CN109829052A - A kind of open dialogue method and system based on human-computer interaction - Google Patents
A kind of open dialogue method and system based on human-computer interaction Download PDFInfo
- Publication number
- CN109829052A CN109829052A CN201910121662.7A CN201910121662A CN109829052A CN 109829052 A CN109829052 A CN 109829052A CN 201910121662 A CN201910121662 A CN 201910121662A CN 109829052 A CN109829052 A CN 109829052A
- Authority
- CN
- China
- Prior art keywords
- answer
- corpus
- human
- document
- word
- 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
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
A kind of open dialogue method and system based on human-computer interaction is claimed in the present invention; user carries out voice input inquiry; search engine is retrieved using Lucene first after carrying out speech recognition, and retrieval content is that the retrieval pair of index is pre-established by corpus combing;Search result passes to mixed vector model, and mixed vector model generates model by trained vector in advance, expands answer, then the answer expanded is passed to order models;Order models pass through preparatory trained parameter, return the result.The present invention passes through the relevant technologies using natural language processing field, highlight semanteme and the syntactic analysis of problem and document databse, comprehensive and careful semantic analysis is compared to problem, and this methods and techniques are used in the analysis of sentence of document databse, establish On The Standardization typelib, the type of problem and the type of deserved answer have been standardized, retrieval time is shortened.
Description
Technical field
The present invention relates to artificial intelligence fields, more particularly to a kind of open dialogue method based on human-computer interaction and are
System.
Background technique
Since some time, the interactive mode between people and computer mostly rests on mouse and clicks Menu, keyboard
Input order, latest developments to touch screen etc..It will be clear that these interactive modes are not optimal selection, it is man-machine directly right
Words are very promising interactive modes.Many conversational systems based on experience are had recently emerged, people passes through voice or keyboard etc.
Mode inputs after natural language, and system can return to corresponding natural language response and execute corresponding operation.But in practice,
Existing conversational system currently entered to user can only be talked about and provide response, can not based on context provide and more accurately answer
It answers, while user needs the working condition of moment concern external equipment, makes rational management to equipment, this gives the use of user
Big inconvenience is brought, the quality of user experience is greatly reduced.
The universal and development of internet just constantly changes people's lives mode, promotes the rapid of artificial intelligence product
Development, more and more intellectual products continue to bring out, as intelligent robot, individual affair assistant, emotion are accompanied and attended to robot, apple
The intelligent Answer Systems such as Siri, the small ice of Microsoft bring a lot of convenience to people's lives.We can pass through apple Siri
It introduces dining room, inquiry weather conditions, voice setting alarm clock, arrange schedule and search data etc..It is inputted by natural language, Siri
Intelligent answer can be provided, and provide corresponding answer.Therefore it studies and improves intelligent answer and tie up to academia and industry
Boundary's value and significance all with higher.
The prior art completes the corresponding automatic test of artificial intelligence response system by the way of following: using being based on
The open source automated test tool of Google: selenium operates web page, completes to put question to and answer, the guarantor of input results
Depositing content only includes problem and answer, and next session operation is directly executed after the completion of execution, and result is recorded.Knot
It only includes problem and actual result that fruit, which records content,.The shortcomings that prior art:
(1) problem result can be because the content of context enters different scenes and does not enter into correctly so as to cause subproblem
Process flow, as a result accuracy is inadequate.(2) session data sequential organization is indefinite, lacks execution logic.(3) implementing result
Save it is not perfect, only by corresponding result export save, be not easy to carry out data statistics and maintenance.
Summary of the invention
The present invention provide a kind of user experience effect it is good, in conjunction with contextual analysis intelligent terminal on based on man-machine friendship
Mutual open dialogue method and system.In order to solve the above technical problems, the technical solution adopted by the present invention is that: a kind of intelligence is provided
Can artificial intelligence natural language operation system in terminal, including user and intelligent terminal, characterized in that the intelligent terminal with
User carries out the dialogue with context logic, completes user according to correct timing and arranges the program executed.
A kind of open dialogue method based on human-computer interaction is claimed in the present invention first, it is characterised in that:
User carries out voice input inquiry, and search engine is retrieved using Lucene first after carrying out speech recognition, and retrieval content is
The retrieval pair of index is pre-established by corpus combing;
Search result passes to mixed vector model, and mixed vector model generates model by trained vector in advance, expands
Answer, then the answer expanded is passed into order models;
Order models pass through preparatory trained parameter, return the result.
Further, the speech recognition, including user are known by voice first using the problem of natural language spoken language proposition
Optimal identification form is not thought by pretreatment, feature extraction and pattern-recognition formation system in part, and language acquisition problems are converted
For text question, for subsequent module analysis and handle;
Later by preconfigured correction parameter, output information classification and feedback for being determined according to the recognition result
Mode, controlling the correction module accordingly and inputting information correction is information after corresponding correction, and information after correction is further
It is converted into the feedback information of the output information classification, and feedback information is fed back into user.
Preferably, the problem that will have been marked when being retrieved using Lucene and answer are matched, and canonical is added
Canonical matching is realized in expression, combs by corpus, typical problem is extended to the template of multiple scaling problems;
Question template and answer are established index by Lucene respectively, and when query statement is input to this module, Lucene calls inspection
Suo Fangfa is inquired from index whether there is and matched answer.
Preferably, the retrieval content is that the retrieval pair of index is pre-established by corpus combing, comprising:
Before establishing index, corpus is pre-processed, including the duplicate document of removal, carries out word segmentation processing, information retrieval
Key in module is determination to document weight and is ranked up to document.
Further, the answer that will expand passes to order models, further comprises:
Text Pretreatment segments experimental data set, removes stop words and part-of-speech tagging;
It is filtered by part of speech, determines initial candidate word set, retain noun, verb, adjective and adverbial word therein as initial candidate
Keyword set;
The average information entropy of each word in initial candidate keyword set is calculated, the inverse of average information entropy is as the first of each vertex
Beginning weight simultaneously constructs TextRank graph model, iterates to calculate the weight of each word in initial candidate set;
Judge whether to restrain, if do not restrained, continues the weight for iterating to calculate each word in initial candidate set, if received
It holds back, is ranked up according to term weighing, export top n word.Preferably, the order models pass through preparatory trained parameter,
It returns the result, specifically includes:
Different feedback models is automatically selected according to different output information classifications and generates feedback information, and is sent to user's progress
It has been shown that, the output of the feedback information, including but not limited to text, audio output, video output, picture output, animation output;
Further, feedback text is obtained by inquiring text corpus;Backchannel is obtained by machine learning and artificial intelligence
Sound;Video database, which is inquired, by neuroid and artificial intelligence obtains feedback video;By natural language understanding processing and
Mapping software handles to obtain feedback picture;It handles to obtain feedback animation by natural language understanding processing and animation producing software.
A kind of open conversational system based on human-computer interaction, the opening based on human-computer interaction is also claimed in the present invention
The system of formula dialogue method is put for executing a kind of above-mentioned open dialogue method based on human-computer interaction, which is characterized in that also
Include:
Corpus library module, language analysis module and retrieval module;
Corpus library module is mainly by corpus management, the management of ontology class, corpus topology, corpus dimension, hot spot management, question and answer history
And part of speech manages seven submodule compositions;
Language analysis module is broadly divided into three pretreatment, session management and post-processing parts, and preprocessing part mainly has intelligence
It can segment, two submodules compositions of Entity recognition, session management is mainly provided with rhetorical question module, and post-processing is mainly provided with intelligent
Correction module;
Retrieving portion is retrieved by Lucene and two submodules that reorder form, wherein the technology used that reorders has sequence to learn
Habit, fuzzy matching, measuring similarity.
The present invention highlights the semanteme and language of problem and document databse by the relevant technologies using natural language processing field
Method analysis, comprehensive and careful semantic analysis is compared to problem, and this methods and techniques are used to the sentence of document databse
In son analysis, On The Standardization typelib is established, the type of problem and the type of deserved answer have been standardized, when shortening retrieval
Between.And joined common problem library in information retrieval and answer extracting stage, further shorten retrieval time.Experiment shows
It is this to apply the method in Chinese spoken interactive open field question answering system that greatly improve natural language processing technique to be
The performance of system.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention without any creative labor, may be used also for those of ordinary skill in the art
To obtain other drawings based on these drawings.
Fig. 1 is a kind of structural module diagram of the open conversational system based on human-computer interaction according to the present invention;
Fig. 2 is a kind of work flow diagram of the open dialogue method based on human-computer interaction according to the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Attached drawing 1 is a kind of structural module diagram of the open conversational system based on human-computer interaction according to the present invention;This
A kind of structural module diagram of open conversational system based on human-computer interaction is claimed it is characterised by comprising: corpus in invention
Library module, language analysis module and retrieval module;
Corpus library module is mainly by corpus management, the management of ontology class, corpus topology, corpus dimension, hot spot management, question and answer history
And part of speech manages seven submodule compositions;
Language analysis module is broadly divided into three pretreatment, session management and post-processing parts, and preprocessing part mainly has intelligence
It can segment, two submodules compositions of Entity recognition, session management is mainly provided with rhetorical question module, and post-processing is mainly provided with intelligent
Correction module;
Retrieving portion is retrieved by Lucene and two submodules that reorder form, wherein the technology used that reorders has sequence to learn
Habit, fuzzy matching, measuring similarity.
Corpus library module is embodied as an abstract class, and corpus management in corpus, the management of ontology class, corpus topology,
Seven corpus dimension, hot spot management, question and answer history and part of speech management submodules respectively correspond the subclass implemented.Son
Method in Similar integral abstract class, and specific method is realized inside oneself.
Wherein corpus class provides the related operation to database, typical problem, model answer, scaling problem, ontology class
And keyword is indicated in this module with specific data type.Corpus management submodule includes asking all kinds of in knowledge base
The increasing of topic and answer deletes, changes function, includes the operation of the creation to ontology class, deletion and modification, language in ontology class management
Material topology exhibits and knowledge dimension show the framework for relying on ontology class, and ontology class is shown to user's observation, question and answer history pipe
Reason management user query log, does not allow to manually add log, only allows to inquire, delete and modification log, part of speech management management
All keywords in system.Common problem needs the requirement of the policy information of real-time release itself according to administrative body, realizes
Permission user adds question and answer pair manually, and system will preferentially inquire question and answer pair in inquiry.
In the artificial intelligence markup language knowledge base of artificial intelligence natural language interaction body, the storage of state reservoir is current
State sends program in addition to providing response, and to program handling system to the natural language input of user, external equipment is allowed to execute
Corresponding operation, finally according to the state of current state reservoir and user's input or program handling system feedback and system
Jumping of doing well is done in response referring to the rule in artificial intelligence markup language knowledge base, can allow system in this way
With context logic, and user can be completed according to correct timing and arrange the program executed.Using artificial intelligence mark
When note language carries out writing knowledge base, use topic variable as state reservoir.
Attached drawing 2 is a kind of workflow of the open dialogue method based on human-computer interaction according to the present invention
Figure.
The open dialogue method based on human-computer interaction, it is characterised in that:
User carries out voice input inquiry, and search engine is retrieved using Lucene first after carrying out speech recognition, and retrieval content is
The retrieval pair of index is pre-established by corpus combing;
Search result passes to mixed vector model, and mixed vector model generates model by trained vector in advance, expands
Answer, then the answer expanded is passed into order models;
Order models pass through preparatory trained parameter, return the result.
Further, before user's progress voice input inquiry, further includes: generated newly by logging in browser
Session inputs problem to artificial intelligence response system, and the practical answer that crawl artificial intelligence response system is returned obtains
Practical answer and expected answer compare after comparing result, after having grabbed corresponding content, by corresponding expected answer
The practical answer actually obtained compares, in practical answer comprising expected answer or with expected answer complete one
The case where cause, then can to this session test result be labeled as Pass (expression passes through), if practical answer do not include or
It then can be Fail to this conversation test result queue that person is not equal to expected answer completely (expression does not pass through).
Further, the speech recognition, including user are known by voice first using the problem of natural language spoken language proposition
Optimal identification form is not thought by pretreatment, feature extraction and pattern-recognition formation system in part, and language acquisition problems are converted
For text question, for subsequent module analysis and handle;
Later by preconfigured correction parameter, output information classification and feedback for being determined according to the recognition result
Mode, controlling the correction module accordingly and inputting information correction is information after corresponding correction, and information after correction is further
It is converted into the feedback information of the output information classification, and feedback information is fed back into user.Business consultation is carried out in user
In the case where, type belonging to the business consultation can also be further determined that by machine learning algorithm model.In current attitude
The type of tendency is to occur in this dialogue to current attitude tendency type secondary in the case that negative attitude is inclined to type
Number adds up.For example, can be from the accumulation result that user's attitude is inclined in acquisition current session in caching, and update user
History attitude be inclined to record.
Belong to the adhesion value of each attitude tendency type according to the current attitude tendency that voice obtains user.For example, attitude is inclined
It may include the negative attitude tendency type such as anger, anxiety to type, can also include the positive attitude tendency such as happiness, calmness
Type.Adhesion value can be referred to as attitude tendency concentration, and adhesion value can be current attitude tendency and belong to each attitude tendency type
Probability is also possible to current attitude tendency for the grade of each attitude tendency type.The highest attitude of adhesion value is inclined to type
Type as current attitude tendency.For example, each attitude tendency type can be ranked up according to the adhesion value of acquisition, it will
The highest attitude tendency type of adhesion value is as candidate type.
Preferably, the problem that will have been marked when being retrieved using Lucene and answer are matched, and canonical is added
Canonical matching is realized in expression, combs by corpus, typical problem is extended to the template of multiple scaling problems;
Question template and answer are established index by Lucene respectively, and when query statement is input to this module, Lucene calls inspection
Suo Fangfa is inquired from index whether there is and matched answer.
Preferably, the retrieval content is that the retrieval pair of index is pre-established by corpus combing, comprising:
Before establishing index, corpus is pre-processed, including the duplicate document of removal, carries out word segmentation processing, information retrieval
Key in module is determination to document weight and is ranked up to document.The weight of document can
To be calculated according to following formula
Wherein:It is weight of i-th keyword in the case study stage that the document includes,It is the pass
The frequency that keyword occurs in this document,It is the inverse frequency that the keyword occurs in a document, D refers to keyword
Distribution density in a document.The frequency that keyword occurs in the document more it is high then it TF it is bigger, keyword more
Occur in more documents then its IDF with regard to smaller, otherwise bigger, the more concentration that keyword is distributed in this document, then D
It is worth bigger.TF*IDF value usually often occurs in a document from the significance level for reflecting the keyword on one side
The word of (TF is big), and the word (IDF is big) in other seldom present documents, the information content contained by the word is more, this word
Also more important.
Further, the search result passes to mixed vector model, and mixed vector model passes through trained in advance
Vector generates model, expands answer and includes:
If all words that all question sentences include in FAQ are, then each of FAQ question sentence is all
The vector that can be tieed up with a nTo indicate.Wherein,Calculation method are as follows:
If n isThe number occurred in this question sentence, m are to contain in FAQQuestion sentence number, M is question sentence in FAQ
Sum, then.Further, the answer that will expand passes to order models, further wraps
It includes:
Text Pretreatment segments experimental data set, removes stop words and part-of-speech tagging;
It is filtered by part of speech, determines initial candidate word set, retain noun, verb, adjective and adverbial word therein as initial candidate
Keyword set;
The average information entropy of each word in initial candidate keyword set is calculated, the inverse of average information entropy is as the first of each vertex
Beginning weight simultaneously constructs TextRank graph model, iterates to calculate the weight of each word in initial candidate set;
Judge whether to restrain, if do not restrained, continues the weight for iterating to calculate each word in initial candidate set, if received
It holds back, is ranked up according to term weighing, export top n word.Preferably, the order models pass through preparatory trained parameter,
It returns the result, specifically includes:
Different feedback models is automatically selected according to different output information classifications and generates feedback information, and is sent to user's progress
It has been shown that, the output of the feedback information, including but not limited to text, audio output, video output, picture output, animation output;
Further, feedback text is obtained by inquiring text corpus;Backchannel is obtained by machine learning and artificial intelligence
Sound;Video database, which is inquired, by neuroid and artificial intelligence obtains feedback video;By natural language understanding processing and
Mapping software handles to obtain feedback picture;It handles to obtain feedback animation by natural language understanding processing and animation producing software.
Program handling system is connect by wired or wireless sensor network with external equipment, according in program storage
The information that program feeds back external equipment judges, and issues instruction to external equipment and controls.Program handling system simultaneously
It is communicated in artificial intelligence natural language operation system with natural language interaction intelligent body, one side artificial intelligence nature language
Say that program is sent to program handling system by interactive intelligence body, program operating condition is sent to by another aspect program handling system
Natural language interaction intelligent body.
The above described is only a preferred embodiment of the present invention, be not intended to limit the present invention in any form, though
So the present invention has been disclosed as a preferred embodiment, and however, it is not intended to limit the invention, any technology people for being familiar with this profession
Member, without departing from the scope of the present invention, when the technology contents using the disclosure above are modified or are modified
It is right according to the technical essence of the invention for the equivalent embodiment of equivalent variations, but without departing from the technical solutions of the present invention
Any simple modification, equivalent change and modification made by above embodiments, all of which are still within the scope of the technical scheme of the invention.
Claims (8)
1. a kind of open dialogue method based on human-computer interaction, it is characterised in that:
User carries out voice input inquiry, and search engine is retrieved using Lucene first after carrying out speech recognition, and retrieval content is
The retrieval pair of index is pre-established by corpus combing;
Search result passes to mixed vector model, and mixed vector model generates model by trained vector in advance, expands
Answer, then the answer expanded is passed into order models;
Order models pass through preparatory trained parameter, return the result.
2. a kind of open dialogue method based on human-computer interaction as described in claim 1, it is characterised in that:
The speech recognition, including user are passed through in advance by speech recognition part first using the problem of natural language spoken language proposition
Reason, feature extraction and pattern-recognition form system and think optimal identification form, and language acquisition problems are converted to text question, for
Subsequent module analysis and processing;
Later by preconfigured correction parameter, output information classification and feedback for being determined according to the recognition result
Mode, controlling the correction module accordingly and inputting information correction is information after corresponding correction, and information after correction is further
It is converted into the feedback information of the output information classification, and feedback information is fed back into user.
3. a kind of open dialogue method based on human-computer interaction as described in claim 1, it is characterised in that:
The problem that will have been marked when being retrieved using Lucene and answer are matched, and regular expressions are added and realize canonical
Matching, combs by corpus, typical problem is extended to the template of multiple scaling problems;
Question template and answer are established index by Lucene respectively, and when query statement is input to this module, Lucene calls inspection
Suo Fangfa is inquired from index whether there is and matched answer.
4. a kind of open dialogue method based on human-computer interaction as described in claim 1, it is characterised in that:
The retrieval content is that the retrieval pair of index is pre-established by corpus combing, comprising:
Before establishing index, corpus is pre-processed, including the duplicate document of removal, carries out word segmentation processing, information retrieval
Key in module is determination to document weight and is ranked up to document;The weight of document can
To be calculated according to following formula
Wherein:It is weight of i-th keyword in the case study stage that the document includes,It is the pass
The frequency that keyword occurs in this document,It is the inverse frequency that the keyword occurs in a document, D refers to keyword
Distribution density in a document;
The frequency that keyword occurs in the document more it is high then it TF it is bigger, keyword occurs then in more documents
Its IDF is with regard to smaller, and on the contrary bigger, the more concentration that keyword is distributed in this document, then D value is bigger;TF*IDF value
From the significance level for reflecting the keyword on one side, often occur the word of (TF is big) usually in a document, and it is seldom
Word (IDF is big) in other present documents, the information content contained by the word is more, this word is also more important.
5. a kind of open dialogue method based on human-computer interaction as described in claim 1, it is characterised in that:
The search result passes to mixed vector model, and mixed vector model generates model by trained vector in advance,
Expanding answer includes:
If all words that all question sentences include in FAQ are, then each of FAQ question sentence all may be used
With the vector tieed up with a nTo indicate;
Wherein,Calculation method are as follows: set n asThe number occurred in this question sentence, m are to contain in FAQQuestion sentence
Number, M are the sum of question sentence in FAQ, then。
6. a kind of open dialogue method based on human-computer interaction as described in claim 1, it is characterised in that:
The answer that will expand passes to order models, further comprises:
Text Pretreatment segments experimental data set, removes stop words and part-of-speech tagging;
It is filtered by part of speech, determines initial candidate word set, retain noun, verb, adjective and adverbial word therein as initial candidate
Keyword set;
The average information entropy of each word in initial candidate keyword set is calculated, the inverse of average information entropy is as the first of each vertex
Beginning weight simultaneously constructs TextRank graph model, iterates to calculate the weight of each word in initial candidate set;
Judge whether to restrain, if do not restrained, continues the weight for iterating to calculate each word in initial candidate set, if received
It holds back, is ranked up according to term weighing, export top n word.
7. a kind of open dialogue method based on human-computer interaction as described in claim 1, it is characterised in that:
The order models pass through preparatory trained parameter, return the result, specifically include:
Different feedback models is automatically selected according to different output information classifications and generates feedback information, and is sent to user's progress
It has been shown that, the output of the feedback information, including but not limited to text, audio output, video output, picture output, animation output;
Further, feedback text is obtained by inquiring text corpus;Backchannel is obtained by machine learning and artificial intelligence
Sound;Video database, which is inquired, by neuroid and artificial intelligence obtains feedback video;By natural language understanding processing and
Mapping software handles to obtain feedback picture;It handles to obtain feedback animation by natural language understanding processing and animation producing software.
8. a kind of open conversational system based on human-computer interaction, the system of the open dialogue method based on human-computer interaction
For executing a kind of open dialogue method based on human-computer interaction such as any one of claim 1-8, which is characterized in that also wrap
It includes:
Corpus library module, language analysis module and retrieval module;
Corpus library module is mainly by corpus management, the management of ontology class, corpus topology, corpus dimension, hot spot management, question and answer history
And part of speech manages seven submodule compositions;
Language analysis module is broadly divided into three pretreatment, session management and post-processing parts, and preprocessing part mainly has intelligence
It can segment, two submodules compositions of Entity recognition, session management is mainly provided with rhetorical question module, and post-processing is mainly provided with intelligent
Correction module;
Retrieving portion is retrieved by Lucene and two submodules that reorder form, wherein the technology used that reorders has sequence to learn
Habit, fuzzy matching, measuring similarity.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910121662.7A CN109829052A (en) | 2019-02-19 | 2019-02-19 | A kind of open dialogue method and system based on human-computer interaction |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910121662.7A CN109829052A (en) | 2019-02-19 | 2019-02-19 | A kind of open dialogue method and system based on human-computer interaction |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109829052A true CN109829052A (en) | 2019-05-31 |
Family
ID=66863571
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910121662.7A Pending CN109829052A (en) | 2019-02-19 | 2019-02-19 | A kind of open dialogue method and system based on human-computer interaction |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109829052A (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110321472A (en) * | 2019-06-12 | 2019-10-11 | 中国电子科技集团公司第二十八研究所 | Public sentiment based on intelligent answer technology monitors system |
CN111125150A (en) * | 2019-12-26 | 2020-05-08 | 成都航天科工大数据研究院有限公司 | Industrial field question-answering system retrieval method |
CN111737435A (en) * | 2020-06-24 | 2020-10-02 | 全球能源互联网研究院有限公司 | Question-answer fault diagnosis assistant decision model construction and decision method and system |
CN112036156A (en) * | 2020-09-25 | 2020-12-04 | 北京小米松果电子有限公司 | Text dialogue method, text dialogue device and storage medium |
CN112380330A (en) * | 2020-11-13 | 2021-02-19 | 四川大学 | Training robot system and method under background of fine yin syndrome |
CN112380329A (en) * | 2020-11-13 | 2021-02-19 | 四川大学 | Training robot system and method under fine positive symptom background |
CN112380231A (en) * | 2020-11-13 | 2021-02-19 | 四川大学 | Training robot system and method with depressive disorder characteristics |
CN113033187A (en) * | 2019-12-25 | 2021-06-25 | 厦门铠甲网络股份有限公司 | Method for establishing iterative corpus |
CN113515616A (en) * | 2021-07-12 | 2021-10-19 | 中国电子科技集团公司第二十八研究所 | Task driving system based on natural language |
CN113656562A (en) * | 2020-11-27 | 2021-11-16 | 话媒(广州)科技有限公司 | Multi-round man-machine psychological interaction method and device |
CN114138755A (en) * | 2022-02-07 | 2022-03-04 | 中国建筑第五工程局有限公司 | Material archive retrieval method aiming at supplier cooperation based on artificial intelligence |
CN115048944A (en) * | 2022-08-16 | 2022-09-13 | 之江实验室 | Open domain dialogue reply method and system based on theme enhancement |
CN115292458A (en) * | 2022-06-29 | 2022-11-04 | 北京梦天门科技股份有限公司 | Investigation information input method and system and electronic equipment |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2008128423A1 (en) * | 2007-04-19 | 2008-10-30 | Shenzhen Institute Of Advanced Technology | An intelligent dialog system and a method for realization thereof |
CN103902652A (en) * | 2014-02-27 | 2014-07-02 | 深圳市智搜信息技术有限公司 | Automatic question-answering system |
CN104485036B (en) * | 2014-12-05 | 2018-08-10 | 沈阳理工大学 | A kind of automatic speech learning system |
CN108681574A (en) * | 2018-05-07 | 2018-10-19 | 中国科学院合肥物质科学研究院 | A kind of non-true class quiz answers selection method and system based on text snippet |
CN109271459A (en) * | 2018-09-18 | 2019-01-25 | 四川长虹电器股份有限公司 | Chat robots and its implementation based on Lucene and grammer networks |
-
2019
- 2019-02-19 CN CN201910121662.7A patent/CN109829052A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2008128423A1 (en) * | 2007-04-19 | 2008-10-30 | Shenzhen Institute Of Advanced Technology | An intelligent dialog system and a method for realization thereof |
CN103902652A (en) * | 2014-02-27 | 2014-07-02 | 深圳市智搜信息技术有限公司 | Automatic question-answering system |
CN104485036B (en) * | 2014-12-05 | 2018-08-10 | 沈阳理工大学 | A kind of automatic speech learning system |
CN108681574A (en) * | 2018-05-07 | 2018-10-19 | 中国科学院合肥物质科学研究院 | A kind of non-true class quiz answers selection method and system based on text snippet |
CN109271459A (en) * | 2018-09-18 | 2019-01-25 | 四川长虹电器股份有限公司 | Chat robots and its implementation based on Lucene and grammer networks |
Non-Patent Citations (2)
Title |
---|
秦兵等: "基于常问问题集的中文问答系统研究", 《哈尔滨工业大学学报》 * |
肖雯娟: "基于MATLAB的孤立词语音识别系统分析", 《电子制作》 * |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110321472A (en) * | 2019-06-12 | 2019-10-11 | 中国电子科技集团公司第二十八研究所 | Public sentiment based on intelligent answer technology monitors system |
CN113033187A (en) * | 2019-12-25 | 2021-06-25 | 厦门铠甲网络股份有限公司 | Method for establishing iterative corpus |
CN113033187B (en) * | 2019-12-25 | 2022-08-05 | 厦门铠甲网络股份有限公司 | Method for establishing iterative corpus |
CN111125150A (en) * | 2019-12-26 | 2020-05-08 | 成都航天科工大数据研究院有限公司 | Industrial field question-answering system retrieval method |
CN111125150B (en) * | 2019-12-26 | 2023-12-26 | 成都航天科工大数据研究院有限公司 | Search method for industrial field question-answering system |
CN111737435A (en) * | 2020-06-24 | 2020-10-02 | 全球能源互联网研究院有限公司 | Question-answer fault diagnosis assistant decision model construction and decision method and system |
CN112036156A (en) * | 2020-09-25 | 2020-12-04 | 北京小米松果电子有限公司 | Text dialogue method, text dialogue device and storage medium |
CN112380330A (en) * | 2020-11-13 | 2021-02-19 | 四川大学 | Training robot system and method under background of fine yin syndrome |
CN112380231A (en) * | 2020-11-13 | 2021-02-19 | 四川大学 | Training robot system and method with depressive disorder characteristics |
CN112380329A (en) * | 2020-11-13 | 2021-02-19 | 四川大学 | Training robot system and method under fine positive symptom background |
CN113656562A (en) * | 2020-11-27 | 2021-11-16 | 话媒(广州)科技有限公司 | Multi-round man-machine psychological interaction method and device |
CN113515616A (en) * | 2021-07-12 | 2021-10-19 | 中国电子科技集团公司第二十八研究所 | Task driving system based on natural language |
CN113515616B (en) * | 2021-07-12 | 2024-05-14 | 中国电子科技集团公司第二十八研究所 | Task driving system based on natural language |
CN114138755A (en) * | 2022-02-07 | 2022-03-04 | 中国建筑第五工程局有限公司 | Material archive retrieval method aiming at supplier cooperation based on artificial intelligence |
CN114138755B (en) * | 2022-02-07 | 2022-04-08 | 中国建筑第五工程局有限公司 | Material archive retrieval method aiming at supplier cooperation based on artificial intelligence |
CN115292458A (en) * | 2022-06-29 | 2022-11-04 | 北京梦天门科技股份有限公司 | Investigation information input method and system and electronic equipment |
CN115048944A (en) * | 2022-08-16 | 2022-09-13 | 之江实验室 | Open domain dialogue reply method and system based on theme enhancement |
CN115048944B (en) * | 2022-08-16 | 2022-12-20 | 之江实验室 | Open domain dialogue reply method and system based on theme enhancement |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109829052A (en) | A kind of open dialogue method and system based on human-computer interaction | |
CN108121829B (en) | Software defect-oriented domain knowledge graph automatic construction method | |
US11823074B2 (en) | Intelligent communication manager and summarizer | |
US10496756B2 (en) | Sentence creation system | |
WO2018153215A1 (en) | Method for automatically generating sentence sample with similar semantics | |
CN105868179B (en) | A kind of intelligent answer method and device | |
US11481387B2 (en) | Facet-based conversational search | |
CN105677795B (en) | Recommended method, recommendation apparatus and the recommender system of abstract semantics | |
CN111598702A (en) | Knowledge graph-based method for searching investment risk semantics | |
US11687826B2 (en) | Artificial intelligence (AI) based innovation data processing system | |
CN109460459A (en) | A kind of conversational system automatic optimization method based on log study | |
CN106326307A (en) | Language interaction method | |
WO2024011813A1 (en) | Text expansion method and apparatus, device, and medium | |
CN111858842A (en) | Judicial case screening method based on LDA topic model | |
CN108920599A (en) | A kind of the request-answer system answer precise positioning and abstracting method of knowledge based ontology library | |
CN112417846A (en) | Text automatic generation method and device, electronic equipment and storage medium | |
CN116821457B (en) | Intelligent consultation and public opinion processing system based on multi-mode large model | |
Bai et al. | Applied research of knowledge in the field of artificial intelligence in the intelligent retrieval of teaching resources | |
CN114707516A (en) | Long text semantic similarity calculation method based on contrast learning | |
CN114911893A (en) | Method and system for automatically constructing knowledge base based on knowledge graph | |
CN113779987A (en) | Event co-reference disambiguation method and system based on self-attention enhanced semantics | |
CN117194628A (en) | Compression technology-based prompt word optimization method, device, equipment and storage medium | |
CN115017271B (en) | Method and system for intelligently generating RPA flow component block | |
Liu et al. | The extension of domain ontology based on text clustering | |
Liu et al. | Keywords extraction method for technological demands of small and medium-sized enterprises based on LDA |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190531 |