CN101030267A - Automatic question-answering method and system - Google Patents
Automatic question-answering method and system Download PDFInfo
- Publication number
- CN101030267A CN101030267A CN 200610059919 CN200610059919A CN101030267A CN 101030267 A CN101030267 A CN 101030267A CN 200610059919 CN200610059919 CN 200610059919 CN 200610059919 A CN200610059919 A CN 200610059919A CN 101030267 A CN101030267 A CN 101030267A
- Authority
- CN
- China
- Prior art keywords
- participle
- speech
- word message
- information
- 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.)
- Granted
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
An automatically question-answering method includes dividing inputted writing information, carrying out seeking according to divided result, updating result set by utilizing matched seek-result, judging whether match of inputted writing information is finished or not and selecting the answer according to result set if said match is finished. The system used for realizing said method is also disclosed.
Description
Technical field
The present invention relates to a kind of computer application system and method, specifically, utilize the automatic question-answering method of language matching technique and the automatically request-answering system of language.
Background technology
In the existing language automatically request-answering system, be to adopt simple mode-matching technique to realize mostly, its method is first speech from sentence, the sentence of user's input and the sentence in the knowledge base are mated, if two identical couplings that just continue next speech of speech, some not too crucial speech that exist in the middle sentence that may utilize any speech asterisk wildcard to neglect user's input, repeat this process and finish, just pairing the replying of the sentence in the knowledge base returned to the user if the match is successful up to the sentence coupling of whole user's input.But for foreign language, Chinese has the advantages that word is flexible, sentence structure is complicated and changeable, and is not suitable for simple mode-matching technique.Existing Chinese automatically request-answering system is with reference to more external English automatically request-answering systems, adopt simple mode-matching technique to realize, this causes the coverage rate of Chinese automatically request-answering system ubiquity Chinese knowledge base narrow, the accuracy rate of system and all very low problem of rate of recalling, and user experience has been caused injury.
Automatically request-answering system claims QA (automatic Question Answering) system again, and it adopts natural language processing technique, finishes the analyzing and processing to customer problem on the one hand, finishes the generation of correct option on the other hand.Automatically request-answering system is a core with the natural language understanding technology, relates to multi-door subjects such as computational linguistics, information science and artificial intelligence, is one of focus of computer utility research.
Natural language understanding is an important research direction in the artificial intelligence field, and it makes computing machine can understand and use human natural language, can realize between people and the computing machine efficient communication based on natural language.
Knowledge base is the key components of automatically request-answering system, has stored a large amount of information with the right form of question and answer statement usually.When the natural language sentences of user input and the some sentences in the knowledge base when the match is successful, replying of its correspondence will be returned to the user.
Chinese word segmentation and part-of-speech tagging speech are the minimum independently significant language elements of activity.In Chinese, there is not separator between speech and the speech, speech itself also lacks tangible morphological markers, and therefore, the peculiar problem of Chinese information processing is exactly how the word string of Chinese to be divided into rational sequence of terms, i.e. Chinese word segmentation.Chinese word segmentation is the basis of deep level of processing such as syntactic analysis, also is the important step that mechanical translation, information retrieval and information extraction etc. are used.And part-of-speech tagging is exactly to the correct part of speech mark of each speech in the sentence according to the information in the sentence context.
The accuracy rate of automatically request-answering system is that automatically request-answering system is made the number of times of correctly replying divided by altogether response times.For example the user has imported 100 sentences to robot, and robot has made 100 times and replied, and it is correct wherein having 20 times, and the accuracy rate of this robot system is exactly 20% so.
The recall rate of automatically request-answering system is that automatically request-answering system is made the number of times of correctly replying divided by there being the number of times of correctly replying in the knowledge base.For example the user has imported 100 sentences to robot, robot has made 100 times and has replied, and it is correct wherein having 20 times, but in 100 sentences of user's input, that has only 25 sentences wherein in the knowledge base correctly replys existence, and the rate of recalling of this robot system is exactly 80% so.
Illustrate the shortcoming of the Chinese automatically request-answering system that adopts simple mode-matching technique realization below.
Suppose to exist in the knowledge base of automatically request-answering system following two groups of question and answer statements right, every group of natural language sentences (hereinafter to be referred as user's sentence) and systems response that all comprises user's input.
First group:
Are user's sentence: you born in Shenzhen?
Systems response: yes, how do you know?
Second group:
Are user's sentence: you born in Beijing?
Systems response: not right, I am born in Shenzhen.
When the user imports " you are born in Shenzhen? " or " you are born in Beijing? " the time, it all is correct replying.But when user input " you are born in Shanghai? ", automatically request-answering system just can't find user's sentence of coupling, thereby has returned wrong reply (may be replying of system default).But the systems response in fact, in second group is only correctly replying of user's input.
Because it is very many to replace the speech in " Shanghai ", so the problems referred to above also can't be by increasing more question and answer statement to solving.In addition, it is also infeasible that " Beijing " is replaced with any speech asterisk wildcard because the user may import " you were born in 76 years? ", the match is successful in meeting equally, causes replying makeing mistakes.
In sum, simple mode-matching technique also is not suitable for Chinese automatically request-answering system, causes the coverage rate of Chinese knowledge base narrow, and the accuracy rate of system and the rate of recalling are all very low, can damage user experience.
Summary of the invention
Technical matters solved by the invention provides a kind of automatically request-answering system, can improve the content coverage rate of Chinese knowledge base, the significant simultaneously accuracy rate of pattern match and the rate of recalling of improving.
Technical scheme of the present invention is as follows:
A kind of automatic question-answering method comprises:
(1) Word message with input carries out cutting;
(2) result according to cutting searches;
(3) refresh results set with the lookup result that is complementary;
(4) whether the Word message coupling of judging input is finished;
(5) finish when the Word message coupling of input, select to reply according to results set.
Preferably, store described stock's Word message in the inferenctial knowledge storehouse.
Preferably, in the described step (2), stock's Word message is handled through participle and part-of-speech tagging.
Preferably, described step (1) is specially, and Chinese word segmentation and part-of-speech tagging module are carried out cutting processing, the participle of output character information and part-of-speech tagging information to the Word message of input.
Preferably, step (2) is specially, and according to the participle and the part-of-speech tagging information of inputting word information, searches the stock's Word message with identical participle in results set; Step (4) is specially, and when inputting word information coupling is not finished, the next participle that extracts inputting word information is as searching foundation, and with results set as searching target, execution in step (2).
Preferably, step (2) is specially:
According to the participle and the part-of-speech tagging information of inputting word information, in results set, search stock's Word message with appointment part of speech asterisk wildcard; Step (4) is specially, and when inputting word information coupling is not finished, the next participle that extracts inputting word information is as searching foundation, and with results set as searching target, execution in step (2).
Preferably, step (2) is specially:
According to the participle and the part-of-speech tagging information of inputting word information, in results set, search stock's Word message of specifying any speech asterisk wildcard; Step (4) is specially, and when inputting word information coupling is not finished, the next participle that extracts inputting word information is as searching foundation, and with results set as searching target, execution in step (2).
Preferably, step (2) is specially:
According to the participle and the part-of-speech tagging information of inputting word information, search stock's Word message with identical participle, and according to the score value integration of setting; Simultaneously,, search stock's Word message of specifying the part of speech asterisk wildcard in results set according to the participle and the part-of-speech tagging information of inputting word information, and according to the score value integration of setting; Step (4) is specially, and when inputting word information coupling is not finished, the next participle that extracts Word message is as searching foundation, and with results set as searching target, execution in step (2).
Preferably, step (2) also comprises:
According to the participle and the part-of-speech tagging information of inputting word information, in results set, search stock's Word message of specifying any speech asterisk wildcard, and according to the score value integration of setting; Step (4) is specially, and when inputting word information coupling is not finished, the next participle that extracts Word message is as searching foundation, and with results set as searching target, execution in step (2).
Preferably, step (2) further comprises:
When not finding the stock's Word message that is complementary, select a conduct to reply in default the replying in the inferenctial knowledge storehouse at random, send and end.
Another technical scheme of the present invention is as follows:
A kind of automatically request-answering system comprises::
Network Interface Module receives the Word message of importing, and will reply transmission;
Participle and part-of-speech tagging module are carried out participle and part-of-speech tagging to the Word message of importing, and participle and the part-of-speech tagging information thereof that cuts out is sent;
Reasoning module according to replying of participle and part-of-speech tagging information searching correspondence thereof, will be replied and send to described Network Interface Module.
Preferably, participle and the part-of-speech tagging method in described participle and the part-of-speech tagging module invokes computational language knowledge base.
Preferably, also store the statistics and the dictionary of word frequency of utilization in the described computational language knowledge base.
Preferably, described reasoning module calls in the inferenctial knowledge storehouse and replys with Word message is corresponding.
Preferably, described inferenctial knowledge library storage is Word message in stock, and described stock's Word message is handled through participle or part-of-speech tagging.
Preferably, described stock's Word message comprises participle, any speech asterisk wildcard or part of speech asterisk wildcard.
The present invention can be applied to different language, the characteristics particularly flexible, that sentence structure is complicated and changeable at Chinese word, in automatically request-answering system, utilize Chinese part of speech asterisk wildcard can improve the content coverage rate of Chinese knowledge base, reduce the workload of building the storehouse, the significant simultaneously accuracy rate and the rate of recalling that improves based on the Chinese automatically request-answering system of pattern match, thereby lifting user's experience.
Description of drawings
Fig. 1 is the operational flowchart of automatic question-answering method;
Fig. 2 is the structural representation of automatically request-answering system.
Embodiment
Below with reference to Fig. 1, the automatic question-answering method of Chinese is described in detail.
The technical program has only provided a specific embodiment, can select different method for mode matching to use the part of speech asterisk wildcard during practical application.
Step S001: receiving end is received the Word message that the user imports.In this preferred embodiment, automatically request-answering system 100 receives the Word message of user's input by Network Interface Module 101.
Step S002: the Word message that receives is carried out cutting handle, export a series of speech and part-of-speech tagging information.In this preferred embodiment, Chinese word segmentation and part-of-speech tagging module 102 are called participle and the part-of-speech tagging method in the computational language knowledge base, the Word message that receives is carried out cutting handle the participle of output character information and part-of-speech tagging information.
Step S003: from these participles and part of speech information, take out current participle and part-of-speech tagging information as searching foundation.In this preferred embodiment, take out first participle and part-of-speech tagging information as searching foundation.
Step S004: search according to current participle and part-of-speech tagging information.
In this preferred embodiment, reasoning module 104, is searched in inferenctial knowledge storehouse 105 according to first participle and part-of-speech tagging information from first speech, and the result that will find combination as a result of.Inferenctial knowledge storehouse 105 internal memories contain stock's Word message of replying and passing through word segmentation processing, and this stock's Word message comprises participle, any speech asterisk wildcard or part of speech asterisk wildcard, and each participle can corresponding a plurality of replying.
The target of searching is to find stock's Word message of following three category features, and stock's Word message is corresponding replys with this:
The first, the stock's Word message in the inferenctial knowledge storehouse 105 is identical with first participle of the Word message of user's input at the participle of current location.Keep the score for stock's Word message of choosing, whenever choose once, with score value increase by 1 (initial value is 0) of this type of stock's Word message.
The second, stock's Word message the part of speech asterisk wildcard occurred in current location, and the specified part of speech of this part of speech asterisk wildcard is identical with the part of speech of the current participle of the Word message of user's input.Keep the score for stock's Word message of choosing, whenever choose once, with the score value increase by 0.5 of this type of user's sentence.
Three, stock's Word message any speech asterisk wildcard occurred in current location.Keep the score for stock's Word message of choosing, whenever choose once, with the score value increase by 0.2 of this type of user's sentence.
Above-mentioned three class match patterns can be chosen one arbitrarily, also can choose several the combination, as match pattern.In this preferred embodiment, above-mentioned three class match patterns are selected for use simultaneously, and will all put into results set according to stock's Word message that three class match patterns are chosen.
Among the present invention, the Word message (for example sentence) to user's input in inferenctial knowledge storehouse 105 has increased the part of speech asterisk wildcard, represents that all have the speech of specifying part of speech.Automatically request-answering system at first carries out participle and part-of-speech tagging after receiving the sentence that the user imports, and then hands to reasoning module 104.When the sentence of the 104 couples of users of reasoning module input and the user's sentence in the inferenctial knowledge storehouse 105 carry out pattern match, the part of speech asterisk wildcard can the match is successful with having any speech of specifying part of speech, but, if the speech of other user's sentences mates fully in speech in the sentence of user's input and the knowledge base, then the priority of part of speech asterisk wildcard is lower than the priority of coupling fully.Can significantly improve based on the accuracy rate of the Chinese automatically request-answering system of pattern match and the rate of recalling by this method
Step S005: if find stock's Word message that the Word message with user input is complementary and reply, then with these stock's Word messages with reply as current results set.Because the follow-up stock's literal letter that finds can constantly refresh a last results set, so results set can in time obtain upgrading.In this preferred embodiment, along with the carrying out of coupling and adding up of integration, the quantity of stock's Word message of this results set is constantly to dwindle, and therefore the accuracy of replying is constantly improving.
Also store default replying in the inferenctial knowledge storehouse 105, if above-mentioned searching all failed, then reasoning module 104 is thought in the inferenctial knowledge storehouse 105 not and the Word messages of user's input conform to replys, default replying can be called by system in inferenctial knowledge storehouse 105, select one (step S009) at random, return to user (step S010).
Step S006: whether the Word message of judging user input has mated finishes.Whether in this preferred embodiment, this step is carried out by reasoning module, finish so that in time judge coupling.
Step S007: if the Word message of user input not coupling do not finish, then extract next participle and part-of-speech tagging information as searching foundation, execution in step S004 continues above-mentioned search procedure, up to all the match is successful, perhaps it fails to match midway.In this preferred embodiment, reasoning module 104 thinks that coupling does not finish, and then carries out the n+1 operation, with next participle as the foundation of searching.
Step S008: if the Word message of user input has mated finish, then from results set, be complementary reply in one of picked at random, return to user (step S010).
In this preferred embodiment, reasoning module 104 is judged to have mated and is finished, and selects the highest the replying of integrated value to send to Network Interface Module 101 from results set, sends to the user by Network Interface Module 101.
Step S009: if in inferenctial knowledge storehouse 105, do not find stock's Word message of coupling, then reasoning module 104 will default the replying from inferenctial knowledge storehouse 105 in one of picked at random, as replying.
Step S010: replying of will receiving sends to the user.
In this preferred embodiment, Network Interface Module 101 receives replying that reasoning modules 104 send, and this is replied sends to the user.
Among the present invention, utilize the part of speech asterisk wildcard of Chinese to improve the content coverage rate in inferenctial knowledge storehouse 105, reduced the workload of building the storehouse, can improve simultaneously the accuracy rate and the rate of recalling significantly based on the automatic question-answering method of pattern match, thereby promoting user's experience, is a very significant innovation.
Example in the reference background technology, in the inferenctial knowledge of the present invention storehouse 105 of supporting Chinese part of speech asterisk wildcard, two groups of question and answer statements are right below having constructed:
First group:
Are the Word message of user's input: you born in Shenzhen?
Replying of system responses: yes, how do you know?
Second group:
Are the Word message of user's input: you born at POSnsPOS?
Replying of system responses: not right, I am born in Shenzhen.
Wherein POSnsPOS is the mode that adopts the part of speech asterisk wildcard to represent in the present embodiment, and wherein POS is the start-stop mark of part of speech information, and ns is the noun part of speech in expression orientation.
When the user imports " you are born in Shenzhen? " the time, the match is successful with user's sentence of first group, system to user's response " yes, how you know? "
When the user imports " you are born in Beijing? " or " you are born in Shanghai? " the time, all with second group in user's sentence the match is successful, " not right, I am born in Shenzhen to user response in system.”
In fact, as long as user input is speech similar Beijing and Shanghai, any ns of possessing part of speech, can with second group of question and answer statement to the match is successful; Similar but " 76 years " this speech does not possess the ns part of speech, so can be not that second group of question and answer statement is right by the mistake coupling.
Among the present invention, can also select different method for mode matching to use the part of speech asterisk wildcard, be used to improve based on the accuracy rate of pattern match and the rate of recalling, for example, the sentence to user's input does not pursue the coupling of speech, but the order of upsetting speech is directly mated.
Below with reference to Fig. 2 the preferred embodiments of the present invention are described in detail.
Different language has different grammers, and making has different match patterns between the speech.In this preferred embodiment, system selects for use Chinese as recognition objective.
Select for use the automatically request-answering system 100 of Chinese to comprise Network Interface Module 101, Chinese word segmentation and part-of-speech tagging module 102, reasoning module 104, and computational language knowledge base 103 and inferenctial knowledge storehouse 105.
Chinese word segmentation and part-of-speech tagging module 102 are called participle and the part-of-speech tagging method in the computational language knowledge base 103, Word message to user's input carries out Chinese word segmentation and part-of-speech tagging, and speech that all are cut out and part-of-speech tagging information thereof are submitted to reasoning module 104 then.
In this preferred embodiment, computational language knowledge base 103 stored be Chinese word segmentation and part-of-speech tagging information necessary, also comprise various statisticss such as dictionary and word frequency, this computational language knowledge base 103 can be upgraded according to actual needs, in time new participle and part-of-speech tagging method are mended into.
Claims (16)
1, a kind of automatic question-answering method comprises:
(1) Word message with input carries out cutting;
(2) result according to cutting searches;
(3) refresh results set with the lookup result that is complementary;
(4) whether the Word message coupling of judging input is finished;
(5) finish when the Word message coupling of input, select to reply according to results set.
2, automatic question-answering method according to claim 1 is characterized in that, stores described stock's Word message in the inferenctial knowledge storehouse.
3, automatic question-answering method according to claim 2 is characterized in that, in the described step (2), stock's Word message is handled through participle and part-of-speech tagging.
4, automatic question-answering method according to claim 1 is characterized in that, described step (1) is specially, and Chinese word segmentation and part-of-speech tagging module are carried out cutting processing, the participle of output character information and part-of-speech tagging information to the Word message of input.
5, automatic question-answering method according to claim 1 is characterized in that, step (2) is specially, and according to the participle and the part-of-speech tagging information of inputting word information, searches the stock's Word message with identical participle in results set; Step (4) is specially, and when inputting word information coupling is not finished, the next participle that extracts inputting word information is as searching foundation, and with results set as searching target, execution in step (2).
6, automatic question-answering method according to claim 1 is characterized in that, step (2) is specially:
According to the participle and the part-of-speech tagging information of inputting word information, in results set, search stock's Word message with appointment part of speech asterisk wildcard; Step (4) is specially, and when inputting word information coupling is not finished, the next participle that extracts inputting word information is as searching foundation, and with results set as searching target, execution in step (2).
7, automatic question-answering method according to claim 1 is characterized in that, step (2) is specially:
According to the participle and the part-of-speech tagging information of inputting word information, in results set, search stock's Word message of specifying any speech asterisk wildcard; Step (4) is specially, and when inputting word information coupling is not finished, the next participle that extracts inputting word information is as searching foundation, and with results set as searching target, execution in step (2).
8, automatic question-answering method according to claim 1 is characterized in that, step (2) is specially:
According to the participle and the part-of-speech tagging information of inputting word information, search stock's Word message with identical participle, and according to the score value integration of setting; Simultaneously,, search stock's Word message of specifying the part of speech asterisk wildcard in results set according to the participle and the part-of-speech tagging information of inputting word information, and according to the score value integration of setting; Step (4) is specially, and when inputting word information coupling is not finished, the next participle that extracts Word message is as searching foundation, and with results set as searching target, execution in step (2).
9, automatic question-answering method according to claim 8 is characterized in that, step (2) also comprises:
According to the participle and the part-of-speech tagging information of inputting word information, in results set, search stock's Word message of specifying any speech asterisk wildcard, and according to the score value integration of setting; Step (4) is specially, and when inputting word information coupling is not finished, the next participle that extracts Word message is as searching foundation, and with results set as searching target, execution in step (2).
10, automatic question-answering method according to claim 9 is characterized in that, step (2) further comprises:
When not finding the stock's Word message that is complementary, select a conduct to reply in default the replying in the inferenctial knowledge storehouse at random, send and end.
11, a kind of automatically request-answering system is characterized in that, comprising::
Network Interface Module receives the Word message of importing, and will reply transmission;
Participle and part-of-speech tagging module are carried out participle and part-of-speech tagging to the Word message of importing, and participle and the part-of-speech tagging information thereof that cuts out is sent;
Reasoning module according to replying of participle and part-of-speech tagging information searching correspondence thereof, will be replied and send to described Network Interface Module.
12, automatically request-answering system according to claim 11 is characterized in that, participle and part-of-speech tagging method in described participle and the part-of-speech tagging module invokes computational language knowledge base.
13, automatically request-answering system according to claim 12 is characterized in that, also stores the statistics and the dictionary of word frequency of utilization in the described computational language knowledge base.
14, automatically request-answering system according to claim 11 is characterized in that, described reasoning module calls in the inferenctial knowledge storehouse and replys with Word message is corresponding.
15, automatically request-answering system according to claim 14 is characterized in that, described inferenctial knowledge library storage is Word message in stock, and described stock's Word message is handled through participle or part-of-speech tagging.
16, automatically request-answering system according to claim 15 is characterized in that, described stock's Word message comprises participle, any speech asterisk wildcard or part of speech asterisk wildcard.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200610059919A CN100578539C (en) | 2006-02-28 | 2006-02-28 | Automatic question-answering method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200610059919A CN100578539C (en) | 2006-02-28 | 2006-02-28 | Automatic question-answering method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101030267A true CN101030267A (en) | 2007-09-05 |
| CN100578539C CN100578539C (en) | 2010-01-06 |
Family
ID=38715593
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200610059919A Active CN100578539C (en) | 2006-02-28 | 2006-02-28 | Automatic question-answering method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100578539C (en) |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101251862B (en) * | 2008-03-25 | 2010-06-09 | 北京百问百答网络技术有限公司 | Content-based problem automatic classifying method and system |
| CN101286161B (en) * | 2008-05-28 | 2010-10-06 | 华中科技大学 | Intelligent Chinese request-answering system based on concept |
| CN102203774A (en) * | 2008-11-03 | 2011-09-28 | 微软公司 | Retrieval using a generalized sentence collocation |
| CN101431573B (en) * | 2007-11-08 | 2013-02-20 | 上海撼世网络科技有限公司 | Method and equipment for implementing automatic customer service through human-machine interaction technology |
| CN103092928A (en) * | 2012-12-31 | 2013-05-08 | 安徽科大讯飞信息科技股份有限公司 | Voice query method and system |
| CN103116577A (en) * | 2013-02-04 | 2013-05-22 | 刘东民 | Method of intelligently processing user natural language order |
| CN104933084A (en) * | 2015-05-04 | 2015-09-23 | 上海智臻网络科技有限公司 | Method, apparatus and device for acquiring answer information |
| CN104933031A (en) * | 2015-06-29 | 2015-09-23 | 华东师范大学 | Unsupervised automatic Q&A method based on semantic web |
| CN105354300A (en) * | 2015-11-05 | 2016-02-24 | 上海智臻智能网络科技股份有限公司 | Information recommendation method and apparatus |
| CN105975531A (en) * | 2016-04-29 | 2016-09-28 | 华南师范大学 | Robot dialogue control method and system based on dialogue knowledge base |
| CN106407271A (en) * | 2016-08-26 | 2017-02-15 | 厦门快商通科技股份有限公司 | Intelligent customer service system, and method for updating intelligent customer service knowledge base of intelligent customer service system |
| CN106855873A (en) * | 2015-12-09 | 2017-06-16 | 财团法人工业技术研究院 | Network question-answering system, method and computer readable recording medium |
| CN107193948A (en) * | 2017-05-22 | 2017-09-22 | 邢加和 | Human-computer dialogue data analysing method and device |
| CN107798140A (en) * | 2017-11-23 | 2018-03-13 | 北京神州泰岳软件股份有限公司 | A kind of conversational system construction method, semantic controlled answer method and device |
| CN109426663A (en) * | 2017-08-29 | 2019-03-05 | 云量科技(北京)有限责任公司 | A kind of recognition methods and system of natural language |
| US10241716B2 (en) | 2017-06-30 | 2019-03-26 | Microsoft Technology Licensing, Llc | Global occupancy aggregator for global garbage collection scheduling |
| CN110347808A (en) * | 2019-05-28 | 2019-10-18 | 成都美美臣科技有限公司 | One e-commerce website intelligent robot customer service construction method |
| CN110457456A (en) * | 2019-07-25 | 2019-11-15 | 阿里巴巴集团控股有限公司 | A kind of chat robots answer method and device |
| CN109189897B (en) * | 2018-07-27 | 2020-07-31 | 什伯（上海）智能技术有限公司 | Chatting method and chatting device based on data content matching |
-
2006
- 2006-02-28 CN CN200610059919A patent/CN100578539C/en active Active
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101431573B (en) * | 2007-11-08 | 2013-02-20 | 上海撼世网络科技有限公司 | Method and equipment for implementing automatic customer service through human-machine interaction technology |
| CN101251862B (en) * | 2008-03-25 | 2010-06-09 | 北京百问百答网络技术有限公司 | Content-based problem automatic classifying method and system |
| CN101286161B (en) * | 2008-05-28 | 2010-10-06 | 华中科技大学 | Intelligent Chinese request-answering system based on concept |
| CN102203774A (en) * | 2008-11-03 | 2011-09-28 | 微软公司 | Retrieval using a generalized sentence collocation |
| CN103092928B (en) * | 2012-12-31 | 2015-12-23 | 安徽科大讯飞信息科技股份有限公司 | Voice inquiry method and system |
| CN103092928A (en) * | 2012-12-31 | 2013-05-08 | 安徽科大讯飞信息科技股份有限公司 | Voice query method and system |
| CN103116577A (en) * | 2013-02-04 | 2013-05-22 | 刘东民 | Method of intelligently processing user natural language order |
| CN104933084A (en) * | 2015-05-04 | 2015-09-23 | 上海智臻网络科技有限公司 | Method, apparatus and device for acquiring answer information |
| US10489435B2 (en) | 2015-05-04 | 2019-11-26 | Shanghai Xiaoi Robot Technology Co., Ltd. | Method, device and equipment for acquiring answer information |
| CN104933031A (en) * | 2015-06-29 | 2015-09-23 | 华东师范大学 | Unsupervised automatic Q&A method based on semantic web |
| CN104933031B (en) * | 2015-06-29 | 2018-11-06 | 华东师范大学 | A kind of automatic question-answering method unsupervised based on semantic net |
| CN105354300A (en) * | 2015-11-05 | 2016-02-24 | 上海智臻智能网络科技股份有限公司 | Information recommendation method and apparatus |
| CN105354300B (en) * | 2015-11-05 | 2019-04-05 | 上海智臻智能网络科技股份有限公司 | A kind of information recommendation method and device |
| CN106855873A (en) * | 2015-12-09 | 2017-06-16 | 财团法人工业技术研究院 | Network question-answering system, method and computer readable recording medium |
| CN106855873B (en) * | 2015-12-09 | 2020-07-31 | 财团法人工业技术研究院 | Network question-answering system, method and computer readable recording medium |
| US10713313B2 (en) | 2015-12-09 | 2020-07-14 | Industrial Technology Research Institute | Internet question answering system and method, and computer readable recording media |
| CN105975531B (en) * | 2016-04-29 | 2019-07-19 | 华南师范大学 | Robot dialog control method and system based on dialogue knowledge base |
| CN105975531A (en) * | 2016-04-29 | 2016-09-28 | 华南师范大学 | Robot dialogue control method and system based on dialogue knowledge base |
| CN106407271B (en) * | 2016-08-26 | 2020-02-14 | 厦门快商通科技股份有限公司 | Intelligent customer service system and updating method of intelligent customer service knowledge base thereof |
| CN106407271A (en) * | 2016-08-26 | 2017-02-15 | 厦门快商通科技股份有限公司 | Intelligent customer service system, and method for updating intelligent customer service knowledge base of intelligent customer service system |
| CN107193948A (en) * | 2017-05-22 | 2017-09-22 | 邢加和 | Human-computer dialogue data analysing method and device |
| CN107193948B (en) * | 2017-05-22 | 2018-04-20 | 邢加和 | Human-computer dialogue data analysing method and device |
| US10241716B2 (en) | 2017-06-30 | 2019-03-26 | Microsoft Technology Licensing, Llc | Global occupancy aggregator for global garbage collection scheduling |
| CN109426663A (en) * | 2017-08-29 | 2019-03-05 | 云量科技(北京)有限责任公司 | A kind of recognition methods and system of natural language |
| CN107798140A (en) * | 2017-11-23 | 2018-03-13 | 北京神州泰岳软件股份有限公司 | A kind of conversational system construction method, semantic controlled answer method and device |
| CN107798140B (en) * | 2017-11-23 | 2020-07-03 | 中科鼎富（北京）科技发展有限公司 | Dialog system construction method, semantic controlled response method and device |
| CN109189897B (en) * | 2018-07-27 | 2020-07-31 | 什伯（上海）智能技术有限公司 | Chatting method and chatting device based on data content matching |
| CN110347808A (en) * | 2019-05-28 | 2019-10-18 | 成都美美臣科技有限公司 | One e-commerce website intelligent robot customer service construction method |
| CN110457456A (en) * | 2019-07-25 | 2019-11-15 | 阿里巴巴集团控股有限公司 | A kind of chat robots answer method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN100578539C (en) | 2010-01-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101030267A (en) | Automatic question-answering method and system | |
| CN1253820C (en) | Device and method for intercrossing language information retrieval | |
| CN1207664C (en) | Error correcting method for voice identification result and voice identification system | |
| CN100565515C (en) | A kind of Chinese auto-answer method and system | |
| CN1492367A (en) | Inquire/response system and inquire/response method | |
| CN1193779A (en) | Method for dividing sentences in Chinese language into words and its use in error checking system for texts in Chinese language | |
| CN111310439B (en) | Intelligent semantic matching method and device based on depth feature dimension changing mechanism | |
| CN1912872A (en) | Method and system for abstracting new word | |
| CN101079028A (en) | On-line translation model selection method of statistic machine translation | |
| CN1942875A (en) | Dialogue supporting apparatus | |
| CN1977261A (en) | Method and system for word sequence processing | |
| CN1352774A (en) | System for Chinese tokenization and named entity recognition | |
| CN1834955A (en) | Multilingual translation memory, translation method, and translation program | |
| CN1692404A (en) | Lexical stress prediction | |
| CN1617134A (en) | System for identifying paraphrases using machine translation techniques | |
| CN1670723A (en) | Systems and methods for improved spell checking | |
| CN1910573A (en) | System for identifying and classifying denomination entity | |
| CN1601520A (en) | System and method for the recognition of organic chemical names in text documents | |
| CN1908935A (en) | Search method and system of a natural language | |
| CN1123432A (en) | Method for self-correction of grammar in machine translation | |
| CN1950819A (en) | A system and method for retrieving information and a system and method for storing information | |
| CN111160014B (en) | Intelligent word segmentation method | |
| CN1224954C (en) | Speech recognition device comprising language model having unchangeable and changeable syntactic block | |
| CN1256650C (en) | Chinese whole sentence input method | |
| CN1102779C (en) | Simplified Chinese character-the original complex form changingover apparatus |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |