CN106295807A - A kind of method and device of information processing - Google Patents
A kind of method and device of information processing Download PDFInfo
- Publication number
- CN106295807A CN106295807A CN201610710565.8A CN201610710565A CN106295807A CN 106295807 A CN106295807 A CN 106295807A CN 201610710565 A CN201610710565 A CN 201610710565A CN 106295807 A CN106295807 A CN 106295807A
- Authority
- CN
- China
- Prior art keywords
- sample
- asked
- standard
- question sentence
- extension
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
- G06N5/022—Knowledge engineering; Knowledge acquisition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
- G06F40/211—Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention relates to human-computer interaction technique field, particularly relate to the method and device of information processing in man-machine interaction.The present invention provides a kind of method of information processing, including: providing model sample storehouse, described model sample storehouse includes that sample standard asks and ask with each sample standard that corresponding sample extension is asked；Thering is provided knowledge base, described knowledge base includes that knowledge base standard asks and ask with each knowledge base standard that corresponding knowledge base extension is asked and answer, and described knowledge base is for furnishing an answer for user's question sentence；Determine that whether there is the sample extension matched with the user's question sentence in man-machine interaction daily record in described model sample storehouse asks；If existing, it is determined that the corresponding standard of user's question sentence described in described man-machine interaction daily record asks that the corresponding sample standard asked with the sample extension mated is asked the most identical；If differing, then optimize described knowledge base.The present invention also provides for device and the system of information processing corresponding to the method for a kind of and above-mentioned information processing.
Description
Technical field
The present invention relates to human-computer interaction technique field, particularly relate to the method and device of information processing in man-machine interaction.
Background technology
Man-machine interaction is the science of the interactive relation between Study system and user.System can be various machine
Device, it is also possible to be computerized system and software.Such as, various artificial intelligence system, example can be realized by man-machine interaction
As, intelligent customer service system, speech control system etc..
Artificial intelligence's semantics recognition is the basis of man-machine interaction, and human language can be identified by it, to be converted into machine
Device it will be appreciated that language.In order to understand human language, artificial intelligence's semantics recognition system needs a set of knowledge base.Magnanimity
Isomeric data is organized into knowledge by knowledge learning system, and is dissolved into existing knowledge hierarchy.
The original question sentence that various artificial intelligence systems use artificial intelligence's semantics recognition technology to propose user processes,
Determine the standard question sentence that this original question sentence is corresponding, more limited based on more incidental in this standard question sentence and original question sentence
Information provide corresponding answer, in artificial intelligence system, record the place for each original question sentence with the form of daily record
Reason situation, the information of each daily record comprises: user propose original question sentence (user's question sentence), standard question sentence (standard is asked) and answer
Case.
Knowledge base it is optimized, to comprise two important steps: picked out by the interactive log that needs optimize；For
Knowledge base is optimized by select daily record.
In prior art, when selecting interactive log, mainly by collect manually and sort out correct daily record storehouse and
Meaningless daily record storehouse, then contrasts with interactive log every day, filters the log content mated completely.Each daily record
All need artificial contrast, need to put into a large amount of hand labor.Meanwhile, when knowledge base is optimized by needs, it is also desirable to specialty
Knowledge operation maintenance personnel need to optimize daily record for every and carry out standard and ask and write, and is costly and inefficient down.
Summary of the invention
It is an object of the invention to provide the method and device of a kind of information processing, overcome present in conventional art with
Lower problem: need to put into a large amount of hand labor and select the interactive log that need to optimize.Meanwhile, when information processing, system can be automatically
Proposed standard is asked, further reduces the input of hand labor, improves the optimization efficiency of knowledge base.
According to above-mentioned purpose, the present invention provides a kind of method of information processing, including: model sample storehouse, described mould are provided
Pattern example storehouse includes that sample standard asks and ask with each sample standard that corresponding sample extension is asked；Knowledge base is provided, described
Knowledge base includes that knowledge base standard asks and ask with each knowledge base standard that the extension of corresponding knowledge base is asked and answer, described in know
Know storehouse for furnishing an answer for user's question sentence；Determine in described model sample storehouse and whether exist and the user in man-machine interaction daily record
The sample extension that question sentence matches is asked；If existing, it is determined that the corresponding mark of user's question sentence described in described man-machine interaction daily record
Standard asks that the corresponding sample standard asked with the sample extension mated is asked the most identical；If differing, then optimize described knowledge base.
In one embodiment, the extension of described sample is asked and is included that knowledge base extension is asked, described sample standard is asked and included knowledge base
Standard is asked.
In one embodiment, determine in described model sample storehouse that whether there is the sample matched with described user's question sentence expands
Exhibition is asked and is included: ask that execution Semantic Similarity Measurement to determine in described model sample storehouse is by described user's question sentence and sample extension
At least one semantic similarity with described user's question sentence of no existence is asked more than the sample extension of first threshold.
In one embodiment, determine that the corresponding standard of described user's question sentence is asked corresponding with what the sample extension mated was asked
Sample standard is asked the most identical including: the institute that relatively the corresponding standard of described user's question sentence is asked with the sample extension mated is asked is right
Sample standard is answered to ask that word is the most completely the same.
In one embodiment, if existing and more than described first threshold and being less than with described user's question sentence semantic similarity
The sample extension of 100% is asked, and the corresponding standard of described user's question sentence ask with semantic similarity more than described first threshold and
The corresponding sample standard question sentence that sample extension less than 100% is asked is identical, then by described user's question sentence and described user's question sentence
Corresponding standard ask and be added into described model sample storehouse explicitly.
In one embodiment, if the sample extension that there is multiple coupling is asked, it is determined that the corresponding mark of described user's question sentence
Standard is asked and is asked with the corresponding sample standard asked of sample extension mated and the most identical comprise determining whether a sample mated
The corresponding sample standard that extension is asked asks that the corresponding standard with described user's question sentence is asked identical.
In one embodiment, the optimization to described knowledge base includes: result based on described Semantic Similarity Measurement, it is recommended that
The corresponding sample standard asked more than the sample extension of Second Threshold with the semantic similarity of described user's question sentence is asked；Will be from being pushed away
The sample standard recommended is asked the sample standard that middle artificial selection goes out to ask with described user's question sentence and is added into described knowledge base explicitly.
In one embodiment, described method also includes: ask what middle artificial selection went out by described from the sample standard recommended
Sample standard is asked and is added into described model sample storehouse explicitly with described user's question sentence.
In one embodiment, if described model sample storehouse does not exist the sample extension matched with described user's question sentence
Ask, then creating the knowledge point corresponding with described user's question sentence in knowledge base, described knowledge point includes: knowledge base standard is asked, known
Know storehouse extension to ask and answer.
In one embodiment, described method also includes: the knowledge point created in knowledge base is added simultaneously to described mould
Pattern example storehouse.
In one embodiment, ask that execution Semantic Similarity Measurement includes by described user's question sentence and sample extension: to sample
Extension is asked and is carried out participle, and calculates word and sentence vector value；Described user's question sentence is carried out participle, and calculates word and sentence vector
Value；Calculate sample and extend the word and sentence vector value and the word of described user's question sentence and the degree of association of sentence vector value asked, with
Go out described user's question sentence and extend, with sample, the semantic similarity asked.
In one embodiment, whether there is, in determining described model sample storehouse, the sample matched with described user's question sentence
Before extension is asked, described method also includes: user's question sentences all in described man-machine interaction daily record are carried out pretreatment, to filter people
Invalid data in machine interactive log user's question sentence.
The present invention also provides for the device of a kind of information processing, including: first analyzes module, is used for determining in model sample storehouse
Whether there is the sample extension matched with the user's question sentence in man-machine interaction daily record to ask；Second analyze module, in response to
There is the sample extension matched with described user's question sentence to ask, it is determined that the institute of user's question sentence described in described man-machine interaction daily record
Corresponding standard asks that the corresponding sample standard asked with the sample extension mated is asked the most identical；And optimization module, it is used for responding
Corresponding standard in described user's question sentence is asked that the corresponding sample standard asked with the sample extension mated is asked and is differed, then optimize
Knowledge base.
In one embodiment, described first analyzes module includes: Semantic Similarity Measurement module, for described user being asked
Sentence asks execution Semantic Similarity Measurement, to determine in described model sample storehouse whether there is at least one with described with sample extension
The semantic similarity of user's question sentence is asked more than the sample extension of first threshold.
In one embodiment, described second analyzes module includes: comparison module, right for the institute of relatively described user's question sentence
The corresponding sample standard asked with the sample extension mated asks that word is the most completely the same to answer standard to ask.
In one embodiment, described second analyzes module also includes: add module, in response to existing and described user
Question sentence semantic similarity is asked more than described first threshold and the extension of the sample less than 100%, and described user's question sentence is corresponding
Standard is asked and is extended, more than described first threshold and the sample less than 100%, the corresponding sample standard question sentence asked with semantic similarity
Identical, then the corresponding standard of described user's question sentence and described user's question sentence is asked and be added into described model sample explicitly
Storehouse.
In one embodiment, if the sample extension that there is multiple coupling is asked, the most described second analyzes module determines whether
The corresponding sample standard that the sample extension of one coupling is asked asks that the corresponding standard with described user's question sentence is asked identical.
In one embodiment, described optimization module includes: recommending module, for knot based on described Semantic Similarity Measurement
Really, it is recommended that the corresponding sample standard asked more than the sample extension of Second Threshold with the semantic matching degree of described user's question sentence is asked；
And interpolation module, for relevant to described user's question sentence by asking that the standard that middle artificial selection goes out is asked from the sample standard recommended
It is added into described knowledge base connection.
In one embodiment, described interpolation module is further used for manually selecting described from the sample standard recommended is asked
The standard selected out is asked and is added into described model sample storehouse explicitly with described user's question sentence.
In one embodiment, if described model sample storehouse does not exist the sample extension matched with described user's question sentence
Asking, the most described interpolation module creates the knowledge point corresponding with described user's question sentence in knowledge base, and described knowledge point includes: knowledge
Library standard is asked, knowledge base extension is asked and answer.
In one embodiment, the knowledge point created in knowledge base is also added simultaneously to described model by described interpolation module
Sample storehouse.
In one embodiment, described Semantic Similarity Measurement module includes: word and vector calculation module, for expanding sample
Exhibition is asked and is carried out participle, and calculates word and sentence vector value, and described user's question sentence carries out participle, and calculate word and sentence to
Value；And relatedness computation module, it is used for calculating sample and extends the word of the word asked and sentence vector value and described user's question sentence
With the degree of association of sentence vector value, the semantic similarity asked to draw described user's question sentence and sample to extend.
Whether in one embodiment, described device also includes: pretreatment module, be used in determining described model sample storehouse
Before the sample extension that existence and described user's question sentence match is asked, user's question sentences all in described man-machine interaction daily record are entered
Row pretreatment, to filter the invalid data in man-machine interaction daily record user's question sentence.
The present invention also provides for the system of a kind of information processing, the device processed including described any information, also includes: model
Sample storehouse, described model sample storehouse includes that sample standard asks and ask with each sample standard that corresponding sample extension is asked；Know
Knowing storehouse, described knowledge base includes that knowledge base standard asks and ask with each knowledge base standard that corresponding knowledge base extension is asked and answers
Case, described knowledge base is for furnishing an answer for user's question sentence.
The present invention choose need to optimize man-machine interaction daily record time, first pass through the model sample storehouse set up and carry out Automatic sieve
Choosing, has filtered out substantial amounts of existing knowledge content, has decreased the input amount of hand labor.Simultaneity factor can need to optimize people from trend
Machine interactive log proposed standard is asked, artificial need to carry out selecting, further reduce hand labor, improve knowledge base
Optimize efficiency.
More preferably understand in order to the above-mentioned and other aspect of the present invention is had, preferred embodiment cited below particularly, and coordinate attached
Figure, is described in detail below:
Accompanying drawing explanation
Fig. 1 is knowledge base schematic diagram of the present invention；
Fig. 2 is model sample storehouse of the present invention schematic diagram；
Fig. 3 be the information processing of one embodiment of the invention method flow in optimize knowledge base flow process schematic diagram；
Fig. 4 is the schematic diagram of the method flow of the information processing of one embodiment of the invention；
Fig. 5 is the schematic diagram of the device of the information processing of one embodiment of the invention.
Detailed description of the invention
User can produce interactive log in intelligent robot interaction, and every interactive log is by user's question sentence, right
The knowledge base standard answered is asked and answer three part composition.Wherein user's question sentence is to be directly inputted acquisition by user, passes through question and answer
After engine is to the parsing identification of user's question sentence, calls corresponding knowledge base standard and ask about the answer of correspondence.At these interactive logs
The middle answer giving corresponding knowledge point according to user's question sentence with robot is replied accuracy and is distinguished, and can be divided into user's question sentence content machine
Device people replies, user question sentence content robot gives correctly to reply, user question sentence content robot gives mistake answer.
Cause robot not reply or give the reason of mistake answer mainly due to robot knowledge base has lacked corresponding knowledge
The way to put questions of point or existing knowledge point is the abundantest.Therefore the analysis of the interactive log by producing every day, extracts because knowledge point lacks
Lose or the daily record of the not abundant incorrect answer of robot caused of way to put questions is a main path to knowledge base Continuous optimization.
The method and apparatus that the present invention provides can greatly reduce the artificial input amount when extracting the man-machine interaction daily record needing to optimize.
Present invention is primarily concerned with the user's question sentence in interactive log and standard is asked.
Refer to Fig. 1 and Fig. 2, figures 1 and 2 show that the partial objects of information processing of the present invention, knowledge base and model sample
Example storehouse.
As it is shown in figure 1, knowledge base 10 includes that at least one knowledge base standard is asked 101 and asks phase with each knowledge base standard
Corresponding knowledge base extension asks 1011 and answer, and the most each knowledge base standard is asked a corresponding answer, can be had multiple knowledge
Storehouse extension asks that the extension of 1011-knowledge base asks 101n that a corresponding knowledge base standard asks 101.Owing to knowledge base standard asks 101 with answering
There is relation one to one in case, present invention is primarily concerned with knowledge base standard and ask and ask corresponding with each knowledge base standard
Knowledge base extends the processing procedure asked.Generally, all can there is multiple knowledge base standard and ask in knowledge base, knowledge base standard is asked
101-knowledge base standard asks 10n.Knowledge base includes that multiple knowledge point, each knowledge point include: knowledge base standard asks,
The extension of multiple knowledge bases is asked and an answer, and the most different knowledge base extensions asks it is all corresponding same answer, a knowledge base
Standard asks also this answer corresponding.It is usually the multiple knowledge bases extension corresponding from each knowledge point and asks middle selection one expression
The knowledge base extension being clearly easily maintained asks that the knowledge base standard as this knowledge point is asked, therefore knowledge base standard is asked and known with one
Know storehouse extension and ask identical.It should be noted that each knowledge base standard asks that corresponding knowledge base extension asks that number can be identical,
Can also be different.
In interactive process, after receiving user's question sentence, can be obtained from knowledge base by Semantic Similarity Measurement
The highest with the semantic similarity of user's question sentence and higher than threshold value knowledge base extension is asked, and asks answering of correspondence by the extension of this knowledge base
Case is sent to user, simultaneously using this user's question sentence and with this knowledge base extension ask corresponding knowledge base standard with asking relatedness as
Article one, interactive log.
As in figure 2 it is shown, model sample storehouse 20 includes that at least one sample standard asks 201 and corresponding one or many
The extension of individual sample asks 2011, is similar to knowledge base data structure, and a sample standard is asked and can be asked correspondence with the extension of multiple samples.
Be usually from the extension of multiple samples ask middle selection one express the extension that is clearly easily maintained ask as with the plurality of sample pair
The sample standard answered is asked, therefore sample standard is asked and asked identical with the extension of one of them sample.Each sample standard asks the sample of correspondence
Example extension asks that number can be identical, it is also possible to different.
Refer to Fig. 3, it is shown that the knowledge base Optimizing Flow 30 of one embodiment of the invention, comprise the steps of
Step 301: start.
Step 302: determine and whether there is the sample matched with the user's question sentence in man-machine interaction daily record in model sample storehouse
Example extension is asked.
Step 303: if exist, it is determined that the corresponding standard of user's question sentence described in described man-machine interaction daily record ask with
The corresponding sample standard that the sample extension joined is asked is asked the most identical.
Step 304: if differing, then optimize described knowledge base.
In step 302, first looked for whether in model sample storehouse with man-machine interaction daily record user's question sentence is semantic near
As sample extension ask, if there being approximation, then referred to as coupling.If there being coupling, now think that this user's question sentence can quilt
Model sample storehouse judges.The most in step 303, if can be determined, it is determined that the standard that this user's question sentence is corresponding is asked and this sample
Example extension asks that the standard of correspondence is asked the most identical, and herein identical refers to that word is completely the same, if identical, shows in knowledge base
Include the knowledge point corresponding with this user's question sentence, it is not necessary to utilize this user journal to optimize knowledge base.If differing, then show
Model sample storehouse with in knowledge base all less than the question sentence corresponding with this interactive log content, now show that this interactive log is new
Content, need to utilize this interactive log Advance data quality knowledge base, namely enter in step 304.Now, due to alternately
User's question sentence in daily record can be determined, can directly will approximated with interactive log user's question sentence semanteme in model sample storehouse
One or more sample standards that individual or multiple sample question sentence is corresponding are asked and are recommended knowledge maintenance personnel, when for one, by knowing
Know attendant to judge whether properly；When for time multiple, by knowledge maintenance personnel the most directly select one most suitable,
Finally will determine that most suitable sample standard that is suitable or that select is asked and user's question sentence is stored in knowledge base explicitly, thus people
The input of work has only to carry out simple supervision and management, and the knowledge maintenance personnel of the management that exercises supervision only need to recognize Chinese, has
Normal logical judgment ability, so manually needs coming of certain knowledge edition experience for needing before to put into
Say, reduce further the requirement to personnel's threshold, and improve optimization efficiency.
Being also advantageous in that of the method, it may be judged whether need optimization knowledge base to be entirely and complete in local model sample storehouse
, and without using the knowledge base in high in the clouds.The most both improve arithmetic speed, save again the spending of high in the clouds knowledge base.
In one embodiment, sample extension is asked and is included that knowledge base extension is asked, sample standard is asked and included that knowledge base standard is asked.More
Further, sample extension asks that all knowledge bases including in knowledge base extension is asked, sample standard asks the institute including in knowledge base
Knowledge base standard is had to ask.In this embodiment, all knowledge base standards that model sample storehouse includes in knowledge base are asked and knowledge
Storehouse extension is asked.The now model sample storehouse judgement to whether optimizing is the most accurate, reduce even further what subsequent artefacts selected
Workload.
In one embodiment, in step 302, if judged result is, model sample storehouse does not exist and asks with described user
The sample extension that sentence matches is asked, then create the knowledge point corresponding with described user's question sentence in knowledge base, and described knowledge point is wrapped
Include: knowledge base standard is asked, knowledge base extension is asked and answer.In this embodiment, it is believed that this interactive log cannot be by model sample storehouse
Judged, i.e. the most relevant to this interactive log in knowledge base information, need to utilize this interactive log to optimize knowledge base.Now
Due to this interactive log undecidable, only actively add a knowledge relevant to this user's question sentence by knowledge maintenance personnel
Point, i.e. needs to add that a knowledge base standard is asked, the extension of multiple knowledge base is asked and an answer, completes the optimization of knowledge base.
In a preferred embodiment, whether step 302 mates and is weighed by semantic similarity, can be set
One threshold value, when semantic similarity is more than first threshold, it is believed that interactive log user's question sentence asks coupling with sample extension.When manually
When input amount can ensure that, described first threshold can be set higher..Otherwise, then first threshold can be set ground
Lower, such that it is able to save human cost.
In one embodiment, described the user's question sentence phase whether existed in model sample storehouse with man-machine interaction daily record is determined
The sample extension of coupling is asked, is completed by semantic matching degree computing, including step: asks sample extension and carries out participle, and
Calculate word and sentence vector value；Described user's question sentence is carried out participle, and calculates word and sentence vector value；Calculate each sample to expand
Open up the word and sentence vector value and the word of described user's question sentence and the degree of association of sentence vector value asked, to draw described user's question sentence
The semantic similarity asked is extended with sample.The operation method of semantic matching degree is a lot, and method of the prior art can also be transported
Use in the present invention.
Owing to the quality in model sample storehouse is most important for the present invention, more preferably, in another embodiment, to model sample
Example storehouse is optimized, including two ways: one, while being optimized knowledge base, and identical content is added into model
Sample storehouse；Two, when existing with described user's question sentence semantic similarity more than described first threshold and the extension of the sample less than 100%
Ask, and the corresponding standard of described user's question sentence is asked with semantic similarity more than described first threshold and the sample less than 100%
The corresponding sample standard question sentence that extension is asked is identical, then ask phase by the corresponding standard of described user's question sentence and described user's question sentence
Associatedly it is added into model sample storehouse.The first optimizes primarily to make model sample storehouse content keep with knowledge base content
Unanimously, and up-to-date question sentence and standard being asked updates into model sample storehouse, to encounter the friendship of the content approximation with updating in next time
Mutually during daily record, directly can be filtered, without artificial judgment optimization by the present invention.Under the second optimal way, due to
From knowledge base, correct answer can be provided for current user's question sentence, i.e. find correct standard to ask, such that it is able to need not be by
Interactive log optimizes into knowledge base, but optimizes and be conducive in model sample storehouse including follow-up more interactive log in can determine that
In the range of, such that it is able to directly process relevant interactive log by the present invention.
Refer to Fig. 4, for the schematic diagram of the information processing method flow process of one embodiment of the invention, compare Fig. 2, shown in Fig. 3
Method flow includes the optimization to model sample storehouse.Specifically include:
Step 401: start.
Step 402: determine and whether there is the sample matched with the user's question sentence in man-machine interaction daily record in model sample storehouse
Example extension is asked, enters step 403 if existing, and otherwise enters step 405.
Step 403: determine that described in described man-machine interaction daily record, the corresponding standard of user's question sentence is asked and the sample mated
The corresponding sample standard that extension is asked is asked the most identical.If identical entrance step 404, otherwise enter step 406.
Step 404: judge whether the semantic similarity that user's question sentence and sample extension are asked is more than first threshold and is less than
100%, if then entering 407, otherwise enter 408.
Step 405: re-create knowledge point, and use knowledge point to optimize knowledge base and model sample storehouse.
Step 406: select creation of knowledge point, and use knowledge point to optimize knowledge base and model sample storehouse.
Step 407: use interactive log content, Optimized model sample storehouse.
Step 408: terminate.
Wherein step 405 content includes: by knowledge maintenance personnel actively add one to this relevant knowing of user's question sentence
Know point, i.e. need to add that a knowledge base standard is asked, the extension of multiple knowledge base is asked and an answer, completes the excellent of knowledge base
Changing, utilize identical Optimized model sample storehouse, knowledge point simultaneously, simply the optimization in model sample storehouse has only been used in knowledge point
Question sentence and standard ask content.Step 406 includes: the one or more standards in recommended models sample storehouse are asked to knowledge maintenance people
Member, the pairing that knowledge maintenance personnel directly carry out selecting being formed user's question sentence and standard is asked, then this pairing is added
Enter knowledge base, this pairing is added into model sample storehouse simultaneously.In step 407, by right with institute for the user's question sentence in interactive log
The standard answered is asked and is added in model sample storehouse, thus forms a pair new sample extension and ask the correspondence asked with sample standard.This
Invention also provides for the device 51 of a kind of information processing, refer to Fig. 5.In one embodiment, described device includes the first analysis mould
Block 501, second is analyzed module 502 and optimizes module 503.Interactive log initially enters the first analysis module 501, and first analyzes mould
Block 501 determines that whether there is the sample extension matched with the user's question sentence in man-machine interaction daily record in model sample storehouse asks, if
Exist, then enter second analysis module 502, determine the corresponding standard of user's question sentence described in described man-machine interaction daily record ask with
The corresponding sample standard that the sample extension of coupling is asked is asked the most identical, if differing, entering optimization module 503 and knowing described
Know storehouse to be optimized.
In another embodiment, refer to Fig. 5, first analyzes module 501 also includes Semantic Similarity Measurement module 5011,
The user's question sentence being used for calculating in man-machine interaction daily record extends, with sample, the semantic similarity asked, and then draws matching degree.Second
Analyzing module 502 and include comparison module 5021, the corresponding standard being used for relatively described user's question sentence is asked and the sample expansion mated
The corresponding sample standard that exhibition is asked asks that word is the most completely the same.Optimize module 503 also include recommending module 5031, for based on
The result of Semantic Similarity Measurement module 5011, it is recommended that with the sample that the semantic matching degree of described user's question sentence is more than Second Threshold
The corresponding sample standard that extension is asked is asked.Optimize module 503 also to include adding module 5032, for by from the sample mark recommended
Standard is asked the standard that middle artificial selection goes out to ask with described user's question sentence and is added into described knowledge base explicitly, simultaneously by foregoing
Optimize and be added into model sample storehouse.
More preferably, while knowledge base is optimized, model sample storehouse 504 is optimized.Second analyzes module 502 also
Including adding module 5022, whether the semantic similarity asked when user's question sentence and sample extension is more than first threshold and is less than
100%, and when the standard of correspondence is asked identical, interactive log content optimization is entered model sample storehouse.Add module 5032 to be additionally operable to
The sample standard recommended from recommending module 5031 is asked the standard that middle artificial selection goes out ask with described user's question sentence add explicitly
Add model sample storehouse.
In another embodiment, first filter the invalid data in interactive log, can pick according to default filtering rule
Except the junk data in daily record data, such as: single English alphabet is repeated 5 times above data.Naive Bayesian can be used afterwards
Algorithm is analyzed, and calculates log content whether in the range of analysis model can determine that.
The present invention also provides for the system 52 of a kind of information processing, refer to Fig. 5.Including described any information processing means,
Include knowledge base 504 and model sample storehouse 505 simultaneously.
The present invention choose need to optimize man-machine interaction daily record time, first pass through the model sample storehouse set up and carry out Automatic sieve
Choosing, has filtered out substantial amounts of existing knowledge content, has decreased the input amount of hand labor.Simultaneity factor can need to optimize people from trend
Machine interactive log proposed standard is asked, artificial need to carry out selecting, further reduce hand labor, improve knowledge base
Optimize efficiency.
Thering is provided of this disclosure being previously described is for making any person skilled in the art all can make or use these public affairs
Open.Various amendment of this disclosure the most all will be apparent from, and as defined herein general
Suitable principle can be applied to other variants spirit or scope without departing from the disclosure.Thus, the disclosure is not intended to be limited
Due to example described herein and design, but should be awarded and principle disclosed herein and novel features phase one
The widest scope caused.
Claims (24)
1. a method for information processing, including:
Thering is provided model sample storehouse, described model sample storehouse includes that sample standard is asked and asks corresponding sample with each sample standard
Example extension is asked；
Thering is provided knowledge base, described knowledge base includes that knowledge base standard is asked and asks corresponding knowledge base with each knowledge base standard
Extension is asked and answer, and described knowledge base is for furnishing an answer for user's question sentence；
Determine that whether there is the sample extension matched with the user's question sentence in man-machine interaction daily record in described model sample storehouse asks；
If existing, it is determined that the corresponding standard of user's question sentence described in described man-machine interaction daily record is asked and the sample extension mated
The corresponding sample standard asked is asked the most identical；
If differing, then optimize described knowledge base.
2. the method for information processing as claimed in claim 1, it is characterised in that the extension of described sample is asked and included that knowledge base extends
Asking, described sample standard is asked and is included that knowledge base standard is asked.
3. the method for information processing as claimed in claim 1, it is characterised in that determine in described model sample storehouse and whether exist
The sample extension matched with described user's question sentence is asked and is included:
Ask whether execution Semantic Similarity Measurement is deposited to determine in described model sample storehouse by described user's question sentence and sample extension
Ask more than the sample extension of first threshold with the semantic similarity of described user's question sentence at least one.
4. the method for information processing as claimed in claim 1, it is characterised in that determine the corresponding standard of described user's question sentence
The corresponding sample standard that the sample extension asked and mate is asked is asked the most identical including:
The relatively corresponding standard of described user's question sentence asks that the corresponding sample standard asked with the sample extension mated asks that word is
No completely the same.
5. the method for information processing as claimed in claim 4, it is characterised in that if existing and described user's question sentence semantic similitude
Degree is asked more than described first threshold and the extension of the sample less than 100%, and the corresponding standard of described user's question sentence is asked with semantic
Similarity is identical, then by institute more than the corresponding sample standard question sentence that described first threshold and the extension of the sample less than 100% are asked
The corresponding standard of user's question sentence and described user's question sentence stated is asked and is added into described model sample storehouse explicitly.
6. the method for information processing as claimed in claim 4, it is characterised in that if the sample extension that there is multiple coupling is asked,
Then determine that the corresponding standard of described user's question sentence asks that the corresponding sample standard asked with the sample extension mated is asked the most identical
Including:
Determine whether that the corresponding sample standard that a sample extension mated is asked is asked and the corresponding mark of described user's question sentence
Standard is asked identical.
7. the method for information processing as claimed in claim 4, it is characterised in that the optimization to described knowledge base includes:
Result based on described Semantic Similarity Measurement, it is recommended that be more than Second Threshold with the semantic similarity of described user's question sentence
The corresponding sample standard that sample extension is asked is asked；
Add asking the sample standard that middle artificial selection goes out to ask with described user's question sentence from the sample standard recommended explicitly
Enter described knowledge base.
8. the method for information processing as claimed in claim 7, it is characterised in that described method also includes:
Ask that the sample standard that middle artificial selection goes out is asked with described user's question sentence explicitly by described from the sample standard recommended
It is added into described model sample storehouse.
9. the method for information processing as claimed in claim 1, it is characterised in that if not existing and institute in described model sample storehouse
State the sample extension that user's question sentence matches to ask, then in knowledge base, create the knowledge point corresponding with described user's question sentence, described
Knowledge point includes: knowledge base standard is asked, knowledge base extension is asked and answer.
10. the method for information processing as claimed in claim 9, it is characterised in that described method also includes: will be in knowledge base
The knowledge point created is added simultaneously to described model sample storehouse.
The method of 11. information processings as claimed in claim 3, it is characterised in that described user's question sentence is asked with sample extension
Execution Semantic Similarity Measurement includes:
Sample extension is asked and carries out participle, and calculate word and sentence vector value；
Described user's question sentence is carried out participle, and calculates word and sentence vector value；
Calculate sample and extend the word and sentence vector value and the word of described user's question sentence and the degree of association of sentence vector value asked, with
Go out described user's question sentence and extend, with sample, the semantic similarity asked.
The method of 12. information processings as claimed in claim 1, it is characterised in that in determining described model sample storehouse whether
Before the sample extension that existence and described user's question sentence match is asked, described method also includes:
User's question sentences all in described man-machine interaction daily record are carried out pretreatment, to filter in man-machine interaction daily record user's question sentence
Invalid data.
The device of 13. 1 kinds of information processings, including:
Whether first analyzes module, match with the user's question sentence in man-machine interaction daily record for determining to exist in model sample storehouse
Sample extension ask；
Second analyzes module, asks for the sample extension matched in response to existence and described user's question sentence, it is determined that described people
The corresponding standard of user's question sentence described in machine interactive log asks that the corresponding sample standard asked with the sample extension mated is asked
No identical；And
Optimize module, extend, with the sample mated, the corresponding sample asked for asking in response to the corresponding standard of described user's question sentence
Example standard is asked and is differed, then optimize knowledge base.
The device of 14. information processings as claimed in claim 13, it is characterised in that described first analyzes module includes:
Semantic Similarity Measurement module, for asking execution Semantic Similarity Measurement, with really by described user's question sentence with sample extension
Whether fixed described model sample storehouse exists at least one semantic similarity with described user's question sentence sample more than first threshold
Example extension is asked.
The device of 15. information processings as claimed in claim 14, it is characterised in that described second analyzes module includes:
Comparison module, the corresponding standard for relatively described user's question sentence is asked and is extended, with the sample mated, the corresponding sample asked
Standard asks that word is the most completely the same.
The device of 16. information processings as claimed in claim 15, it is characterised in that described second analyzes module also includes:
Add module, for more than described first threshold and being less than with described user's question sentence semantic similarity in response to existing
The sample extension of 100% is asked, and the corresponding standard of described user's question sentence ask with semantic similarity more than described first threshold and
The corresponding sample standard question sentence that sample extension less than 100% is asked is identical, then by described user's question sentence and described user's question sentence
Corresponding standard ask and be added into described model sample storehouse explicitly.
The device of 17. information processings as claimed in claim 14, it is characterised in that if there is the sample extension of multiple coupling
Asking, the most described second analyzes module determines whether that a sample mated extends the corresponding sample standard asked and asks and described use
The corresponding standard of family question sentence is asked identical.
The device of 18. information processings as claimed in claim 14, it is characterised in that described optimization module includes:
Recommending module, for result based on described Semantic Similarity Measurement, it is recommended that with the semantic matching degree of described user's question sentence
The corresponding sample standard asked more than the extension of the sample of Second Threshold is asked；And
Add module, for relevant to described user's question sentence by asking that the standard that middle artificial selection goes out is asked from the sample standard recommended
It is added into described knowledge base connection.
The device of 19. information processings as claimed in claim 18, it is characterised in that described interpolation module is further used for institute
State and ask the standard that middle artificial selection goes out to ask with described user's question sentence from the sample standard recommended to be added into described mould explicitly
Pattern example storehouse.
The device of 20. information processings as claimed in claim 18, it is characterised in that if described model sample storehouse does not exist with
The sample extension that described user's question sentence matches is asked, the most described interpolation module creates corresponding with described user's question sentence in knowledge base
Knowledge point, described knowledge point includes: knowledge base standard is asked, knowledge base extension ask and answer.
The device of 21. information processings as claimed in claim 20, it is characterised in that described interpolation module also will be in knowledge base
The knowledge point created is added simultaneously to described model sample storehouse.
The device of 22. information processings as claimed in claim 14, it is characterised in that described Semantic Similarity Measurement module bag
Include:
Participle and vector calculation module, carry out participle for asking sample extension, and calculate word and sentence vector value, and to institute
State user's question sentence and carry out participle, and calculate word and sentence vector value；And
Relatedness computation module, is used for calculating sample and extends the word and sentence vector value and the word of described user's question sentence and sentence asked
The degree of association of vector value, to show that described user's question sentence extends, with sample, the semantic similarity asked.
The device of 23. information processings as claimed in claim 13, it is characterised in that described device also includes:
Pretreatment module, for whether there is the sample matched with described user's question sentence in determining described model sample storehouse expands
Before exhibition is asked, user's question sentences all in described man-machine interaction daily record are carried out pretreatment, ask filtering man-machine interaction daily record user
Invalid data in Ju.
The system of 24. 1 kinds of information processings, it is characterised in that described system includes:
The device of the information processing any one of claim 13-23；
Model sample storehouse, described model sample storehouse includes that sample standard asks and ask with each sample standard that corresponding sample expands
Zhan Wen；
Knowledge base, described knowledge base includes that corresponding knowledge base extension is asked and asked with each knowledge base standard to knowledge base standard
Asking and answer, described knowledge base is for furnishing an answer for user's question sentence.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810620088.5A CN108764480B (en) | 2016-08-23 | 2016-08-23 | Information processing system |
| CN201610710565.8A CN106295807B (en) | 2016-08-23 | 2016-08-23 | A kind of method and device of information processing |
| CN201811074893.9A CN109344237B (en) | 2016-08-23 | 2016-08-23 | Information processing method and device for man-machine interaction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610710565.8A CN106295807B (en) | 2016-08-23 | 2016-08-23 | A kind of method and device of information processing |
Related Child Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811074893.9A Division CN109344237B (en) | 2016-08-23 | 2016-08-23 | Information processing method and device for man-machine interaction |
| CN201810620088.5A Division CN108764480B (en) | 2016-08-23 | 2016-08-23 | Information processing system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106295807A true CN106295807A (en) | 2017-01-04 |
| CN106295807B CN106295807B (en) | 2018-12-21 |
Family
ID=57615826
Family Applications (3)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810620088.5A Active CN108764480B (en) | 2016-08-23 | 2016-08-23 | Information processing system |
| CN201610710565.8A Active CN106295807B (en) | 2016-08-23 | 2016-08-23 | A kind of method and device of information processing |
| CN201811074893.9A Active CN109344237B (en) | 2016-08-23 | 2016-08-23 | Information processing method and device for man-machine interaction |
Family Applications Before (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810620088.5A Active CN108764480B (en) | 2016-08-23 | 2016-08-23 | Information processing system |
Family Applications After (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811074893.9A Active CN109344237B (en) | 2016-08-23 | 2016-08-23 | Information processing method and device for man-machine interaction |
Country Status (1)
| Country | Link |
|---|---|
| CN (3) | CN108764480B (en) |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106886820A (en) * | 2017-02-08 | 2017-06-23 | 深圳市科迈爱康科技有限公司 | Intelligent information processing method and system |
| CN108345644A (en) * | 2018-01-15 | 2018-07-31 | 阿里巴巴集团控股有限公司 | A kind of method and device of data processing |
| CN109189912A (en) * | 2018-10-09 | 2019-01-11 | 阿里巴巴集团控股有限公司 | The update method and device of user's consulting statement library |
| CN109213847A (en) * | 2018-09-14 | 2019-01-15 | 广州神马移动信息科技有限公司 | Layered approach and its device, electronic equipment, the computer-readable medium of answer |
| CN109325040A (en) * | 2018-07-13 | 2019-02-12 | 众安信息技术服务有限公司 | A kind of extensive method, device and equipment in FAQ question and answer library |
| CN109934347A (en) * | 2017-12-18 | 2019-06-25 | 上海智臻智能网络科技股份有限公司 | Extend the device of question and answer knowledge base |
| CN109933777A (en) * | 2017-12-18 | 2019-06-25 | 上海智臻智能网络科技股份有限公司 | Knowledge base expanding unit |
| CN109947651A (en) * | 2019-03-21 | 2019-06-28 | 上海智臻智能网络科技股份有限公司 | Artificial intelligence engine optimization method and device |
| CN109992675A (en) * | 2019-01-30 | 2019-07-09 | 阿里巴巴集团控股有限公司 | Information processing method and device |
| CN110019304A (en) * | 2017-12-18 | 2019-07-16 | 上海智臻智能网络科技股份有限公司 | Extend the method and storage medium, terminal of question and answer knowledge base |
| CN110019305A (en) * | 2017-12-18 | 2019-07-16 | 上海智臻智能网络科技股份有限公司 | Knowledge base extended method and storage medium, terminal |
| CN110347807A (en) * | 2019-05-20 | 2019-10-18 | 平安科技（深圳）有限公司 | Problem information processing method and processing device |
| WO2020047779A1 (en) * | 2018-09-05 | 2020-03-12 | 西门子（中国）有限公司 | Fault analysis method and device and computer readable medium |
| CN110928991A (en) * | 2019-11-20 | 2020-03-27 | 上海智臻智能网络科技股份有限公司 | Method and device for updating question-answer knowledge base |
| CN111125379A (en) * | 2019-12-26 | 2020-05-08 | 科大讯飞股份有限公司 | Knowledge base expansion method and device, electronic equipment and storage medium |
| CN111144098A (en) * | 2019-12-26 | 2020-05-12 | 支付宝(杭州)信息技术有限公司 | Recall method and device for expanded question sentence |
| CN111382239A (en) * | 2018-12-27 | 2020-07-07 | 上海智臻智能网络科技股份有限公司 | Method and device for optimizing interaction flow |
| CN111382235A (en) * | 2018-12-27 | 2020-07-07 | 上海智臻智能网络科技股份有限公司 | Question-answer knowledge base optimization method and device |
| CN111400458A (en) * | 2018-12-27 | 2020-07-10 | 上海智臻智能网络科技股份有限公司 | Automatic generalization method and device |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109829051B (en) * | 2019-01-30 | 2023-01-17 | 科大讯飞股份有限公司 | Method and device for screening similar sentences of database |
| CN110362665B (en) * | 2019-06-12 | 2021-04-30 | 深圳追一科技有限公司 | Question-answering system and method based on semantic similarity |
| CN112936304B (en) * | 2021-02-02 | 2022-09-16 | 浙江大学 | Self-evolution type service robot system and learning method thereof |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090210218A1 (en) * | 2008-02-07 | 2009-08-20 | Nec Laboratories America, Inc. | Deep Neural Networks and Methods for Using Same |
| CN104679815A (en) * | 2014-12-08 | 2015-06-03 | 北京云知声信息技术有限公司 | Method and system for screening question and answer pairs and updating question and answer database in real time |
| CN105488185A (en) * | 2015-12-01 | 2016-04-13 | 上海智臻智能网络科技股份有限公司 | Optimization method and apparatus for knowledge base |
| CN105550361A (en) * | 2015-12-31 | 2016-05-04 | 上海智臻智能网络科技股份有限公司 | Log processing method and apparatus, and ask-answer information processing method and apparatus |
| CN105678324A (en) * | 2015-12-31 | 2016-06-15 | 上海智臻智能网络科技股份有限公司 | Similarity calculation-based questions and answers knowledge base establishing method, device and system |
| CN105677783A (en) * | 2015-12-31 | 2016-06-15 | 上海智臻智能网络科技股份有限公司 | Information processing method and device for intelligent question-answering system |
| CN105824797A (en) * | 2015-01-04 | 2016-08-03 | 华为技术有限公司 | Method, device and system evaluating semantic similarity |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101178705A (en) * | 2007-12-13 | 2008-05-14 | 中国电信股份有限公司 | Free-running speech comprehend method and man-machine interactive intelligent system |
| CN104199825A (en) * | 2014-07-23 | 2014-12-10 | 清华大学 | Information inquiry method and system |
| CN104360994A (en) * | 2014-12-04 | 2015-02-18 | 科大讯飞股份有限公司 | Natural language understanding method and natural language understanding system |
| US20160196490A1 (en) * | 2015-01-02 | 2016-07-07 | International Business Machines Corporation | Method for Recommending Content to Ingest as Corpora Based on Interaction History in Natural Language Question and Answering Systems |
| CN105591882B (en) * | 2015-12-10 | 2018-03-06 | 北京中科汇联科技股份有限公司 | A kind of intelligence machine person to person mixes the method and system of customer service |
| CN105631022B (en) * | 2015-12-29 | 2019-03-05 | 上海智臻智能网络科技股份有限公司 | Information processing method and device |
-
2016
- 2016-08-23 CN CN201810620088.5A patent/CN108764480B/en active Active
- 2016-08-23 CN CN201610710565.8A patent/CN106295807B/en active Active
- 2016-08-23 CN CN201811074893.9A patent/CN109344237B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090210218A1 (en) * | 2008-02-07 | 2009-08-20 | Nec Laboratories America, Inc. | Deep Neural Networks and Methods for Using Same |
| CN104679815A (en) * | 2014-12-08 | 2015-06-03 | 北京云知声信息技术有限公司 | Method and system for screening question and answer pairs and updating question and answer database in real time |
| CN105824797A (en) * | 2015-01-04 | 2016-08-03 | 华为技术有限公司 | Method, device and system evaluating semantic similarity |
| CN105488185A (en) * | 2015-12-01 | 2016-04-13 | 上海智臻智能网络科技股份有限公司 | Optimization method and apparatus for knowledge base |
| CN105550361A (en) * | 2015-12-31 | 2016-05-04 | 上海智臻智能网络科技股份有限公司 | Log processing method and apparatus, and ask-answer information processing method and apparatus |
| CN105678324A (en) * | 2015-12-31 | 2016-06-15 | 上海智臻智能网络科技股份有限公司 | Similarity calculation-based questions and answers knowledge base establishing method, device and system |
| CN105677783A (en) * | 2015-12-31 | 2016-06-15 | 上海智臻智能网络科技股份有限公司 | Information processing method and device for intelligent question-answering system |
Non-Patent Citations (1)
| Title |
|---|
| 葛丽萍: "交互式问答系统中的待改进问题自动识别方法", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106886820A (en) * | 2017-02-08 | 2017-06-23 | 深圳市科迈爱康科技有限公司 | Intelligent information processing method and system |
| CN110019304B (en) * | 2017-12-18 | 2024-01-05 | 上海智臻智能网络科技股份有限公司 | Method for expanding question-answering knowledge base, storage medium and terminal |
| CN109934347B (en) * | 2017-12-18 | 2024-02-02 | 上海智臻智能网络科技股份有限公司 | Device for expanding question-answer knowledge base |
| CN109933777B (en) * | 2017-12-18 | 2024-02-06 | 上海智臻智能网络科技股份有限公司 | Knowledge base expanding device |
| CN109934347A (en) * | 2017-12-18 | 2019-06-25 | 上海智臻智能网络科技股份有限公司 | Extend the device of question and answer knowledge base |
| CN109933777A (en) * | 2017-12-18 | 2019-06-25 | 上海智臻智能网络科技股份有限公司 | Knowledge base expanding unit |
| CN110019305B (en) * | 2017-12-18 | 2024-03-15 | 上海智臻智能网络科技股份有限公司 | Knowledge base expansion method, storage medium and terminal |
| CN110019304A (en) * | 2017-12-18 | 2019-07-16 | 上海智臻智能网络科技股份有限公司 | Extend the method and storage medium, terminal of question and answer knowledge base |
| CN110019305A (en) * | 2017-12-18 | 2019-07-16 | 上海智臻智能网络科技股份有限公司 | Knowledge base extended method and storage medium, terminal |
| CN108345644A (en) * | 2018-01-15 | 2018-07-31 | 阿里巴巴集团控股有限公司 | A kind of method and device of data processing |
| CN109325040A (en) * | 2018-07-13 | 2019-02-12 | 众安信息技术服务有限公司 | A kind of extensive method, device and equipment in FAQ question and answer library |
| CN109325040B (en) * | 2018-07-13 | 2020-11-10 | 众安信息技术服务有限公司 | FAQ question-answer library generalization method, device and equipment |
| WO2020047779A1 (en) * | 2018-09-05 | 2020-03-12 | 西门子（中国）有限公司 | Fault analysis method and device and computer readable medium |
| CN109213847A (en) * | 2018-09-14 | 2019-01-15 | 广州神马移动信息科技有限公司 | Layered approach and its device, electronic equipment, the computer-readable medium of answer |
| CN109189912A (en) * | 2018-10-09 | 2019-01-11 | 阿里巴巴集团控股有限公司 | The update method and device of user's consulting statement library |
| CN111382239A (en) * | 2018-12-27 | 2020-07-07 | 上海智臻智能网络科技股份有限公司 | Method and device for optimizing interaction flow |
| CN111382235A (en) * | 2018-12-27 | 2020-07-07 | 上海智臻智能网络科技股份有限公司 | Question-answer knowledge base optimization method and device |
| CN111400458A (en) * | 2018-12-27 | 2020-07-10 | 上海智臻智能网络科技股份有限公司 | Automatic generalization method and device |
| CN111382239B (en) * | 2018-12-27 | 2023-06-23 | 上海智臻智能网络科技股份有限公司 | Interaction flow optimization method and device |
| CN109992675A (en) * | 2019-01-30 | 2019-07-09 | 阿里巴巴集团控股有限公司 | Information processing method and device |
| CN109947651A (en) * | 2019-03-21 | 2019-06-28 | 上海智臻智能网络科技股份有限公司 | Artificial intelligence engine optimization method and device |
| CN110347807A (en) * | 2019-05-20 | 2019-10-18 | 平安科技（深圳）有限公司 | Problem information processing method and processing device |
| CN110347807B (en) * | 2019-05-20 | 2023-08-08 | 平安科技（深圳）有限公司 | Problem information processing method and device |
| CN110928991A (en) * | 2019-11-20 | 2020-03-27 | 上海智臻智能网络科技股份有限公司 | Method and device for updating question-answer knowledge base |
| CN111125379A (en) * | 2019-12-26 | 2020-05-08 | 科大讯飞股份有限公司 | Knowledge base expansion method and device, electronic equipment and storage medium |
| CN111144098B (en) * | 2019-12-26 | 2023-05-30 | 支付宝(杭州)信息技术有限公司 | Recall method and device for extended question |
| CN111125379B (en) * | 2019-12-26 | 2022-12-06 | 科大讯飞股份有限公司 | Knowledge base expansion method and device, electronic equipment and storage medium |
| CN111144098A (en) * | 2019-12-26 | 2020-05-12 | 支付宝(杭州)信息技术有限公司 | Recall method and device for expanded question sentence |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106295807B (en) | 2018-12-21 |
| CN108764480A (en) | 2018-11-06 |
| CN109344237B (en) | 2020-11-17 |
| CN109344237A (en) | 2019-02-15 |
| CN108764480B (en) | 2020-07-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106295807A (en) | A kind of method and device of information processing | |
| CN110008288B (en) | Construction method and application of knowledge map library for network fault analysis | |
| CN107329967B (en) | Question answering system and method based on deep learning | |
| CN107766511A (en) | Intelligent answer method, terminal and storage medium | |
| CN107464568A (en) | Based on the unrelated method for distinguishing speek person of Three dimensional convolution neutral net text and system | |
| CN104537341B (en) | Face picture information getting method and device | |
| CN106294774A (en) | User individual data processing method based on dialogue service and device | |
| CN109408821B (en) | Corpus generation method and device, computing equipment and storage medium | |
| CN102509001B (en) | Method for automatically removing time sequence data outlier point | |
| CN110442700A (en) | Man-machine more wheel dialogue methods and system, smart machine for human-computer interaction | |
| CN111767385A (en) | Intelligent question and answer method and device | |
| CN108388553A (en) | Talk with method, electronic equipment and the conversational system towards kitchen of disambiguation | |
| US20210073216A1 (en) | Business intelligence system based on artificial intelligence and analysis method thereof | |
| CN110110095A (en) | A kind of power command text matching technique based on shot and long term memory Recognition with Recurrent Neural Network | |
| CN111460145A (en) | Learning resource recommendation method, device and storage medium | |
| TW201820172A (en) | System, method and non-transitory computer readable storage medium for conversation analysis | |
| CN116719917A (en) | Knowledge question-answering system, method and storage medium combining large model with external knowledge | |
| CN115114421A (en) | Question-answer model training method | |
| CN105373568B (en) | Problem answers Auto-learning Method and device | |
| CN113807462B (en) | Network equipment fault cause positioning method and system based on AI | |
| CN116467459A (en) | Internet of things equipment fault reporting method and device, computer equipment and storage medium | |
| CN110134775A (en) | Question and answer data creation method and device, storage medium | |
| CN112966084B (en) | Knowledge graph-based answer query method, device, equipment and storage medium | |
| CN114862233A (en) | Intelligent decision method and intelligent decision system | |
| CN113609268A (en) | Intelligent psychological knowledge question-answering method and device based on knowledge graph |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20181018 Address after: 201803 7, 398 Lane 1555, Jiangxi Road, Jinsha, Jiading District, Shanghai. Applicant after: SHANGHAI ZHIZHEN INTELLIGENT NETWORK SCIENCE & TECHNOLOGY CO., LTD. Applicant after: Guizhou little love robot technology Co., Ltd. Address before: 201803 7, 398 Lane 1555, Jiangxi Road, Jinsha, Jiading District, Shanghai. Applicant before: SHANGHAI ZHIZHEN INTELLIGENT NETWORK SCIENCE & TECHNOLOGY CO., LTD. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |