CN106663128A - Extended learning method of chat dialogue system and chat conversation system - Google Patents
Extended learning method of chat dialogue system and chat conversation system Download PDFInfo
- Publication number
- CN106663128A CN106663128A CN201680001735.3A CN201680001735A CN106663128A CN 106663128 A CN106663128 A CN 106663128A CN 201680001735 A CN201680001735 A CN 201680001735A CN 106663128 A CN106663128 A CN 106663128A
- Authority
- CN
- China
- Prior art keywords
- answer
- question answering
- learning method
- answering system
- chat conversations
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Abstract
The invention discloses an extended learning method of a chat dialogue system and a chat conversation system. The extended learning method comprises the following steps : the problem of the user input question-answering system, generating an answer question answering system; questions and answers to the user performs confidence judgment of relevance, judging whether the answer is correct; if the answer question answering system judges whether the error, the user is found by a correction module of the correct answer., and the problem and the correct answer to the user is added to the question answering system in the present invention can reduce the problem of the user cannot be recognized or recognition errors.
Description
Technical field
The present invention relates to computer technology, the expansion learning method of more particularly to a kind of chat conversations system and chat right
Telephone system.
Background technology
Man-machine interaction is interactional technology between the research people and computing device produced since computer is born. its mesh
Mark be make machine help people efficiently, it is comfortable, be safely completed mission requirements.
And wherein, automatic chatting conversational system is a kind of man-machine interactive system, by receive user natural language form
Input, provides respective feedback；But there is customer problem None- identified, the problem of identification mistake often in this kind of conversational system.
It should be noted that the introduction of technical background is intended merely to above the convenient technical scheme to the application carry out it is clear,
Complete explanation, and facilitate the understanding of those skilled in the art and illustrate.Can not be merely because these schemes be the application's
Background section is set forth and thinks that above-mentioned technical proposal is known to those skilled in the art.
The content of the invention
In view of the drawbacks described above of prior art, the technical problem to be solved there is provided a kind of reduction user
The expansion learning method and chat conversations system of the chat conversations system of problem None- identified or identification Problem-Error.
For achieving the above object, the invention provides a kind of expansion learning method of chat conversations system, including step：
Customer problem is input into question answering system, question answering system generates answer；
Confidence level judgement is carried out to customer problem and the degree of association of answer, judges whether correct to answer；
If judging question answering system erroneous answers, the correct option of customer problem is found by correcting module, and this is used
Family problem and correct option are added in question answering system.
Further, the customer problem is automatically generated by problem generation module.The presence of problem generation module, will
So that customer problem can be input in question answering system with substantial amounts of generation, a large number of users problem, be conducive to quickly expanding question and answer
System.
Further, the problem generation module is based on arbitrary in state machine, speech model and RNN neutral nets
Plant what is realized.Certainly, if conditions permit, arrange multiple corresponding to question answering system, and the problem realized based on different mechanisms
Generation module is also possible to help question answering system to carry out expansion study.Further, if judging to answer system answer correctly,
Then question answering system is strengthened, allows question answering system to confirm, lift the confidence level of the problem output.If answer is correct, improve
Corresponding confidence level, it is possible to reduce the resource occupation of confidence level judge module, as soon as possible into the question and answer judgement of next round, carries
The learning efficiency of high question answering system.
Further, it is described that problem is input into question answering system, according to problem system generate answer the step of before also include
Step：
Problem to generating is filtered, and removal meets pre-conditioned problem.For some insignificant problems, carry out
Filter, efficiency is expanded in the study that not only can improve question answering system, and in the man-machine interaction in question answering system later stage, reduce and use
Family carries out the situation of meaningless question and answer；For example, there is not the related keyword of question answering system in the customer problem of the generation, or
Obstructed up time, or the situation of other wrong and meaningless problems, it can be assumed that to be to meet pre-conditioned problem.
Further, it is described it is pre-conditioned be by algorithm arrange.It is pre-conditioned to be configured and change,
The different times demand different with different regions is adapted to, the applicability of the question answering system is improved.
Further, the correction module includes man-machine interactively unit or corrects unit automatically.For be judged as answer
The customer problem of mistake, we can be with by the way of manual correction, it would however also be possible to employ the mode corrected automatically, or even, can be with
By the way of manual correction and automatically correction combination, the correct option of customer problem is matched with preferably to provide, and is added
To in question answering system.
Further, the customer problem is input into by man-machine interaction unit.Customer problem can formally transported
Quickly generated and expanded using the problem generation module based on state machine, speech model and RNN neutral nets before row
Practise；During user is actually used, study expansion can also be gradually carried out after question answering system operation.
Present invention also offers a kind of chat conversations system of the expansion learning method using as described in the present invention is arbitrary, bag
Include：Question answering system, for receive user problem and provides answer；
Confidence level judge module, for carrying out confidence level judgement to customer problem and answer；
Correct module, for carrying out answer correction to the customer problem for being judged as erroneous answers, and by customer problem and
Correct option is added in question answering system.
Further, the chat conversations system also includes problem filtering module, for filtering to customer problem, goes
Except meeting pre-conditioned problem.The setting of problem filtering module, it is possible to reduce the resource occupation problem of meaningless problem, improves
Efficiency is expanded in the practicality of question answering system and study.
The invention has the beneficial effects as follows：The present invention is due to increased correction function so that question answering system is in the use to being input into
When family problem cannot feed back correct option, correct option can be corrected and be provided, ask so that question answering system learns the user
Topic is simultaneously added to corresponding correct option in question answering system, expands the question and answer storehouse of question answering system, it is to avoid the customer problem is again
The problem that still cannot correctly feed back during appearance, study progressively and expansion, by causing, question answering system is gradually perfect, and then reduces
Customer problem None- identified or identification mistake, the situation that the intention of user cannot be understood.
With reference to explanation hereinafter and accompanying drawing, the particular implementation of the application is disclose in detail, specify the original of the application
Reason can be in adopted mode.It should be understood that presently filed embodiment is not so limited in scope.In appended power
In the range of the spirit and terms that profit is required, presently filed embodiment includes many changes, modifications and equivalent.
The feature for describing for a kind of embodiment and/or illustrating can be in same or similar mode one or more
It is combined with the feature in other embodiment used in individual other embodiment, or substitute the feature in other embodiment.
It should be emphasized that term "comprises/comprising" refers to the presence of feature, one integral piece, step or component when using herein, but and
It is not excluded for the presence of one or more further features, one integral piece, step or component or additional.
Description of the drawings
Included accompanying drawing is used for providing being further understood from the embodiment of the present application, which constitutes of specification
Point, for illustrating presently filed embodiment, and come together to explain the principle of the application with word description.It should be evident that under
Accompanying drawing in the description of face is only some embodiments of the present application, for those of ordinary skill in the art, is not paying wound
On the premise of the property made is laborious, can be with according to these other accompanying drawings of accompanying drawings acquisition.In the accompanying drawings：
Fig. 1 is a kind of expansion learning method of chat conversations system；
Fig. 2 is a kind of chat conversations system.
Specific embodiment
In order that those skilled in the art more fully understand the technical scheme in the application, below in conjunction with the application reality
The accompanying drawing in example is applied, the technical scheme in the embodiment of the present application is clearly and completely described, it is clear that described enforcement
Example is only some embodiments of the present application, rather than the embodiment of whole.Based on the embodiment in the application, this area is common
All other embodiment that technical staff is obtained under the premise of creative work is not made, should all belong to the application protection
Scope.
Embodiment one：
Fig. 1 is a kind of expansion learning method of chat conversations system, and referring to Fig. 1, the expansion learning method includes step：
S1:Customer problem is input into question answering system, question answering system generates answer；
S2:Confidence level judgement is carried out to customer problem and the degree of association of answer, judges whether correct to answer；
S3:If judging question answering system erroneous answers, the correct option of customer problem is found by correcting module, and should
Customer problem and correct option are added in question answering system.
The invention has the beneficial effects as follows：The present invention is due to increased correction function so that question answering system is in the use to being input into
When family problem cannot feed back correct option, correct option can be corrected and be provided, ask so that question answering system learns the user
Topic is simultaneously added to corresponding correct option in question answering system, expands the question and answer storehouse of question answering system, it is to avoid the customer problem is again
The problem that still cannot correctly feed back during appearance, study progressively and expansion, by causing, question answering system is gradually perfect, and then reduces
Customer problem None- identified or identification mistake, the situation that the intention of user cannot be understood.
The present embodiment is preferred, and customer problem is automatically generated by problem generation module.The presence of problem generation module,
To cause the customer problem can be with substantial amounts of generation, a large number of users problem is input in question answering system, is conducive to quickly expanding and is asked
Answer system.
The present embodiment is preferred, and problem generation module is based on arbitrary in state machine, speech model and RNN neutral nets
It is a kind of to realize.Certainly, if conditions permit, arrange multiple corresponding to question answering system, and based on asking that different mechanisms are realized
Topic generation module is also possible to help question answering system to carry out expansion study.
The present embodiment is preferred, if judging to answer system answer correct, question answering system is strengthened, and allows question answering system
Confirm, lift the confidence level of the problem output.If answer is correct, corresponding confidence level is improved, it is possible to reduce confidence level is sentenced
The resource occupation of disconnected module, as soon as possible into the question and answer judgement of next round, improves the learning efficiency of question answering system.
The present embodiment is preferred, and problem is input into question answering system, the step of generate answer according to problem system before also wrap
Include step：
Problem to generating is filtered, and removal meets pre-conditioned problem.For some insignificant problems, carry out
Filter, efficiency is expanded in the study that not only can improve question answering system, and in the man-machine interaction in question answering system later stage, reduce and use
Family carries out the situation of meaningless question and answer；For example, there is not the related keyword of question answering system in the customer problem of the generation, or
Obstructed up time, or the situation of other wrong and meaningless problems, it can be assumed that to be to meet pre-conditioned problem.
The present embodiment is preferred, and pre-conditioned arranged by algorithm.It is pre-conditioned to be configured and more
Change, adapt to the different times demand different with different regions, improve the applicability of the question answering system.
The present embodiment is preferred, and correcting module includes that unit is corrected in man-machine interactively unit or networking.For being judged as back
The customer problem of mistake is answered, we can be with by the way of manual correction, it would however also be possible to employ the mode that networking is corrected, or even, can
By the way of correcting and combine using manual correction and networking, the correct option of customer problem is matched with preferably to provide, and is added
In being added to question answering system.
The present embodiment is preferred, and customer problem is input into by man-machine interaction unit.Customer problem can be formal
Quickly generated and expanded using based on the problem generation module of state machine, speech model and RNN neutral nets before operation
Study；During user is actually used, study expansion can also be gradually carried out after question answering system operation.
Embodiment two：
Fig. 2 is a kind of chat conversations system of the expansion learning method using as described in the present invention is arbitrary of the present invention, should be chatted
Its conversational system 100 includes：Question answering system 1, for receive user problem and provides answer；
Confidence level judge module 2, for carrying out confidence level judgement to customer problem and answer；
Correct module 3, for carrying out answer correction to the customer problem for being judged as erroneous answers, and by customer problem and
Correct option is added in question answering system.
The present embodiment is preferred, and chat conversations system also includes problem filtering module, for filtering to customer problem,
Removal meets pre-conditioned problem.The setting of problem filtering module, it is possible to reduce the resource occupation problem of meaningless problem, carries
Efficiency is expanded in the practicality of high question answering system and study.
The invention has the beneficial effects as follows：The present invention increased correction function so that question and answer system due to being provided with correction module
Unite when the customer problem to being input into cannot feed back correct option, correct option can be corrected and provide, so as to question and answer system
System learns the customer problem and corresponding correct option is added in question answering system, expands the question and answer storehouse of question answering system, it is to avoid
The problem that still cannot correctly feed back when the customer problem occurs again, study progressively and expansion, will cause question answering system by
It is gradually perfect, and then reduce customer problem None- identified or identification mistake, the situation that the intention of user cannot be understood.
The preferred embodiment of the present invention described in detail above.It should be appreciated that one of ordinary skill in the art without
Need creative work just can make many modifications and variations with design of the invention.Therefore, all technologies in the art
Personnel are available by logical analysis, reasoning, or a limited experiment on the basis of existing technology under this invention's idea
Technical scheme, all should be in the protection domain being defined in the patent claims.
Claims (10)
1. a kind of expansion learning method of chat conversations system, is characterized in that：Including step：
Customer problem is input into question answering system, question answering system generates answer；
Confidence level judgement is carried out to customer problem and the degree of association of answer, judges whether correct to answer；
If judging question answering system erroneous answers, the correct option of customer problem is found by correcting module, and the user is asked
Topic and correct option are added in question answering system.
2. the expansion learning method of chat conversations system as claimed in claim 1, is characterized in that：The customer problem is by asking
Topic generation module is automatically generated.
3. the expansion learning method of chat conversations system as claimed in claim 3, is characterized in that：The problem generation module is
Realized based on any in state machine, speech model and RNN neutral nets.
4. the expansion learning method of chat conversations system as claimed in claim 1, is characterized in that：If judging to answer system answer
Correctly, then question answering system strengthened, allows question answering system to confirm, lift the confidence level of problem output.
5. the expansion learning method of chat conversations system as claimed in claim 1, is characterized in that：It is described that problem is input into question and answer
System, according to problem system generate answer the step of before also include step：
Problem to generating is filtered, and removal meets pre-conditioned problem.
6. the expansion learning method of chat conversations system as claimed in claim 5, is characterized in that：It is described it is pre-conditioned be to pass through
What algorithm was arranged.
7. the expansion learning method of chat conversations system as claimed in claim 1, is characterized in that：The correction module includes people
Work interactive unit corrects unit automatically.
8. the expansion learning method of chat conversations system as claimed in claim 1, is characterized in that：The customer problem is to pass through
Man-machine interaction unit input.
9. a kind of chat conversations system of the expansion learning method using as described in claim 1-8 is arbitrary, is characterized in that：Bag
Include：Question answering system, for receive user problem and provides answer；
Confidence level judge module, for carrying out confidence level judgement to customer problem and answer；
Module is corrected, for carrying out answer correction to the customer problem for being judged as erroneous answers, and by customer problem and correctly
Answer is added in question answering system.
10. chat conversations system as claimed in claim 8, is characterized in that：The chat conversations system is also filtered including problem
Module, for filtering to customer problem, removal meets pre-conditioned problem.
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2016/087774 WO2018000282A1 (en) | 2016-06-29 | 2016-06-29 | Extended learning method of chat dialogue system and chat dialogue system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106663128A true CN106663128A (en) | 2017-05-10 |
Family
ID=58838104
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201680001735.3A Pending CN106663128A (en) | 2016-06-29 | 2016-06-29 | Extended learning method of chat dialogue system and chat conversation system |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN106663128A (en) |
| WO (1) | WO2018000282A1 (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107463301A (en) * | 2017-06-28 | 2017-12-12 | 北京百度网讯科技有限公司 | Conversational system construction method, device, equipment and computer-readable recording medium based on artificial intelligence |
| CN107562863A (en) * | 2017-08-30 | 2018-01-09 | 深圳狗尾草智能科技有限公司 | Chat robots reply automatic generation method and system |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109271495B (en) * | 2018-08-14 | 2023-02-17 | 创新先进技术有限公司 | Question-answer recognition effect detection method, device, equipment and readable storage medium |
| CN109741824B (en) * | 2018-12-21 | 2023-08-04 | 质直(上海)教育科技有限公司 | Medical inquiry method based on machine learning |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103744843A (en) * | 2013-12-25 | 2014-04-23 | 北京百度网讯科技有限公司 | Online voice translation method and device |
| CN104714954A (en) * | 2013-12-13 | 2015-06-17 | 中国电信股份有限公司 | Information searching method and system based on context understanding |
| CN105591882A (en) * | 2015-12-10 | 2016-05-18 | 北京中科汇联科技股份有限公司 | Method and system for mixed customer services of intelligent robots and human beings |
| US20160154889A1 (en) * | 2014-12-02 | 2016-06-02 | International Business Machines Corporation | Persona-Based Conversation |
| CN105677783A (en) * | 2015-12-31 | 2016-06-15 | 上海智臻智能网络科技股份有限公司 | Information processing method and device for intelligent question-answering system |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101257512A (en) * | 2008-02-02 | 2008-09-03 | 黄伟才 | Inquiry answer matching method used for inquiry answer system as well as inquiry answer method and system |
| CN102903008B (en) * | 2011-07-29 | 2016-05-18 | 国际商业机器公司 | For the method and system of computer question and answer |
| CN103279528A (en) * | 2013-05-31 | 2013-09-04 | 俞志晨 | Question-answering system and question-answering method based on man-machine integration |
| US9275115B2 (en) * | 2013-07-16 | 2016-03-01 | International Business Machines Corporation | Correlating corpus/corpora value from answered questions |
| CN104050256B (en) * | 2014-06-13 | 2017-05-24 | 西安蒜泥电子科技有限责任公司 | Initiative study-based questioning and answering method and questioning and answering system adopting initiative study-based questioning and answering method |
-
2016
- 2016-06-29 WO PCT/CN2016/087774 patent/WO2018000282A1/en active Application Filing
- 2016-06-29 CN CN201680001735.3A patent/CN106663128A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104714954A (en) * | 2013-12-13 | 2015-06-17 | 中国电信股份有限公司 | Information searching method and system based on context understanding |
| CN103744843A (en) * | 2013-12-25 | 2014-04-23 | 北京百度网讯科技有限公司 | Online voice translation method and device |
| US20160154889A1 (en) * | 2014-12-02 | 2016-06-02 | International Business Machines Corporation | Persona-Based Conversation |
| CN105591882A (en) * | 2015-12-10 | 2016-05-18 | 北京中科汇联科技股份有限公司 | Method and system for mixed customer services of intelligent robots and human beings |
| CN105677783A (en) * | 2015-12-31 | 2016-06-15 | 上海智臻智能网络科技股份有限公司 | Information processing method and device for intelligent question-answering system |
Non-Patent Citations (1)
| Title |
|---|
| 吴瑞红: "互动问答社区中回答可信性分析", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107463301A (en) * | 2017-06-28 | 2017-12-12 | 北京百度网讯科技有限公司 | Conversational system construction method, device, equipment and computer-readable recording medium based on artificial intelligence |
| CN107562863A (en) * | 2017-08-30 | 2018-01-09 | 深圳狗尾草智能科技有限公司 | Chat robots reply automatic generation method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2018000282A1 (en) | 2018-01-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106663128A (en) | Extended learning method of chat dialogue system and chat conversation system | |
| CN108000526B (en) | Dialogue interaction method and system for intelligent robot | |
| CN105447207B (en) | A kind of question and answer exchange method and system towards intelligent robot | |
| CN106096717B (en) | Information processing method towards intelligent robot and system | |
| CN105704013A (en) | Context-based topic updating data processing method and apparatus | |
| Griffin | Russell on the Nature of Logic (1903-1913) | |
| CN106815311A (en) | A kind of problem matching process and device | |
| CN109389870A (en) | A kind of data adaptive method of adjustment and its device applied in electronic instruction | |
| US11765104B2 (en) | Method and system for chatbot-enabled web forms and workflows | |
| CN109671434A (en) | A kind of speech ciphering equipment and self study audio recognition method | |
| CN108874935A (en) | A kind of review content recommendation method and electronic equipment based on phonetic search | |
| Duplessis et al. | Purely corpus-based automatic conversation authoring | |
| CN106776926A (en) | Improve the method and system of responsibility when robot talks with | |
| CN106708950B (en) | Data processing method and device for intelligent robot self-learning system | |
| CN105488370A (en) | Children mode unlock method and device | |
| Wagner et al. | Wittgenstein, Carnap, and Turing: Contrasting notions of analysis | |
| Mills et al. | Semantic negotiation in dialogue: the mechanisms of alignment | |
| CN109726002B (en) | Processing flow adjusting method and device | |
| CN106251717A (en) | Intelligent robot speech follow read learning method and device | |
| CN109934347B (en) | Device for expanding question-answer knowledge base | |
| Partee | Changing Notions of Linguistic Competence in the History | |
| CN106326208B (en) | A kind of system and method that robot is trained by voice | |
| Malchanau et al. | Multidimensional dialogue management for tutoring systems | |
| CN106649752A (en) | Answer acquisition method and device | |
| CN117493582B (en) | Model result output method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Address after: Room 301, Building 39, 239 Renmin Road, Gusu District, Suzhou City, Jiangsu Province, 215000 Applicant after: Suzhou Dogweed Intelligent Technology Co., Ltd. Address before: 518000 Dongfang Science and Technology Building 1307-09, 16 Keyuan Road, Yuehai Street, Nanshan District, Shenzhen City, Guangdong Province Applicant before: Shenzhen green bristlegrass intelligence Science and Technology Ltd. |
|
| CB02 | Change of applicant information | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170510 |
|
| WD01 | Invention patent application deemed withdrawn after publication |