judging operation corresponding to one of the intentions;
step S340: if the function situation analysis cannot judge the operation corresponding to one of the intentions, word segmentation is carried out on the corpus data;
step S350: judging whether new vocabulary or new corpus data exists according to the word segmentation result; and
step S360: if new vocabulary exists, the natural language processing model is updated according to the meaning of the new vocabulary, and if new corpus data exists, the function context analysis is updated according to the intention of the new corpus data.
In step S310, an audio input is received and audio is recognized to form corpus data. In one embodiment, the audio received through the input/output device 150 can be voice-recognized by the voice recognition module 111 of the processor 110, so as to convert the natural language of the user into corpus data. In another embodiment, the voice recognition may also transmit the audio to the cloud voice recognition system through the internet, and the cloud voice recognition system may be implemented as a voice recognition system of google, for example, after recognizing the audio by the cloud voice recognition system, the recognition result is used as corpus data.
Before executing step S320, a common vocabulary model and a common semantic model are established. Thus, referring to fig. 4, fig. 4 is a flow chart of a training data model according to some embodiments of the present disclosure. As shown in fig. 4, the training data model phase comprises the steps of:
step S410: generating a system domain vocabulary set according to the application knowledge database and the domain knowledge database;
step S420: forming a system domain vocabulary set and a plurality of service application parameters into a key entity set;
step S430: classifying the plurality of training corpora into one of query data operations and execute instruction operations;
step S440: differentiating the intentions of the training corpuses corresponding to the query data operation according to the categories in the enterprise database to form a plurality of query data operation intentions, and differentiating the intentions of the training corpuses corresponding to the execution instruction operation according to the service behaviors provided by the enterprise resource system to form a plurality of execution instruction operation intentions;
step S450: establishing a template of query data operation intents and executing instruction operation intents;
step S460: establishing an overall database according to the key entity set, the template of the query data operation intention and the template of the execution instruction operation intention;
Step S470: identifying a plurality of first probabilities of the system domain vocabulary in the key entity set in the training corpus, analyzing a plurality of sentence pattern structures of the training corpus through the identified system domain vocabulary, and a plurality of correlations among the system domain vocabulary, and establishing a common vocabulary model according to the first probabilities and the correlations; and
step S480: and analyzing a plurality of second probabilities of the system domain vocabulary in the query data operation intention and the execution instruction operation intention, and establishing a common semantic model according to the sentence pattern structure and the second probabilities.
In step S410 and step S420, a system domain vocabulary set is generated according to the application knowledge database 132 and the domain knowledge database 133, and the system domain vocabulary set and the plurality of service application parameters are utilized to form a key entity set, wherein the key entity set comprises a plurality of system domain vocabularies. For example, the set of key entities includes information such as business domain vocabulary and service application parameters of the enterprise system. The vocabulary in the enterprise domain refers to the vocabulary that may be needed by each enterprise in different domains, for example, the vocabulary used in the medical industry and the vocabulary used in the transportation industry are necessarily different, so that the vocabulary in the enterprise domain may be changed according to each enterprise using the ERP system. The service application parameters of the enterprise system are parameters corresponding to various services provided by the enterprise system, for example, the leave-leave function in the enterprise system may need leave-leave time, leave-leave information, and the system domain vocabulary in the key entity set may need to include leave-leave, annual leave, sick leave, business leave information, and the like.
Specifically, the key entity set further includes a data field name, a service name provided to the user by the enterprise system, a parameter value of a limiting condition set by the user during inquiry, a parameter value of a service application, an operation function of the enterprise system, and the like, where the operation function of the enterprise system may be an operation function such as leave application, overtime application, business trip application, report, and the like. The above information may have corresponding aliases, and may need to be input together when training the database, for example: the shipment may be given different names such as shipment details or sales bills for a manufacturer in a particular area.
In step S430, the plurality of training corpora are classified into one of a query data operation and an execution instruction operation. The training corpus may be natural language data such as instructions or questions that a user may possibly issue, and after the key entity set is established, the training corpus is classified according to intent. For example, if the user is for a virtual assistant: "please help me find the XX company's invoice". The intent classification of the present invention is classified as a query data operation, and the virtual assistant will help the user to query the XX company's invoice in the enterprise database. If the user is for a virtual assistant: "help me go on business for 1 month and 30 days", which is classified as executing instruction operation in the intention classification of the present invention, the virtual assistant will enter the enterprise resource system to help the user ask for the false.
In step S440, the intents of the training corpora corresponding to the query data operation are differentiated according to the categories in the enterprise database to form a plurality of query data operation intents, and the intents of the training corpora corresponding to the execution instruction operation are differentiated according to the service behaviors provided by the enterprise resource system to form a plurality of execution instruction operation intents. In one embodiment, intent is first differentiated for query data according to enterprise databases for each of the different domains. For example, the data fields stored in the enterprise database of the medical industry must not be identical to the enterprise database of the transportation industry, so that the user requirements of both are not necessarily identical. For example, a user in the healthcare industry may have different intentions to query medical record data, query room space, etc. for data, and a user in the transportation industry may have different intentions to query shipment records, query package shipping status, etc. for data. The service provided by the enterprise resource system of the medical industry is different from the transportation industry, and the query data operation or the service behavior operation provided by the enterprises of the different fields are not necessarily universal, so that the service provided by the enterprises of the different fields needs to be distinguished, for example, the user of the medical industry may have different intentions of providing registered service, providing service of ordering health food in hospital, and the like, and the user of the transportation industry may have different intentions of providing service of automatically classifying goods, arranging goods delivery order, and the like.
In step S450 and step S460, a template of query data operation intention and a template of execution instruction operation intention are established, and the overall database 131 is established according to the set of key entities, the template of query data operation intention and the template of execution instruction operation intention. For example, after the user distinguishes the operation intention of the query data and the operation intention of the execution instruction of the virtual assistant of the enterprise in a certain area, a corresponding template can be generated for each intention, according to the above example, the medical industry has 4 templates corresponding to the query of medical record data, the query of ward room, the provision of registered services and the provision of inpatient ordered health services, and the transportation industry has 4 templates corresponding to the query of delivery records, the query of package delivery status, the provision of services for automatically classifying goods and the provision of goods delivery sequence, and then an overall database 131 is established according to the templates and the key entity set.
In step S470, a plurality of first probabilities of occurrence of the system domain vocabulary in the key entity set in the training corpus are identified, a plurality of sentence pattern structures of the training corpus and a plurality of correlations between the system domain vocabulary are analyzed through the identified system domain vocabulary, and a common vocabulary model is established according to the first probabilities and the correlations. In one embodiment, the probability of each system domain vocabulary appearing in the training corpus is calculated using two algorithms, n-GRAM and Context-free grammar (CFG), and the sentence pattern structure of the training corpus and the relevance of the system domain vocabularies to each other are analyzed by the system domain vocabularies to build a common vocabulary model. For example, if there are "i want to query the quotation of XX company" and "i want to query the delivery of XX company" in the training corpus, and "XX company", "quotation" and "delivery" are all system domain vocabularies, but in the above example, since "XX company" may appear on average in each intention of query data operation, the probability of "XX company" is almost the same in each intention of query data operation, and "quotation" and "delivery" appear only in large amounts in the training corpus of the intention of querying some specific data, but not in the training corpus of the intention of querying other data, so the probability of "quotation" and "delivery" will be particularly high in the corresponding intention, but lower in other intentions.
In step S480, a plurality of second probabilities of occurrence of the vocabulary of the system domain in the query data operation intention and the execution instruction operation intention are analyzed, and a common semantic model is established according to the sentence structure and the second probabilities. In one embodiment, the probability of occurrence of the system domain vocabulary in the query data operation intent and the execution instruction operation intent is calculated using a hidden Markov model (Hidden Markov Model, HMM) algorithm to build a common semantic model, for example, a plurality of training corpus are input during the training data model stage, and the hidden Markov model algorithm must calculate the probability of occurrence of the system domain vocabulary in the disagreement map. In combination with the above example, if the corpus has "i want to query the shipment form of XX company", the "XX company" and "shipment form" can be found out according to the n-gram and the context-free grammar, and the hidden markov model algorithm can further determine that the "shipment form" is associated with the intent of querying shipment data according to the relationships between the probabilities of all the recognized system domain words in the query data operation intent and the execution instruction operation intent and the system domain words, and then combine the system domain words of "XX company" to automatically help the user query the shipment-related data of XX company in the enterprise database.
After the common vocabulary model and the common semantic model are established, step S320 is performed to analyze the corpus data by using the natural language processing model to generate language feature information corresponding to the corpus data, wherein the language feature information includes a plurality of intentions, probabilities corresponding to the intentions, and a plurality of vocabularies. Referring to fig. 5 for details of step S320, fig. 5 is a flowchart of step S320 according to some embodiments of the present disclosure. As shown in fig. 5, step S320 includes the steps of:
step S321: identifying whether the corpus data has the vocabulary which accords with the system field in the key entity set by utilizing the common vocabulary model, setting an identification result as the vocabulary in the language characteristic information, and analyzing the occurrence probability of the vocabulary in the language characteristic information;
step S322: analyzing sentence pattern structures of corpus data according to vocabulary in the feature information; and
step S323: and identifying the intention of the corpus data and the probability corresponding to the intention by utilizing the common semantic model according to the probability of occurrence of the vocabulary in the feature information and the sentence pattern structure of the corpus data.
In step S321 and step S322, whether the corpus data has the system domain vocabulary in the key entity set is identified by using the common vocabulary model, the identification result is set as the vocabulary in the language feature information, the occurrence probability of the vocabulary in the language feature information is analyzed, and then the sentence pattern structure of the corpus data is analyzed according to the vocabulary in the feature information. For example, the corpus data input by the user is identified by using the common vocabulary model, and the sentence pattern structure of the corpus data is further determined. For example, if the user is for a virtual assistant: the term "I want to search the order of the last month of XX" can be identified according to the common term model, and the terms such as "XX", "last month" and "order" can be identified.
In step S323, the intent of the corpus data and the probability corresponding to the intent are identified according to the probability of occurrence of the vocabulary in the feature information and the sentence pattern structure of the corpus data by using the common semantic model. According to the above example, the words of "XX company", "last month" and "delivery bill" are identified, and then the probability of these words in all intentions is further determined. All intents referred to herein include all query data operational intents and probabilities of executing instruction operational intents.
In step S330, the functional context analysis is performed on the language feature information according to the functional context information, so as to determine an operation corresponding to one of the intentions. Before performing the function context analysis, a function context model and a function vocabulary model are first established, the function context model firstly converts the data in the history database 134 into feature vectors when performing the function context analysis, then uses a machine learning algorithm to classify the data in the history database 134 according to various different contexts, then calculates the strong and weak relations between the feature vectors and each context, and then generates the function context model. Machine learning algorithms suitable for establishing the above-mentioned role scenarios include: support vector machines (Support Vector Machine, SVM) commonly used in traditional machine Learning, and algorithms such as convolutional neural networks (Convolutional Neural Networks, CNN), recurrent neural networks (Recurrent Neural Networks, RNN) and Long Short-Term Memory models (LSTM) related to Deep Learning at present.
In the above-mentioned method, the functional vocabulary model is analyzed by using a hidden markov model algorithm according to a large amount of input training corpus, then word segmentation is performed, and then the occurrence frequency of the segmented words is counted to generate a word segmentation frequency table, so as to establish the functional vocabulary model. Referring to fig. 6 for the detailed flow of step S330, fig. 6 is a flow chart of step S330 according to some embodiments of the present disclosure. As shown in fig. 6, step S330 includes the steps of:
step S331: comparing the function situation information with the function situation model by utilizing the corpus data and the function situation information, and generating a function situation identification result; and
step S332: and judging that one of the intentions corresponds to one of the operation of inquiring the data and the operation of executing the instruction according to the function situation recognition result.
In step S331, the corpus data, the functional context information and the functional context model are used for comparison, and a functional context recognition result is generated. The job context information includes the identity of the user, the job position of the user, the department of the user, the time and the place. Some of the functional context information may be sensed by the I/O device 150, for example, to detect the current status of the user (e.g., whether to go on business or not). According to the probability and vocabulary corresponding to all intentions obtained after the corpus data of the user are identified, and the similarity degree of the corpus data of the user and the data in the training data model can be further estimated by combining the functional context information, and the probability is taken as the probability of the corresponding intentions.
In step S332, it is determined that one of the intents corresponds to one of the query data operation and the execution instruction operation according to the function context recognition result. Since there are multiple query data operation intents and multiple execution instruction operation intents in the training data model, and the probability corresponding to each intention is generated after the calculation of the common semantic model, the intention with a lower probability value can be filtered by using a threshold value to obtain the most likely intention and confirm the corresponding operation. As can be seen from the above examples, when words such as "XX company", "last month" and "delivery bill" are identified, these words are judged to match the functional context information to find the most suitable operation intention of the query data or the operation intention of the execution instruction, and after the above operation, it is judged that the user is speaking to the virtual assistant: "I want to check out the order of the last month of XX company", most likely will check out the order of XX company, so it can correspond to the operation that the user wants to perform to query data. The judgment of the functional situation is required because different demands are caused by different information such as the position, department, operation time, operation place, etc. of the user, for example, the purchasing personnel and the financial staff can see [ vendor monthly statistics ], but the statistical objectives of [ vendor monthly statistics ] of the two are different: one is to count the incoming goods status of the manufacturer, and the other is to count the status of paying the manufacturer by the own company. The user, when talking to the virtual assistant, does not necessarily specify what is needed [ vendor per month statistics ], perhaps only: the simple sentence pattern is "i need to count up every month" for the manufacturer of the last month, so that more accurate judgment is needed to be performed in cooperation with the functional situation information of the user.
In step S340, if the function context analysis cannot determine the operation corresponding to one of the intentions, word segmentation is performed on the corpus data. Referring to fig. 7 for the detailed flow of step S340, fig. 7 is a flow chart of step S340 according to some embodiments of the present disclosure. As shown in fig. 7, step S340 includes the steps of:
step S341: word segmentation is carried out on the corpus data according to the functional vocabulary model so as to generate a plurality of word segments; and
step S342: the frequency of these segmentations is calculated.
In step S341 and step S342, word segmentation is performed on the corpus data according to the functional vocabulary model to generate a plurality of segmented words; the frequency of these segmentations is then calculated. If the job context analysis in step S330 cannot determine the operation corresponding to the input corpus data, word segmentation processing is required for the corpus data. Firstly, word breaking is carried out on the material data according to words stored in a previously pre-established functional word model, and then the frequency of a plurality of word breaking generated after word breaking is calculated.
In step S350 and step S360, according to the word segmentation result, whether new vocabulary or new corpus data exists is determined; if new vocabulary exists, the natural language processing model is updated according to the meaning of the new vocabulary, and if new corpus data exists, the function context analysis is updated according to the intention of the new corpus data. Referring to fig. 8 for the detailed flow of step S360, fig. 8 is a flow chart of step S360 according to some embodiments of the present disclosure. As shown in fig. 8, step S360 includes the steps of:
Step S361: judging whether the frequency of the word segmentation calculated by the word segmentation processing is lower than a threshold value or not;
step S362: if one of the segmented words is lower than a threshold value, one of the segmented words is a new vocabulary and receives the definition of the new vocabulary so as to update the common vocabulary model and the common semantic model; and
step S363: if the word segmentation is higher than the threshold value, the corpus data is new corpus data, and the intention of the new corpus data is received so as to update the functional context model.
In step S361 and step S362, it is determined whether the frequency of the words calculated by the word segmentation process is lower than a threshold value, and if one of the words is lower than the threshold value, one of the words is a new vocabulary and a definition of the new vocabulary is received to update the common vocabulary model and the common semantic model. In one embodiment, after the frequency of the word segmentation is calculated through word segmentation processing, the word segmentation below the threshold value is set as a new vocabulary, the virtual assist asks the user for the definition of the new vocabulary, and the new vocabulary and the definition of the new vocabulary are stored in the common vocabulary model and the common semantic model together. For example, if the corpus input by the user is "i want to find the contact of XX company", and if the virtual assistant cannot judge the meaning of "i want to find the contact of XX company", words such as "i", "want to find", "XX company", "contact", etc. are separated after the word segmentation process, if "XX company" is lower than the threshold value, the virtual assistant asks the user what the meaning of "XX company" is, and then the answer of the user and "XX company" are stored in the common vocabulary model and the common semantic model together; and the new vocabulary is also required to be stored in the vocabulary set in the system domain together and shared with all people.
In step S363, if the word segmentation is higher than the threshold, the corpus data is new corpus data, and the intent of the new corpus data is received to update the functional context model. Continuing the example of "I want to find contact of XX company" above, after word segmentation, we separate words such as "I", "want to find", "XX company", "contact", etc., if none of the words is lower than threshold value, we show that the virtual assistant does not understand the intent of corpus, and possibly the training corpus during training of the intelligent assistant is the description of "help I find contact of XX company", so that the virtual assistant cannot understand the intent of "I want to find contact of XX company", and the virtual assistant needs to ask what meaning the user "I want to find contact of XX company" is, and then store the answer of user and the new corpus of "I want to find contact of XX company" together into the functional context model. Before the function model is stored, whether the new corpus is the common corpus is judged, if so, the new corpus is used by other people when the virtual assistant is used, so that the new corpus is required to be stored into a vocabulary set in the system field, and all people share the new corpus; if not, the new corpus is represented by different expressions corresponding to the speaking habit of the user, so that only the functional situation model is required to be updated, and the vocabulary set in the system field is not required to be updated.
According to the embodiment of the present invention, the virtual assistant has an automatic learning function, so that the virtual assistant can update the database of the virtual assistant after inquiring the user if the vocabulary which is not understood by the intelligent assistant is in communication with the user, so that the virtual assistant can automatically learn the speaking habit of the user or the words of special words in the industry, and the effect that the user can use the ERP system more quickly and conveniently is achieved.
Additionally, the above illustration includes exemplary steps in a sequence, but the steps need not be performed in the order shown. It is within the contemplation of the present disclosure that these steps be performed in a different order. It is contemplated that sequences may be added, substituted, altered, and/or omitted within the spirit and scope of the embodiments of the disclosure.
While the present invention has been described with reference to the embodiments, it should be understood that the invention is not limited thereto, but may be variously modified and modified by those skilled in the art without departing from the spirit and scope of the present invention, and the scope of the present invention is accordingly defined by the appended claims.
Claims (14)
1. An automatic learning method of a virtual assistant, comprising:
Receiving an audio input and identifying the audio to form corpus data;
analyzing the corpus data by using a natural language processing model to generate language characteristic information corresponding to the corpus data, wherein the language characteristic information comprises a plurality of intentions, probabilities corresponding to the intentions and a plurality of words;
performing a function context analysis on the language feature information according to function context information, and judging an operation corresponding to one of the plurality of intents, wherein the function context information comprises at least one of an identity, a position, a department, a time and a place of a user, and the function context analysis further comprises: comparing the corpus data and the function situation information with a function situation model, and generating a function situation identification result; and judging that one of the intents corresponds to one of a query data operation and an execution instruction operation according to the function situation recognition result, wherein the function situation model is related to the strong and weak relation between the feature vector converted by the data in a history database and each of the situations;
if the function context analysis cannot judge the operation corresponding to one of the plurality of intentions, performing word segmentation on the corpus data;
Judging whether a new vocabulary or new corpus data exists or not according to the word segmentation result; and
if the new vocabulary exists, updating the natural language processing model according to the meaning of the new vocabulary, and if the new corpus data exists, updating the function situation analysis according to the intention of the new corpus data;
wherein the operation includes one of a query data operation and an execute instruction operation.
2. The method for automatic learning of a virtual assistant according to claim 1, further comprising:
generating a system domain vocabulary set according to an application knowledge database and a domain knowledge database;
the system domain vocabulary set and the plurality of service application parameters form a key entity set, wherein the key entity set comprises a plurality of system domain vocabularies;
classifying a plurality of training corpora into one of the query data operations and the execution instruction operations;
differentiating the intentions of the plurality of training corpuses corresponding to the query data operation according to the category in an enterprise database to form a plurality of query data operation intentions, and differentiating the intentions of the plurality of training corpuses corresponding to the execution instruction operation according to the service behavior provided by an enterprise resource system to form a plurality of execution instruction operation intentions;
Establishing a template of the operation intents of the plurality of query data and a template of the operation intents of the plurality of execution instructions;
establishing a general database according to the key entity set, the templates of the operation intents of the plurality of query data and the templates of the operation intents of the plurality of execution instructions;
identifying a plurality of first probabilities of the plurality of system domain words in the key entity set in the plurality of training corpuses, analyzing a plurality of sentence pattern structures of the plurality of training corpuses through the identified plurality of system domain words, and a plurality of correlations among the plurality of system domain words, and establishing a common word model according to the plurality of first probabilities and the plurality of correlations; and
analyzing a plurality of second probabilities of the plurality of system domain vocabularies in the plurality of query data operation intentions and the plurality of execution instruction operation intentions, and establishing a common semantic model according to the plurality of sentence pattern structures and the plurality of second probabilities.
3. The method for automatic learning of a virtual assistant according to claim 2, further comprising:
classifying the relationship strength of the data in the historical database by using a classifier to generate the function situation model; and
And breaking and analyzing the plurality of training corpuses, and generating a functional vocabulary model according to the data in the historical database.
4. The automatic learning method of a virtual assistant of claim 3, wherein the word segmentation process further comprises:
word segmentation is carried out on the corpus data according to the functional vocabulary model so as to generate a plurality of word segmentation; and
and calculating the frequency of the plurality of segmentation words.
5. The method for automatic learning of a virtual assistant of claim 4, further comprising:
judging whether the frequency of the plurality of word segmentation calculated by the word segmentation processing is lower than a threshold value or not;
if one of the plurality of word segments is lower than the threshold value, one of the plurality of word segments is the new vocabulary and receives the definition of the new vocabulary so as to update the common vocabulary model and the common semantic model; and
if the plurality of word segments are higher than the threshold value, the corpus data is the new corpus data, and the intention of the new corpus data is received so as to update the functional situation model.
6. The method for automatic learning of a virtual assistant of claim 5, further comprising:
judging whether the new corpus data is a common corpus, if so, updating the vocabulary set in the system field according to the new corpus data; and
And updating the vocabulary set in the system domain according to the new vocabulary.
7. The method of automatic learning of a virtual assistant of claim 2, wherein the natural language processing model analyzing the corpus data further comprises:
identifying whether the corpus data has the plurality of system domain vocabularies conforming to the key entity set by utilizing the common vocabulary model, setting an identification result as the plurality of vocabularies, and analyzing the occurrence probability of the plurality of vocabularies;
analyzing sentence pattern structures of the corpus data according to the plurality of vocabularies; and
and identifying the multiple intentions of the corpus data and the probabilities corresponding to the multiple intentions by utilizing the common semantic model according to the occurrence probabilities of the multiple vocabularies and the sentence pattern structure of the corpus data.
8. An automatic learning system for a virtual assistant, respectively connected to an enterprise database and an enterprise resource system, comprising:
a processor;
the storage device is electrically connected to the processor and used for storing a general database, an application knowledge database, a domain knowledge database and a history database;
an input/output device electrically connected to the processor for providing an interface for inputting an audio;
Wherein the processor comprises:
a voice recognition module for recognizing the audio to form a corpus data;
the language characteristic information comprises a plurality of intentions, probabilities corresponding to the intentions and a plurality of words;
the situation identification module is electrically connected with the corpus analysis module and is used for carrying out a function situation analysis on the language characteristic information according to a function situation information, judging an operation corresponding to one of the plurality of intentions, wherein the function situation information comprises at least one of an identity, a position, a department, a time and a place of a user, and the situation analysis module is further used for comparing the corpus data and the function situation information with a function situation model and generating a function situation identification result; and judging that one of the intents corresponds to one of a query data operation and an execution instruction operation according to the function situation recognition result, wherein the function situation model is related to the strong and weak relation between the feature vector converted by the data in a history database and each of the situations;
The unknown corpus judging module is electrically connected with the situation identifying module and is used for carrying out word segmentation processing on the corpus data when the situation identifying module cannot identify the operation corresponding to one of the plurality of intentions and judging whether a new vocabulary or new corpus data exists according to the word segmentation processed result; and
the updating information module is electrically connected with the unknown corpus judging module and is used for updating the natural language processing model according to the meaning of the new vocabulary when the new vocabulary is generated and updating the function situation analysis according to the intention of the new corpus data when the new corpus data is generated;
wherein the operation includes one of a query data operation and an execute instruction operation.
9. The automatic learning system of a virtual assistant of claim 8, wherein the processor further comprises:
the training module is electrically connected with the corpus analysis module and is used for generating a system domain vocabulary set according to the application knowledge database and the domain knowledge database, the system domain vocabulary set and a plurality of service application parameters form a key entity set, the key entity set comprises a plurality of system domain vocabularies, a plurality of training corpuses are classified into one of the query data operation and the execution instruction operation, a plurality of query data operation intents are formed by differentiating intents of the plurality of training corpuses corresponding to the query data operation according to categories in the enterprise database, and a plurality of execution instruction operation intents are formed by differentiating intents of the plurality of training corpuses corresponding to the execution instruction operation according to service behaviors provided by the enterprise resource system;
The model establishing module is electrically connected with the training module, establishes a model of the operation intents of the plurality of query data and a model of the operation intents of the plurality of execution instructions, and establishes the overall database according to the key entity set, the model of the operation intents of the plurality of query data and the model of the operation intents of the plurality of execution instructions;
the vocabulary model building module is electrically connected with the model building module, and is used for identifying a plurality of first probabilities of the plurality of system domain vocabularies in the key entity set in the plurality of training corpuses, analyzing a plurality of sentence pattern structures of the plurality of training corpuses through the identified plurality of system domain vocabularies, and a plurality of correlations among the plurality of system domain vocabularies, and building a common vocabulary model according to the plurality of first probabilities and the plurality of correlations; and
and the semantic model building module is electrically connected with the template building module, analyzes a plurality of second probabilities of the plurality of system domain vocabularies in the plurality of query data operation intentions and the plurality of execution instruction operation intentions, and builds a common semantic model according to the plurality of sentence structures and the plurality of second probabilities.
10. The automatic learning system of a virtual assistant of claim 9, wherein the processor further comprises:
the situation training module is electrically connected with the situation identification module and is used for classifying the relation strength of the data in the history database by using a classifier to generate the function situation model; and
and the vocabulary training module is electrically connected with the unknown corpus judging module and is used for carrying out word breaking and analysis on the plurality of training corpuses and generating a functional vocabulary model according to the data in the historical database.
11. The automatic learning system of claim 10, wherein the unknown corpus judging module is further configured to perform word segmentation on the corpus data according to the functional vocabulary model to generate a plurality of segmented words, so as to calculate the frequency of the plurality of segmented words.
12. The automatic learning system of claim 11, wherein the update information module is further configured to determine whether the frequency of the plurality of tokens calculated by the token processing is below a threshold; if one of the plurality of word segments is lower than the threshold value, one of the plurality of word segments is the new vocabulary and receives the definition of the new vocabulary so as to update the common vocabulary model and the common semantic model; if the plurality of word segments are higher than the threshold value, the corpus data is the new corpus data, and the intention of the new corpus data is received so as to update the functional situation model.
13. The automatic learning system of claim 12, wherein the update information module is further configured to determine whether the new corpus data is a common corpus, and if so, update the system domain vocabulary set according to the new corpus data; and updating the vocabulary set of the system domain according to the new vocabulary.
14. The automatic learning system of a virtual assistant according to claim 9, wherein the corpus analysis module is further configured to identify whether the corpus data has the plurality of system domain vocabularies corresponding to the set of key entities according to the common vocabulary model, set the identification result as the plurality of vocabularies, analyze probabilities of occurrence of the plurality of vocabularies, analyze a sentence pattern structure of the corpus data according to the plurality of vocabularies, and identify the plurality of intentions of the corpus data and probabilities corresponding to the plurality of intentions according to the probabilities of occurrence of the plurality of vocabularies and the sentence pattern structure of the corpus data according to the common vocabulary model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810436639.2A CN110489517B (en) | 2018-05-09 | 2018-05-09 | Automatic learning method and system of virtual assistant |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810436639.2A CN110489517B (en) | 2018-05-09 | 2018-05-09 | Automatic learning method and system of virtual assistant |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110489517A CN110489517A (en) | 2019-11-22 |
| CN110489517B true CN110489517B (en) | 2023-10-31 |
Family
ID=68543225
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810436639.2A Active CN110489517B (en) | 2018-05-09 | 2018-05-09 | Automatic learning method and system of virtual assistant |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110489517B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109949797B (en) | 2019-03-11 | 2021-11-12 | 北京百度网讯科技有限公司 | Method, device, equipment and storage medium for generating training corpus |
| CN112559699A (en) * | 2020-11-09 | 2021-03-26 | 联想(北京)有限公司 | Information interaction method, device and equipment |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2007094291A (en) * | 2005-09-30 | 2007-04-12 | Tetsuo Suga | Learning system of linguistic knowledge of natural language learning system and recording medium which records natural language learning program |
| TW201140559A (en) * | 2010-05-10 | 2011-11-16 | Univ Nat Cheng Kung | Method and system for identifying emotional voices |
| CN103226949A (en) * | 2011-09-30 | 2013-07-31 | 苹果公司 | Using context information to facilitate processing of commands in a virtual assistant |
| CN104346406A (en) * | 2013-08-08 | 2015-02-11 | 北大方正集团有限公司 | Training corpus expanding device and training corpus expanding method |
| CN104360994A (en) * | 2014-12-04 | 2015-02-18 | 科大讯飞股份有限公司 | Natural language understanding method and natural language understanding system |
| TW201516756A (en) * | 2013-10-28 | 2015-05-01 | Univ Kun Shan | Intelligent voice control system and method therefor |
| CN104778945A (en) * | 2005-08-05 | 2015-07-15 | 沃伊斯博克斯科技公司 | Systems and methods for responding to natural language speech utterance |
| CN106057200A (en) * | 2016-06-23 | 2016-10-26 | 广州亿程交通信息有限公司 | Semantic-based interaction system and interaction method |
| CN107015969A (en) * | 2017-05-19 | 2017-08-04 | 四川长虹电器股份有限公司 | Can self-renewing semantic understanding System and method for |
| WO2017212783A1 (en) * | 2016-06-09 | 2017-12-14 | 真由美 稲場 | Program for realizing function to assist in understanding personality and preferences of other party and communicating |
| CN107818781A (en) * | 2017-09-11 | 2018-03-20 | 远光软件股份有限公司 | Intelligent interactive method, equipment and storage medium |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103186523B (en) * | 2011-12-30 | 2017-05-10 | 富泰华工业（深圳）有限公司 | Electronic device and natural language analyzing method thereof |
| US10217059B2 (en) * | 2014-02-04 | 2019-02-26 | Maluuba Inc. | Method and system for generating natural language training data |
| US10402453B2 (en) * | 2014-06-27 | 2019-09-03 | Nuance Communications, Inc. | Utilizing large-scale knowledge graphs to support inference at scale and explanation generation |
| US9754207B2 (en) * | 2014-07-28 | 2017-09-05 | International Business Machines Corporation | Corpus quality analysis |
| KR102447513B1 (en) * | 2016-01-22 | 2022-09-27 | 한국전자통신연구원 | Self-learning based dialogue apparatus for incremental dialogue knowledge, and method thereof |
-
2018
- 2018-05-09 CN CN201810436639.2A patent/CN110489517B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104778945A (en) * | 2005-08-05 | 2015-07-15 | 沃伊斯博克斯科技公司 | Systems and methods for responding to natural language speech utterance |
| JP2007094291A (en) * | 2005-09-30 | 2007-04-12 | Tetsuo Suga | Learning system of linguistic knowledge of natural language learning system and recording medium which records natural language learning program |
| TW201140559A (en) * | 2010-05-10 | 2011-11-16 | Univ Nat Cheng Kung | Method and system for identifying emotional voices |
| CN103226949A (en) * | 2011-09-30 | 2013-07-31 | 苹果公司 | Using context information to facilitate processing of commands in a virtual assistant |
| CN104346406A (en) * | 2013-08-08 | 2015-02-11 | 北大方正集团有限公司 | Training corpus expanding device and training corpus expanding method |
| TW201516756A (en) * | 2013-10-28 | 2015-05-01 | Univ Kun Shan | Intelligent voice control system and method therefor |
| CN104360994A (en) * | 2014-12-04 | 2015-02-18 | 科大讯飞股份有限公司 | Natural language understanding method and natural language understanding system |
| WO2017212783A1 (en) * | 2016-06-09 | 2017-12-14 | 真由美 稲場 | Program for realizing function to assist in understanding personality and preferences of other party and communicating |
| CN106057200A (en) * | 2016-06-23 | 2016-10-26 | 广州亿程交通信息有限公司 | Semantic-based interaction system and interaction method |
| CN107015969A (en) * | 2017-05-19 | 2017-08-04 | 四川长虹电器股份有限公司 | Can self-renewing semantic understanding System and method for |
| CN107818781A (en) * | 2017-09-11 | 2018-03-20 | 远光软件股份有限公司 | Intelligent interactive method, equipment and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| 智能语音助手抢占AI入口市场;孟晋;《新经济导刊》;20170405(第04期);全文 * |
| 电子词典与词汇知识表达;陈克健;中文信息学报(第04期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110489517A (en) | 2019-11-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9575936B2 (en) | Word cloud display | |
| US9672490B2 (en) | Procurement system | |
| US11164564B2 (en) | Augmented intent and entity extraction using pattern recognition interstitial regular expressions | |
| US20150039292A1 (en) | Method and system of classification in a natural language user interface | |
| US20170364504A1 (en) | Method and system for data processing for real-time text analysis | |
| US20160132830A1 (en) | Multi-level score based title engine | |
| CN108038208B (en) | Training method and device of context information recognition model and storage medium | |
| US20220398598A1 (en) | Facilitating an automated, interactive, conversational troubleshooting dialog regarding a product support issue via a chatbot and associating product support cases with a newly identified issue category | |
| WO2020139865A1 (en) | Systems and methods for improved automated conversations | |
| CN110827797A (en) | Voice response event classification processing method and device | |
| US9910909B2 (en) | Method and apparatus for extracting journey of life attributes of a user from user interactions | |
| CA3153056A1 (en) | Intelligently questioning and answering method, device, computer, equipment and storage medium | |
| CN110489517B (en) | Automatic learning method and system of virtual assistant | |
| CN111190973A (en) | Method, device, equipment and storage medium for classifying statement forms | |
| US20240062016A1 (en) | Systems and Methods for Textual Classification Using Natural Language Understanding Machine Learning Models for Automating Business Processes | |
| JP2018156418A (en) | Correction device, correction method and correction program | |
| TWI674530B (en) | Method and system for operating a virtual assistant | |
| US20230351121A1 (en) | Method and system for generating conversation flows | |
| CN110209776B (en) | Method and system for operating virtual assistant | |
| TWI679548B (en) | Method and system for automated learning of a virtual assistant | |
| US20230178073A1 (en) | Systems and methods for parsing and correlating solicitation video content | |
| CN114546326A (en) | Virtual human sign language generation method and system | |
| CN110298372B (en) | Method and system for automatically training virtual assistant | |
| CN112115248A (en) | Method and system for extracting conversation strategy structure from conversation corpus | |
| TWI652587B (en) | Method and system for automatically training virtual assistant |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |