CN106874441A - Intelligent answer method and apparatus - Google Patents
Intelligent answer method and apparatus Download PDFInfo
- Publication number
- CN106874441A CN106874441A CN201710066973.9A CN201710066973A CN106874441A CN 106874441 A CN106874441 A CN 106874441A CN 201710066973 A CN201710066973 A CN 201710066973A CN 106874441 A CN106874441 A CN 106874441A
- Authority
- CN
- China
- Prior art keywords
- answer
- question
- keywords
- confidence level
- answering system
- 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
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000004044 response Effects 0.000 claims abstract description 27
- 238000004458 analytical method Methods 0.000 claims description 10
- 239000000284 extract Substances 0.000 claims description 8
- 238000010586 diagram Methods 0.000 description 5
- 238000009825 accumulation Methods 0.000 description 3
- 238000003058 natural language processing Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Human Computer Interaction (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention relates to a kind of intelligent answer method and apparatus.Methods described includes:Obtain and wait to answer a question;By it is described wait to answer a question be sent respectively to the question answering system of question answering system and knowledge based storehouse based on frequently asked questions and corresponding answer;Obtain candidate answers and corresponding confidence level of the question answering system based on frequently asked questions and corresponding answer to the response to be answered a question, and candidate answers and corresponding confidence level of the question answering system to the response to be answered a question for obtaining the knowledge based storehouse;Highest confidence level in the confidence level is obtained, the highest confidence level is compared with believability threshold;If the highest confidence level is more than or equal to the believability threshold, the corresponding candidate answers of the highest confidence level are treated into corresponding answer of answering a question as described.The answer obtained based on two kinds of different question answering systems carries out Reliability ratio compared with the accuracy of the answer to be answered a question for obtaining is high.
Description
Technical field
The present invention relates to data processing field, more particularly to a kind of intelligent answer method and apparatus.
Background technology
It is automatic to answer FAQ (the Frequently Asked that intelligence system is normally based on restricted domain historical accumulation
Question, frequently asked questions and corresponding answer) data are built, and are limited to the completeness of FAQ data sets, and FAQ data are more, system energy
Enough problem typeses answered and quantity are also more, otherwise fewer.However, factor data accumulation deficiency or the neck without data accumulation
Domain, the answer accuracy for being given is relatively low.
The content of the invention
Based on this, it is necessary to for the inaccurate problem of traditional FAQ system answer problems, there is provided a kind of intelligent answer side
Method and device.
A kind of intelligent answer method, including:
Obtain and wait to answer a question;
By it is described wait to answer a question be sent respectively to based on the question answering system of frequently asked questions and corresponding answer and asking for knowledge based storehouse
Answer system;
The question answering system based on frequently asked questions and corresponding answer is obtained to the candidate answers of the response to be answered a question and right
The confidence level answered, and the question answering system in the knowledge based storehouse is obtained to the candidate answers of the response to be answered a question and right
The confidence level answered;
Highest confidence level in the confidence level is obtained, the highest confidence level is compared with believability threshold;
If the highest confidence level is more than or equal to the believability threshold, by the corresponding candidate of the highest confidence level
Answer waits corresponding answer of answering a question as described.
A kind of intelligent answer device, including:
Problem acquisition module, waits to answer a question for obtaining;
Sending module, for by the question answering system and base waited to answer a question and be sent respectively to based on frequently asked questions and corresponding answer
In the question answering system of knowledge base;
Candidate answers acquisition module, solution question and answer is treated for obtaining the question answering system based on frequently asked questions and corresponding answer to described
The candidate answers and corresponding confidence level of response are inscribed, and the question answering system in the acquisition knowledge based storehouse treats solution question and answer to described
Inscribe the candidate answers and corresponding confidence level of response;
Comparison module, for obtaining highest confidence level in the confidence level, by the highest confidence level and believability threshold
Compare;
Answer determining module, if being more than or equal to the believability threshold for the highest confidence level, by described in most
The corresponding candidate answers of confidence level high wait corresponding answer of answering a question as described.
Above-mentioned intelligent answer method and apparatus, the question and answer system being sent to based on frequently asked questions and corresponding answer by will wait to answer a question
System and the question answering system in knowledge based storehouse, get the candidate answers and correspondence of the question answering system feedback based on frequently asked questions and corresponding answer
Confidence level, and knowledge based storehouse question answering system feedback candidate answers and corresponding confidence level, filter out highest credible
Degree, if highest confidence level be more than or equal to believability threshold, using the corresponding candidate answers of highest confidence level as treat solution question and answer
The answer of topic, the answer obtained based on two kinds of different question answering systems carries out Reliability ratio compared with what is obtained to be answered a question answers
The accuracy of case is high.
Brief description of the drawings
Fig. 1 is the applied environment schematic diagram of intelligent answer method in one embodiment;
Fig. 2 is the internal structure schematic diagram of server in one embodiment;
Fig. 3 is the flow chart of intelligent answer method in one embodiment;
Fig. 4 is based on the question answering system of frequently asked questions and corresponding answer to the sound to be answered a question described in being obtained in one embodiment
The candidate answers and the flow chart of corresponding confidence level answered;
Fig. 5 is to obtain the time of the question answering system to the response to be answered a question in the knowledge based storehouse in one embodiment
Select the flow chart of answer and corresponding confidence level;
Fig. 6 is the structured flowchart of intelligent answer device in one embodiment;
Fig. 7 is the structured flowchart of intelligent answer device in one embodiment.
Specific embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
It is appreciated that term " first " used in the present invention, " second " etc. can be used to describe various elements herein,
But these elements should not be limited by these terms.These terms are only used for distinguishing first element and another element.Citing comes
Say, without departing from the scope of the invention, the first client can be referred to as the second client, and similarly, can be by
Second client is referred to as the first client.First client and the second client both client, but it is not same visitor
Family end.
Fig. 1 is the applied environment schematic diagram of intelligent answer method in one embodiment.As shown in figure 1, the applied environment bag
Include terminal 110 and server 120.Terminal 110 conversates with server 120 and communicates.Include session management on server 120
The question answering system of device, the question answering system based on frequently asked questions and corresponding answer and knowledge based storehouse.Session manager is used to obtain to wait to answer
Problem, will wait to answer a question is sent respectively to the question answering system of question answering system and knowledge based storehouse based on frequently asked questions and corresponding answer,
And obtain answer and the question and answer system in corresponding confidence level and knowledge based storehouse of the question answering system return based on frequently asked questions and corresponding answer
Answer and corresponding confidence level that system is returned, filter out highest confidence level, highest confidence level are compared with believability threshold, if greatly
In or equal to believability threshold, then using the corresponding answer of highest confidence level as the answer to be answered a question.
Fig. 2 is the internal structure schematic diagram of server (or high in the clouds etc.) in one embodiment.As shown in Fig. 2 the server
Including the processor, non-volatile memory medium, built-in storage and the network interface that are connected by system bus.Wherein, the service
The non-volatile memory medium of device is stored with operating system, database and intelligent answer device, is stored with database based on normal
See the question answering system of answer and the question answering system in knowledge based storehouse, the intelligent answer device is used to realize being applied to server
A kind of intelligent answer method.The processor of the server is used to provide calculating and control ability, supports the fortune of whole server
OK.The built-in storage of the server provides environment for the operation of the intelligent answer device in non-volatile memory medium, the internal memory
Computer-readable instruction can be stored in reservoir, when the computer-readable instruction is by the computing device, the place is may be such that
Reason device performs a kind of intelligent answer method.The network interface of the server is used for logical by network connection with outside terminal according to this
What letter, such as receiving terminal sent waits to answer a question and returns to answer etc. to terminal.Server can use independent server
Or multiple server groups into server cluster realize.It will be understood by those skilled in the art that the knot shown in Fig. 2
Structure, only the block diagram of the part-structure related to application scheme, does not constitute what application scheme was applied thereon
The restriction of server, specific server can include than more or less part shown in figure, or combine some parts,
Or arranged with different parts.
Fig. 3 is the flow chart of intelligent answer method in one embodiment.As shown in figure 3, a kind of intelligent answer method, bag
Include:
Step 302, obtains and waits to answer a question.
In the present embodiment, wait to answer a question refer to user consulting problem.Waiting to answer a question can be input into by web portal,
Or be input into by application App etc..Form to be answered a question can be at least one form such as voice, text, picture.
If providing webpage version consulting entrance, start webpage conversation window, problem to be answered is input into webpage conversation window.
If providing application program entry, start application program conversation window, be input into application program conversation window and wait to answer
Problem.
Step 304, by it is described wait to answer a question be sent respectively to question answering system based on frequently asked questions and corresponding answer and based on knowing
Know the question answering system in storehouse.
In the present embodiment, the question answering system based on frequently asked questions and corresponding answer refers to the question answering system based on FAQ.Common problem encountered is that
The number of times that finger is suggested exceedes the problem of frequency threshold value.Frequency threshold value can set as needed, such as 100 times, 10 inferior.It is common to ask
The key to exercises answer refer to FAQs answer.Question answering system based on frequently asked questions and corresponding answer refers to asking of being answered FAQs
Answer system.
Knowledge base refers to the structural knowledge set in restriction field.
Step 306, obtains candidate of the question answering system based on frequently asked questions and corresponding answer to the response to be answered a question
Answer and corresponding confidence level, and obtain the candidate of the question answering system to the response to be answered a question in the knowledge based storehouse
Answer and corresponding confidence level.
In the present embodiment, the question answering system based on frequently asked questions and corresponding answer is treated to answer a question and retrieved and lookup obtains right
The candidate answers answered, and calculate the confidence level of candidate answers.The question answering system in knowledge based storehouse is treated to answer a question carries out semanteme
Analysis, then to analysis after wait to answer a question match and obtain corresponding candidate answers, and calculate the confidence level of candidate answers.
The confidence level of the answer of the question answering system based on FAQ can be calculated Similarity value using the method for measuring similarity between text,
Similarity value is normalized between 0 to 1, used as confidence level, 1 is most credible.The answer of the question answering system in knowledge based storehouse
Confidence level, if there is answer in knowledge base, confidence level is 1, if not existing, confidence level is 0.
Step 308, obtains highest confidence level in the confidence level, and the highest confidence level is compared with believability threshold.
Step 310, if the highest confidence level is more than or equal to the believability threshold, by the highest confidence level pair
The candidate answers answered wait corresponding answer of answering a question as described.
In the present embodiment, the question answering system based on frequently asked questions and corresponding answer is to the candidate answers of the response to be answered a question
The question answering system in confidence level and knowledge based storehouse is compared to the confidence level of the candidate answers of the response to be answered a question, and obtains
Highest confidence level is taken, then highest confidence level and believability threshold are compared, if highest confidence level is more than or equal to confidence level
Threshold value, then using the corresponding candidate answers of highest believability threshold as answer to be answered a question.
Believability threshold refers to the minimum that confidence level needs to meet.Confidence level is more than or equal to believability threshold, then table
Show that answer is credible, otherwise answer is insincere.
Above-mentioned intelligent answer method, the question answering system and base being sent to based on frequently asked questions and corresponding answer by will wait to answer a question
In the question answering system of knowledge base, candidate answers of the question answering system feedback based on frequently asked questions and corresponding answer and corresponding credible are got
Degree, and knowledge based storehouse question answering system feedback candidate answers and corresponding confidence level, highest confidence level is filtered out, if most
Confidence level high is more than or equal to believability threshold, then the corresponding candidate answers of highest confidence level are answered as to be answered a question
Case, the answer obtained based on two kinds of different question answering systems carries out Reliability ratio compared with the standard of the answer to be answered a question for obtaining
True property is high.Additionally, for complex class problem, being retrieved by the question answering system based on frequently asked questions and corresponding answer and statistics can obtain correspondence
Answer, realize and quickly search corresponding answer, save manpower;For simple class problem, by asking for knowledge based storehouse
The accuracy of system is answered, more accurate answer is can obtain.Furthermore, above-mentioned intelligent answer method is effectively alleviated numerous limited
The problem that field FAQ problem sets are not enough, on the other hand effectively reduces the implementation complexity of the question answering system in knowledge based storehouse, drops
The cost of the low question answering system that knowledge base is built to complex class problem.
In one embodiment, above-mentioned intelligent answer method also includes:If the highest confidence level is less than the confidence level
Threshold value, then obtain artificial answer, and the artificial answer is treated into corresponding answer of answering a question as described.
In the present embodiment, highest confidence level is less than believability threshold, represents that candidate answers are all insincere, it is necessary to it is artificial to point out
Answer.Artificial answer is obtained as answer to be answered a question, to ensure to wait to answer a question having corresponding answer.
In one embodiment, above-mentioned intelligent answer method also includes:Wait described to answer a question and corresponding manually answer
Case is updated in the question answering system based on frequently asked questions and corresponding answer.
In the present embodiment, will wait to answer a question and corresponding artificial answer is updated to the question and answer system based on frequently asked questions and corresponding answer
In system, so as to run into next time it is same or similar can be answered by the question answering system of frequently asked questions and corresponding answer when answering a question,
Reduce cost of labor.
Fig. 4 is based on the question answering system of frequently asked questions and corresponding answer to the sound to be answered a question described in being obtained in one embodiment
The candidate answers and the flow chart of corresponding confidence level answered.As shown in figure 4, in one embodiment, being based on described in the acquisition
The question answering system of frequently asked questions and corresponding answer to the candidate answers and corresponding confidence level of the response to be answered a question, including:
Step 402, to it is described wait to answer a question carry out participle, extract keyword, and shape is extended to the keyword
Into the first set of keywords.
In the present embodiment, treating to answer a question carries out participle, extracts keyword, then based on restricted domain thesaurus to closing
Key word is extended and obtains the first set of keywords.It is " which model Magotan has " for example to wait to answer a question, and word segmentation result is " advanced in years
Rise ", " having ", " which ", " model", keyword extraction result is " Magotan ", " model ".Based on restricted domain thesaurus pair
Keyword is extended, and " model " is expanded to " model vehicle ", and the first set of keywords to be answered a question can be represented【Step
Rise model | vehicle】.
Step 404, participle is carried out by each problem in the question answering system based on frequently asked questions and corresponding answer, extracts crucial
Word, generates corresponding second set of keywords of each problem.
In the present embodiment, each problem in the question answering system based on frequently asked questions and corresponding answer is carried out into participle respectively, extracted
The keyword of problem, obtains the second each corresponding set of keywords of each problem.For example, will have in the question answering system of FAQ
" which vehicle Magotan has to problem", word segmentation result is " Magotan ", " having ", " which ", " vehicle", keyword extraction result is
" Magotan ", " vehicle ", the second set of keywords for obtaining is combined into【Magotan vehicle】.
Step 406, obtains the Similarity value of first set of keywords and each second set of keywords.
In the present embodiment, can be using Jie Kade distances, editing distance, Jie Kade distances and word frequency-reverse document-frequency power
The Similarity value for calculating the first set of keywords and the second set of keywords is waited again.
Step 408, chooses second set of keywords maximum with the Similarity value of first set of keywords corresponding
Answer, to the candidate answers of the response to be answered a question, and obtains institute as the question answering system based on frequently asked questions and corresponding answer
State the confidence level of candidate answers.
In the present embodiment, maximum corresponding second set of keywords of Similarity value is represented in existing second set of keywords
In it is most like with the first set of keywords.Can using Similarity value as candidate answers confidence level.
Participle is carried out by that will wait to answer a question and extracts keyword, and extend the first set of keywords of generation, and to base
Each problem in the question answering system of frequently asked questions and corresponding answer carries out participle and extracts corresponding second set of keywords of keyword generation
Close, calculate the Similarity value of the first set of keywords and the second set of keywords, choose the second maximum keyword of Similarity value
Gather corresponding answer as candidate answers, Similarity value as the confidence level of candidate answers is searched simple and convenient.
In one embodiment, the phase for obtaining first set of keywords and the second set of keywords each described
Like angle value, including:The Jie Kade distances of first set of keywords and each second set of keywords are obtained, according to institute
The Similarity value that Jie Kade distances obtain first set of keywords and each second set of keywords is stated, it is described similar
Angle value is directly proportional to the Jie Kade distances.
In the present embodiment, Jie Kade distances are, for weighing the similitude between two set, can be gathered using two
The element number that the element number of common factor is gathered and concentrated divided by two.Computing formula is as follows:
Wherein, A and B is set, and J (A, B) is Jie Kade distances.
For example, the first set of keywords is combined into【Magotan model | vehicle】, the second set of keywords is combined into【Magotan vehicle】, extension
Synonym afterwards may have multiple, also serve as an overall calculation.The common factor of the first set of keywords and the second set of keywords
In element number be 2, the first set of keywords and the second set of keywords and the element number concentrated is 2, then |【Step
Rise vehicle】|/|【Magotan vehicle】|=2/2.
Can be using the Jie Kade distances of the first set of keywords and the second set of keywords as the first set of keywords and the
The Similarity value of two set of keywords.Or, the first set of keywords can be multiplied with the Jie Kade distances of the second set of keywords
The result that will be obtained using positive number is used as the first set of keywords and the Similarity value of the second set of keywords.Jie Kade distances are got over
Greatly, Similarity value is bigger.
In one embodiment, the phase for obtaining first set of keywords and the second set of keywords each described
Like angle value, including:The Jie Kade distances of first set of keywords and each second set of keywords are obtained, and is obtained
The word frequency of each second set of keywords-reverse document-frequency weight is taken, according to the Jie Kade distances and word frequency-reverse
Document-frequency weight obtains the Similarity value of first set of keywords and each second set of keywords, described similar
Angle value is directly proportional to the product of the Jie Kade distances and word frequency-reverse document-frequency weight.
In the present embodiment, word frequency-reverse document-frequency weight refers to TF-IDF (term frequency-inverse
Document frequency) weight.Word frequency (term frequency, TF) refers to some given word in this document
The frequency of middle appearance.The word frequency of certain word can use word occurrence number hereof going out divided by all words hereof
Occurrence number sum is obtained.Reverse document-frequency (inverse document frequency, IDF) is that a word is generally important
The measurement of property.The IDF of a certain particular words, can by general act number divided by the file comprising the word number, then will
To business take the logarithm and obtain IDF values.The product for calculating TF and IDF obtains word frequency-reverse document-frequency weight.
Assuming that document is concentrated with N document, f (i, j) is the frequency (number of times) that lexical item i occurs in document j, then, word
Lexical item frequencies of the item i in document j can be defined as:
Formula is lexical item i normalized results in document j, wherein normalization by f (i, j) divided by going out in same document
The frequency of the most lexical item of occurrence number is calculated, therefore, TF values are be smaller than or equal to 1.
Assuming that document is concentrated with N document, if lexical item i occurs in n document, then IDF can be defined as:
But, when lexical item i does not occur in any document, formula above will occur the situation that denominator is zero,
So generally IDF is defined as:
Based on the definition of TF and IDF, scores of the lexical item i in document j can be defined as TF (i, j) * IDF (i) namely:
First set of keywords and the second set of keywords are obtained with word frequency-reverse document-frequency weight by Jie Kade distances
The Similarity value of conjunction, more accurately, the candidate answers for obtaining are more accurate.
In one embodiment, the phase for obtaining first set of keywords and the second set of keywords each described
Like angle value, including:The editing distance of first set of keywords and each second set of keywords is obtained, according to described
Editing distance obtains the Similarity value of first set of keywords and each second set of keywords, the Similarity value
It is inversely proportional with the editing distance.
In the present embodiment, editing distance refers to Levenshtein distances, refers between two word strings, to be changed into by one another
Minimum edit operation number of times needed for one.Using (string length-edit operation number of times most long)/string length most long obtain editing away from
From.Selection editing distance is smaller, and Similarity value is bigger, and score is higher.
Fig. 5 is to obtain the time of the question answering system to the response to be answered a question in the knowledge based storehouse in one embodiment
Select the flow chart of answer and corresponding confidence level.As shown in figure 5, in one embodiment, it is described to obtain the knowledge based storehouse
Question answering system to the candidate answers and corresponding confidence level of the response to be answered a question, including:
Step 502, to it is described wait to answer a question carry out semantic analysis generation question sentence vector.
In the present embodiment, multiple question templates can be pre-set, treating to answer a question using multiple question templates carries out language
Justice analysis generation question sentence vector.Question template for example,【Vehicle】*【Attribute】Which hasMatching problem such as the " appearance color of Magotan
Which has”【Vehicle】*【Attribute】It is how manyMatching problem is as " minimum ground clearance of Magotan is how many”.Question template can
Represented using regular expression.Problem can carry out template matches using canonical matching way.
In one embodiment, can be right using NLP (Natural Language Processing, natural language processing)
Waiting to answer a question carries out semantic analysis and obtains corresponding question sentence vector.
Step 504, query statement is converted into by the question sentence vector.
In the present embodiment, question sentence vector is converted into query statement sparQL.Based on vectorial sentence, groove fill method is used
Generation query statement.Generating sparQL sentences is:“selectvalue{<http://autohome/serie/ Magotans><
http://autohome/property/ vehicles>value}”.
Step 506, searches and the query statement pair according to the query statement from the question answering system in knowledge based storehouse
The Query Result answered.
In the present embodiment, query statement is the question answering system for meeting knowledge based storehouse.Can be from being based on according to query statement
Searched in the question answering system of knowledge base, if in the presence of finding corresponding Query Result.If not existing, return without result.
Step 508, using the Query Result as the knowledge based storehouse question answering system to the sound to be answered a question
The candidate answers answered, obtain the corresponding confidence level of the candidate answers.
If finding candidate answers, the corresponding confidence level of candidate answers is 1, if searching less than candidate answers, is based on
The question answering system of knowledge base is treated and answered a question without candidate answers.
Because the corresponding answer degree of accuracy of problem in the question answering system in knowledge based storehouse is high, by the question answering system of knowledge base
Corresponding answer is found, the accuracy of corresponding answer of waiting to answer a question is improve, for complex class problem, base can not be set up
It is cost-effective in the question answering system of knowledge base.
It should be noted that above-mentioned intelligent answer method, can be applied to the intelligent customer service system framework of all restricted domains,
Both service can be provided with BS structures (Browser/Server, Browser/Server Mode), it is also possible in input and output side
Speech recognition and TTS (Text To Speech, from text to language) is added to provide service with the communication networks such as phone.Tool
There is very strong transplantability, quickly can move to another restricted domain from a restricted domain.
Fig. 6 is the structured flowchart of intelligent answer device in one embodiment.As shown in fig. 6, a kind of intelligent answer device
600, run on server, including problem acquisition module 602, sending module 604, candidate answers acquisition module 606, compare mould
Block 608 and answer determining module 610.Wherein:
Problem acquisition module 602 is used to obtain to wait to answer a question.
In the present embodiment, wait to answer a question refer to user consulting problem.Waiting to answer a question can be input into by web portal,
Or be input into by application App etc..Form to be answered a question can be at least one form such as voice, text, picture.
Sending module 604 be used for by it is described wait to answer a question be sent respectively to question answering system based on frequently asked questions and corresponding answer and
The question answering system in knowledge based storehouse.
Candidate answers acquisition module 606 is waited to answer for obtaining the question answering system based on frequently asked questions and corresponding answer to described
The candidate answers and corresponding confidence level of problem response, and the question answering system in the acquisition knowledge based storehouse are waited to answer to described
The candidate answers and corresponding confidence level of problem response.
In the present embodiment, the question answering system based on frequently asked questions and corresponding answer is treated to answer a question and retrieved and lookup obtains right
The candidate answers answered, and calculate the confidence level of candidate answers.The question answering system in knowledge based storehouse is treated to answer a question carries out semanteme
Analysis, then to analysis after wait to answer a question match and obtain corresponding candidate answers, and calculate the confidence level of candidate answers.
The confidence level of the answer of the question answering system based on FAQ can be calculated Similarity value using the method for measuring similarity between text,
Similarity value is normalized between 0 to 1, used as confidence level, 1 is most credible.The answer of the question answering system in knowledge based storehouse
Confidence level, if there is answer in knowledge base, confidence level is 1, if not existing, confidence level is 0.
Comparison module 608 is used to obtain highest confidence level in the confidence level, by the highest confidence level and confidence level threshold
Value compares.
If answer determining module 610 is used for the highest confidence level is more than or equal to the believability threshold, will be described
The corresponding candidate answers of highest confidence level wait corresponding answer of answering a question as described.
Above-mentioned intelligent answer device, the question answering system and base being sent to based on frequently asked questions and corresponding answer by will wait to answer a question
In the question answering system of knowledge base, candidate answers of the question answering system feedback based on frequently asked questions and corresponding answer and corresponding credible are got
Degree, and knowledge based storehouse question answering system feedback candidate answers and corresponding confidence level, highest confidence level is filtered out, if most
Confidence level high is more than or equal to believability threshold, then the corresponding candidate answers of highest confidence level are answered as to be answered a question
Case, the answer obtained based on two kinds of different question answering systems carries out Reliability ratio compared with the standard of the answer to be answered a question for obtaining
True property is high.Additionally, for complex class problem, being retrieved by the question answering system based on frequently asked questions and corresponding answer and statistics can obtain correspondence
Answer, realize and quickly search corresponding answer, save manpower;For simple class problem, by asking for knowledge based storehouse
The accuracy of system is answered, more accurate answer is can obtain.Furthermore, above-mentioned intelligent answer method is effectively alleviated numerous limited
The problem that field FAQ problem sets are not enough, on the other hand effectively reduces the implementation complexity of the question answering system in knowledge based storehouse, drops
The cost of the low question answering system that knowledge base is built to complex class problem.
In one embodiment, if the answer determining module 610 is additionally operable to the highest confidence level less than described credible
Degree threshold value, then obtain artificial answer, and the artificial answer is treated into corresponding answer of answering a question as described.
Fig. 7 is the structured flowchart of intelligent answer device in one embodiment.As shown in fig. 7, a kind of intelligent answer device
600, run on server, except including problem acquisition module 602, sending module 604, candidate answers acquisition module 606, ratio
Compared with module 608 and answer determining module 610, also including update module 612.Wherein:
Update module 612 is used to wait to answer a question and corresponding artificial answer is updated to described based on FAQs by described
In the question answering system of answer.
In the present embodiment, will wait to answer a question and corresponding artificial answer is updated to the question and answer system based on frequently asked questions and corresponding answer
In system, so as to run into next time it is same or similar can be answered by the question answering system of frequently asked questions and corresponding answer when answering a question,
Reduce cost of labor.
In one embodiment, the candidate answers acquisition module 606 be additionally operable to it is described wait to answer a question carry out participle,
Keyword is extracted, and the keyword is extended to form the first set of keywords;By described based on frequently asked questions and corresponding answer
Each problem in question answering system carries out participle, extracts keyword, generates the second set of keywords corresponding to each problem;Obtain
Take the Similarity value of first set of keywords and each second set of keywords;And choose crucial with described first
The maximum corresponding answer of the second set of keywords of the Similarity value of word set is used as the question and answer based on frequently asked questions and corresponding answer
System obtains the confidence level of the candidate answers to the candidate answers of the response to be answered a question.
In one embodiment, the candidate answers acquisition module 606 be additionally operable to obtain first set of keywords with
The Jie Kade distances of each second set of keywords, according to the Jie Kade distances obtain first set of keywords with
The Similarity value of each second set of keywords, the Similarity value is directly proportional to the Jie Kade distances.
In one embodiment, the candidate answers acquisition module 606 be additionally operable to obtain first set of keywords with
The Jie Kade distances of each second set of keywords, and obtain the word frequency of each second set of keywords-reverse
Document-frequency weight, first set of keywords is obtained according to the Jie Kade distances and word frequency-reverse document-frequency weight
With the Similarity value of the second set of keywords each described, the Similarity value and the Jie Kade distances and word frequency-reverse text
The product of part frequency weight is directly proportional.
In one embodiment, the candidate answers acquisition module 606 be additionally operable to obtain first set of keywords with
The editing distance of each second set of keywords, according to the editing distance obtain first set of keywords and each
The Similarity value of second set of keywords, the Similarity value is inversely proportional with the editing distance.
In one embodiment, the candidate answers acquisition module 606 be additionally operable to it is described wait to answer a question carry out semanteme
Analysis generation question sentence vector;The question sentence vector is converted into query statement;According to the query statement from knowledge based storehouse
Query Result corresponding with the query statement is searched in question answering system;And using the Query Result as the knowledge based
The question answering system in storehouse obtains the corresponding confidence level of the candidate answers to the candidate answers of the response to be answered a question.
One of ordinary skill in the art will appreciate that all or part of flow in realizing above-described embodiment method, can be
The hardware of correlation is instructed to complete by computer program, described program can be stored in a non-volatile computer and can read
In storage medium, the program is upon execution, it may include such as the flow of the embodiment of above-mentioned each method.Wherein, described storage is situated between
Matter can be magnetic disc, CD, read-only memory (Read-Only Memory, ROM) etc..
Embodiment described above only expresses several embodiments of the invention, and its description is more specific and detailed, but simultaneously
Therefore the limitation to the scope of the claims of the present invention can not be interpreted as.It should be pointed out that for one of ordinary skill in the art
For, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to guarantor of the invention
Shield scope.Therefore, the protection domain of patent of the present invention should be determined by the appended claims.
Claims (12)
1. a kind of intelligent answer method, including:
Obtain and wait to answer a question;
By it is described wait to answer a question be sent respectively to the question and answer system of question answering system and knowledge based storehouse based on frequently asked questions and corresponding answer
System;
The question answering system based on frequently asked questions and corresponding answer is obtained to the candidate answers of the response to be answered a question and corresponding
Confidence level, and the question answering system in the knowledge based storehouse is obtained to the candidate answers of the response to be answered a question and corresponding
Confidence level;
Highest confidence level in the confidence level is obtained, the highest confidence level is compared with believability threshold;
If the highest confidence level is more than or equal to the believability threshold, by the corresponding candidate answers of the highest confidence level
Corresponding answer of answering a question is waited as described.
2. method according to claim 1, it is characterised in that methods described also includes:
If the highest confidence level is less than the believability threshold, artificial answer is obtained, using the artificial answer as described
Wait corresponding answer of answering a question.
3. method according to claim 2, it is characterised in that methods described also includes:
Wait to answer a question and corresponding artificial answer is updated in the question answering system based on frequently asked questions and corresponding answer by described.
4. according to the method in any one of claims 1 to 3, it is characterised in that FAQs is based on described in the acquisition
The question answering system of answer to the candidate answers and corresponding confidence level of the response to be answered a question, including:
To it is described wait to answer a question carry out participle, extract keyword, and the keyword is extended to form the first keyword
Set;
Each problem in the question answering system based on frequently asked questions and corresponding answer is carried out into participle, keyword is extracted, each is generated
The second set of keywords corresponding to problem;
Obtain the Similarity value of first set of keywords and each second set of keywords;And
Answer corresponding with the second set of keywords of the Similarity value maximum of first set of keywords is chosen as described
Question answering system based on frequently asked questions and corresponding answer obtains the candidate answers to the candidate answers of the response to be answered a question
Confidence level.
5. method according to claim 4, it is characterised in that acquisition first set of keywords with described in each
The Similarity value of the second set of keywords, including:
The Jie Kade distances of first set of keywords and each second set of keywords are obtained, according to the Jie Kade
Distance obtains the Similarity value of first set of keywords and each second set of keywords, the Similarity value and institute
Jie Kade distances are stated to be directly proportional;
Or, the Jie Kade distances of first set of keywords and each second set of keywords are obtained, and obtain
The word frequency of each second set of keywords-reverse document-frequency weight, according to the Jie Kade distances and word frequency-reverse text
Part frequency weight obtains the Similarity value of first set of keywords and each second set of keywords, the similarity
Value is directly proportional to the product of the Jie Kade distances and word frequency-reverse document-frequency weight;
Or, the editing distance of first set of keywords and each second set of keywords is obtained, according to the volume
Volume distance obtains the Similarity value of first set of keywords and each second set of keywords, the Similarity value and
The editing distance is inversely proportional.
6. according to the method in any one of claims 1 to 3, it is characterised in that the acquisition knowledge based storehouse
Question answering system to the candidate answers and corresponding confidence level of the response to be answered a question, including:
To it is described wait to answer a question carry out semantic analysis generation question sentence vector;
The question sentence vector is converted into query statement;
Query Result corresponding with the query statement is searched from the question answering system in knowledge based storehouse according to the query statement;
And
Using the Query Result as the knowledge based storehouse question answering system to the candidate answers of the response to be answered a question,
Obtain the corresponding confidence level of the candidate answers.
7. a kind of intelligent answer device, it is characterised in that including:
Problem acquisition module, waits to answer a question for obtaining;
Sending module, for by it is described wait to answer a question be sent respectively to question answering system based on frequently asked questions and corresponding answer and based on knowing
Know the question answering system in storehouse;
Candidate answers acquisition module, for obtaining the question answering system based on frequently asked questions and corresponding answer to the sound to be answered a question
The candidate answers and corresponding confidence level answered, and the question answering system in the knowledge based storehouse is obtained to the sound to be answered a question
The candidate answers and corresponding confidence level answered;
Comparison module, for obtaining highest confidence level in the confidence level, the highest confidence level is compared with believability threshold;
Answer determining module, if being more than or equal to the believability threshold for the highest confidence level, can by the highest
The corresponding candidate answers of reliability wait corresponding answer of answering a question as described.
8. device according to claim 7, it is characterised in that if to be additionally operable to the highest credible for the answer determining module
Degree is less than the believability threshold, then obtain artificial answer, using the artificial answer as it is described wait to answer a question it is corresponding
Answer.
9. device according to claim 8, it is characterised in that described device also includes:
Update module, for waiting to answer a question and corresponding artificial answer is updated to described based on frequently asked questions and corresponding answer by described
In question answering system.
10. the device according to any one of claim 7 to 9, it is characterised in that the candidate answers acquisition module is also used
In to it is described wait to answer a question carry out participle, extract keyword, and the keyword is extended to form the first set of keywords
Close;Each problem in the question answering system based on frequently asked questions and corresponding answer is carried out into participle, keyword is extracted, each is generated and is asked
The second corresponding set of keywords of topic;Obtain first set of keywords to each described second set of keywords it is similar
Angle value;And choose answer conduct corresponding with the second set of keywords of the Similarity value maximum of first set of keywords
The question answering system based on frequently asked questions and corresponding answer obtains the candidate and answers to the candidate answers of the response to be answered a question
The confidence level of case.
11. devices according to claim 10, it is characterised in that the candidate answers acquisition module is additionally operable to obtain described
The Jie Kade distances of the first set of keywords and each second set of keywords, according to the Jie Kade distances are obtained
The Similarity value of the first set of keywords and each second set of keywords, the Similarity value and the Jie Kade distances
It is directly proportional;
Or, the Jie Kade distances of first set of keywords and each second set of keywords are obtained, and obtain
The word frequency of each second set of keywords-reverse document-frequency weight, according to the Jie Kade distances and word frequency-reverse text
Part frequency weight obtains the Similarity value of first set of keywords and each second set of keywords, the similarity
Value is directly proportional to the product of the Jie Kade distances and word frequency-reverse document-frequency weight;
Or, the editing distance of first set of keywords and each second set of keywords is obtained, according to the volume
Volume distance obtains the Similarity value of first set of keywords and each second set of keywords, the Similarity value and
The editing distance is inversely proportional.
12. device according to any one of claim 7 to 9, it is characterised in that the candidate answers acquisition module is also used
In to it is described wait to answer a question carry out semantic analysis generation question sentence vector;The question sentence vector is converted into query statement;According to
The query statement searches Query Result corresponding with the query statement from the question answering system in knowledge based storehouse;And by institute
Query Result is stated as the question answering system in the knowledge based storehouse to the candidate answers of the response to be answered a question, obtains described
The corresponding confidence level of candidate answers.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710066973.9A CN106874441B (en) | 2017-02-07 | 2017-02-07 | Intelligent question-answering method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710066973.9A CN106874441B (en) | 2017-02-07 | 2017-02-07 | Intelligent question-answering method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106874441A true CN106874441A (en) | 2017-06-20 |
CN106874441B CN106874441B (en) | 2024-03-05 |
Family
ID=59167443
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710066973.9A Active CN106874441B (en) | 2017-02-07 | 2017-02-07 | Intelligent question-answering method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106874441B (en) |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108170780A (en) * | 2017-12-26 | 2018-06-15 | 北京邦邦共赢网络科技有限公司 | A kind of the problem of self-service question and answer matching process and device |
CN108932323A (en) * | 2018-06-29 | 2018-12-04 | 北京百度网讯科技有限公司 | Determination method, apparatus, server and the storage medium of entity answer |
CN108959421A (en) * | 2018-06-08 | 2018-12-07 | 三角兽(北京)科技有限公司 | Candidate replys evaluating apparatus and inquiry reverting equipment and its method, storage medium |
CN109063035A (en) * | 2018-07-16 | 2018-12-21 | 哈尔滨工业大学 | A kind of man-machine more wheel dialogue methods towards trip field |
CN109376298A (en) * | 2018-09-14 | 2019-02-22 | 广州神马移动信息科技有限公司 | Data processing method, device, terminal device and computer storage medium |
WO2019114512A1 (en) * | 2017-12-14 | 2019-06-20 | 株式会社日立制作所 | Method and apparatus for customer service, electronic device, and computer readable storage medium |
CN110019838A (en) * | 2017-12-25 | 2019-07-16 | 上海智臻智能网络科技股份有限公司 | Intelligent Answer System and intelligent terminal |
CN110209768A (en) * | 2019-05-31 | 2019-09-06 | 中国联合网络通信集团有限公司 | The problem of automatic question answering treating method and apparatus |
CN110263051A (en) * | 2019-06-11 | 2019-09-20 | 出门问问信息科技有限公司 | Question and answer for question answering system are to update method, device, equipment and storage medium |
CN110799970A (en) * | 2017-06-27 | 2020-02-14 | 华为技术有限公司 | Question-answering system and question-answering method |
CN111339254A (en) * | 2020-02-26 | 2020-06-26 | 常州市贝叶斯智能科技有限公司 | Intelligent voice processing method and device, intelligent equipment and medium |
CN111782794A (en) * | 2020-05-29 | 2020-10-16 | 北京沃东天骏信息技术有限公司 | Question-answer response method and device |
CN111984703A (en) * | 2020-08-19 | 2020-11-24 | 中国银行股份有限公司 | Method and device for positioning problems in knowledge base |
CN112800177A (en) * | 2020-12-31 | 2021-05-14 | 北京智源人工智能研究院 | FAQ knowledge base automatic generation method and device based on complex data types |
CN114238611A (en) * | 2021-12-23 | 2022-03-25 | 北京百度网讯科技有限公司 | Method, apparatus, device and storage medium for outputting information |
WO2023273170A1 (en) * | 2021-06-30 | 2023-01-05 | 同济人工智能研究院(苏州)有限公司 | Welcoming robot conversation method |
US20230055991A1 (en) * | 2021-08-09 | 2023-02-23 | Samsung Electronics Co., Ltd. | System and method for interactive dialogue |
CN116860951A (en) * | 2023-09-04 | 2023-10-10 | 贵州中昂科技有限公司 | Information consultation service management method and management system based on artificial intelligence |
Citations (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060206481A1 (en) * | 2005-03-14 | 2006-09-14 | Fuji Xerox Co., Ltd. | Question answering system, data search method, and computer program |
CN101695082A (en) * | 2009-09-30 | 2010-04-14 | 北京航空航天大学 | Service organization method based on relation mining and device thereof |
CN102262634A (en) * | 2010-05-24 | 2011-11-30 | 北京大学深圳研究生院 | Automatic questioning and answering method and system |
CN102760134A (en) * | 2011-04-28 | 2012-10-31 | 北京百度网讯科技有限公司 | Method and device for mining synonyms |
CN103177108A (en) * | 2013-03-26 | 2013-06-26 | 中山大学 | Medical treatment automatic question answering method based on internet and system thereof |
US20140006012A1 (en) * | 2012-07-02 | 2014-01-02 | Microsoft Corporation | Learning-Based Processing of Natural Language Questions |
CN104216913A (en) * | 2013-06-04 | 2014-12-17 | Sap欧洲公司 | Problem answering frame |
CN104573028A (en) * | 2015-01-14 | 2015-04-29 | 百度在线网络技术(北京)有限公司 | Intelligent question-answer implementing method and system |
CN104850539A (en) * | 2015-05-28 | 2015-08-19 | 宁波薄言信息技术有限公司 | Natural language understanding method and travel question-answering system based on same |
CN105022827A (en) * | 2015-07-23 | 2015-11-04 | 合肥工业大学 | Field subject-oriented Web news dynamic aggregation method |
CN105068661A (en) * | 2015-09-07 | 2015-11-18 | 百度在线网络技术(北京)有限公司 | Man-machine interaction method and system based on artificial intelligence |
US20150347587A1 (en) * | 2014-05-29 | 2015-12-03 | International Business Machines Corporation | Managing documents in question answering systems |
US20160110459A1 (en) * | 2014-10-18 | 2016-04-21 | International Business Machines Corporation | Realtime Ingestion via Multi-Corpus Knowledge Base with Weighting |
CN105760523A (en) * | 2016-02-29 | 2016-07-13 | 百度在线网络技术(北京)有限公司 | Information push method and information push device |
CN105843897A (en) * | 2016-03-23 | 2016-08-10 | 青岛海尔软件有限公司 | Vertical domain-oriented intelligent question and answer system |
CN105912645A (en) * | 2016-04-08 | 2016-08-31 | 上海智臻智能网络科技股份有限公司 | Intelligent question and answer method and apparatus |
KR101662450B1 (en) * | 2015-05-29 | 2016-10-05 | 포항공과대학교 산학협력단 | Multi-source hybrid question answering method and system thereof |
CN106055628A (en) * | 2016-05-27 | 2016-10-26 | 大连楼兰科技股份有限公司 | Intelligent communication method, device, system and application for automobile maintenance direction |
-
2017
- 2017-02-07 CN CN201710066973.9A patent/CN106874441B/en active Active
Patent Citations (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060206481A1 (en) * | 2005-03-14 | 2006-09-14 | Fuji Xerox Co., Ltd. | Question answering system, data search method, and computer program |
CN101695082A (en) * | 2009-09-30 | 2010-04-14 | 北京航空航天大学 | Service organization method based on relation mining and device thereof |
CN102262634A (en) * | 2010-05-24 | 2011-11-30 | 北京大学深圳研究生院 | Automatic questioning and answering method and system |
CN102760134A (en) * | 2011-04-28 | 2012-10-31 | 北京百度网讯科技有限公司 | Method and device for mining synonyms |
US20140006012A1 (en) * | 2012-07-02 | 2014-01-02 | Microsoft Corporation | Learning-Based Processing of Natural Language Questions |
CN103177108A (en) * | 2013-03-26 | 2013-06-26 | 中山大学 | Medical treatment automatic question answering method based on internet and system thereof |
CN104216913A (en) * | 2013-06-04 | 2014-12-17 | Sap欧洲公司 | Problem answering frame |
US20150347587A1 (en) * | 2014-05-29 | 2015-12-03 | International Business Machines Corporation | Managing documents in question answering systems |
US20160110459A1 (en) * | 2014-10-18 | 2016-04-21 | International Business Machines Corporation | Realtime Ingestion via Multi-Corpus Knowledge Base with Weighting |
CN104573028A (en) * | 2015-01-14 | 2015-04-29 | 百度在线网络技术(北京)有限公司 | Intelligent question-answer implementing method and system |
CN104850539A (en) * | 2015-05-28 | 2015-08-19 | 宁波薄言信息技术有限公司 | Natural language understanding method and travel question-answering system based on same |
KR101662450B1 (en) * | 2015-05-29 | 2016-10-05 | 포항공과대학교 산학협력단 | Multi-source hybrid question answering method and system thereof |
CN105022827A (en) * | 2015-07-23 | 2015-11-04 | 合肥工业大学 | Field subject-oriented Web news dynamic aggregation method |
CN105068661A (en) * | 2015-09-07 | 2015-11-18 | 百度在线网络技术(北京)有限公司 | Man-machine interaction method and system based on artificial intelligence |
CN105760523A (en) * | 2016-02-29 | 2016-07-13 | 百度在线网络技术(北京)有限公司 | Information push method and information push device |
CN105843897A (en) * | 2016-03-23 | 2016-08-10 | 青岛海尔软件有限公司 | Vertical domain-oriented intelligent question and answer system |
CN105912645A (en) * | 2016-04-08 | 2016-08-31 | 上海智臻智能网络科技股份有限公司 | Intelligent question and answer method and apparatus |
CN106055628A (en) * | 2016-05-27 | 2016-10-26 | 大连楼兰科技股份有限公司 | Intelligent communication method, device, system and application for automobile maintenance direction |
Non-Patent Citations (1)
Title |
---|
程贤禄 等: "《北京市农林科学院新品种、新技术、新产品科技成果汇编》", vol. 1, 西安电子科技大学出版社, pages: 335 - 336 * |
Cited By (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110799970A (en) * | 2017-06-27 | 2020-02-14 | 华为技术有限公司 | Question-answering system and question-answering method |
WO2019114512A1 (en) * | 2017-12-14 | 2019-06-20 | 株式会社日立制作所 | Method and apparatus for customer service, electronic device, and computer readable storage medium |
CN110019838A (en) * | 2017-12-25 | 2019-07-16 | 上海智臻智能网络科技股份有限公司 | Intelligent Answer System and intelligent terminal |
CN108170780A (en) * | 2017-12-26 | 2018-06-15 | 北京邦邦共赢网络科技有限公司 | A kind of the problem of self-service question and answer matching process and device |
CN108959421B (en) * | 2018-06-08 | 2021-04-13 | 腾讯科技(深圳)有限公司 | Candidate reply evaluation device, query reply device, method thereof, and storage medium |
CN108959421A (en) * | 2018-06-08 | 2018-12-07 | 三角兽(北京)科技有限公司 | Candidate replys evaluating apparatus and inquiry reverting equipment and its method, storage medium |
CN108932323A (en) * | 2018-06-29 | 2018-12-04 | 北京百度网讯科技有限公司 | Determination method, apparatus, server and the storage medium of entity answer |
CN109063035A (en) * | 2018-07-16 | 2018-12-21 | 哈尔滨工业大学 | A kind of man-machine more wheel dialogue methods towards trip field |
CN109063035B (en) * | 2018-07-16 | 2021-11-09 | 哈尔滨工业大学 | Man-machine multi-turn dialogue method for trip field |
CN109376298A (en) * | 2018-09-14 | 2019-02-22 | 广州神马移动信息科技有限公司 | Data processing method, device, terminal device and computer storage medium |
CN110209768B (en) * | 2019-05-31 | 2021-08-10 | 中国联合网络通信集团有限公司 | Question processing method and device for automatic question answering |
CN110209768A (en) * | 2019-05-31 | 2019-09-06 | 中国联合网络通信集团有限公司 | The problem of automatic question answering treating method and apparatus |
CN110263051A (en) * | 2019-06-11 | 2019-09-20 | 出门问问信息科技有限公司 | Question and answer for question answering system are to update method, device, equipment and storage medium |
CN111339254A (en) * | 2020-02-26 | 2020-06-26 | 常州市贝叶斯智能科技有限公司 | Intelligent voice processing method and device, intelligent equipment and medium |
CN111782794A (en) * | 2020-05-29 | 2020-10-16 | 北京沃东天骏信息技术有限公司 | Question-answer response method and device |
CN111984703A (en) * | 2020-08-19 | 2020-11-24 | 中国银行股份有限公司 | Method and device for positioning problems in knowledge base |
CN112800177A (en) * | 2020-12-31 | 2021-05-14 | 北京智源人工智能研究院 | FAQ knowledge base automatic generation method and device based on complex data types |
WO2023273170A1 (en) * | 2021-06-30 | 2023-01-05 | 同济人工智能研究院(苏州)有限公司 | Welcoming robot conversation method |
US20230055991A1 (en) * | 2021-08-09 | 2023-02-23 | Samsung Electronics Co., Ltd. | System and method for interactive dialogue |
CN114238611A (en) * | 2021-12-23 | 2022-03-25 | 北京百度网讯科技有限公司 | Method, apparatus, device and storage medium for outputting information |
CN116860951A (en) * | 2023-09-04 | 2023-10-10 | 贵州中昂科技有限公司 | Information consultation service management method and management system based on artificial intelligence |
CN116860951B (en) * | 2023-09-04 | 2023-11-14 | 贵州中昂科技有限公司 | Information consultation service management method and management system based on artificial intelligence |
Also Published As
Publication number | Publication date |
---|---|
CN106874441B (en) | 2024-03-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106874441A (en) | Intelligent answer method and apparatus | |
US10956464B2 (en) | Natural language question answering method and apparatus | |
US10997370B2 (en) | Hybrid classifier for assigning natural language processing (NLP) inputs to domains in real-time | |
US11017178B2 (en) | Methods, devices, and systems for constructing intelligent knowledge base | |
US11409813B2 (en) | Method and apparatus for mining general tag, server, and medium | |
US9183511B2 (en) | System and method for universal translating from natural language questions to structured queries | |
CN111241237B (en) | Intelligent question-answer data processing method and device based on operation and maintenance service | |
CN111931500B (en) | Search information processing method and device | |
KR20180078318A (en) | Methods and Apparatus for Determining the Agents | |
CN110619050B (en) | Intention recognition method and device | |
CN109829045A (en) | A kind of answering method and device | |
WO2018227930A1 (en) | Method and device for intelligently prompting answers | |
US20230094730A1 (en) | Model training method and method for human-machine interaction | |
US11699034B2 (en) | Hybrid artificial intelligence system for semi-automatic patent infringement analysis | |
CN108536665A (en) | A kind of method and device of determining sentence consistency | |
CN113342944B (en) | Corpus generalization method, apparatus, device and storage medium | |
KR101955920B1 (en) | Search method and apparatus using property language | |
CN114116997A (en) | Knowledge question answering method, knowledge question answering device, electronic equipment and storage medium | |
CN109472032A (en) | A kind of determination method, apparatus, server and the storage medium of entity relationship diagram | |
CN114391142A (en) | Parsing queries using structured and unstructured data | |
CN113792230B (en) | Service linking method, device, electronic equipment and storage medium | |
CN114861639B (en) | Question information generation method and device, electronic equipment and storage medium | |
CN116501841B (en) | Fuzzy query method, system and storage medium for data model | |
CN113971216B (en) | Data processing method and device, electronic equipment and memory | |
TW201824251A (en) | Global spoken dialogue system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |