CN106528522A - Scenarized semantic comprehension and dialogue generation method and system - Google Patents
Scenarized semantic comprehension and dialogue generation method and system Download PDFInfo
- Publication number
- CN106528522A CN106528522A CN201610747188.5A CN201610747188A CN106528522A CN 106528522 A CN106528522 A CN 106528522A CN 201610747188 A CN201610747188 A CN 201610747188A CN 106528522 A CN106528522 A CN 106528522A
- Authority
- CN
- China
- Prior art keywords
- dialogue
- scene
- user
- model
- result
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/237—Lexical tools
- G06F40/247—Thesauruses; Synonyms
-
- 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
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
Abstract
The invention provides a scenarized semantic comprehension and dialogue generation method and a scenarized semantic comprehension and dialogue generation system. The method comprises the steps of: establishing a user scene model, selecting and determining the field of a current conversation according to a result of a word segmentation system; understanding an interactive content of the round based on a corresponding scenarized user model by use of a selected scene semantic parser; and calling a dialogue generator in a corresponding scene, carrying out a dialogue synthesis and generating a dialogue result after interaction of the round in combination with an intermediate state of dialogue management. According to the scenarized semantic comprehension and dialogue generation method and the scenarized semantic comprehension and dialogue generation system, natural language understanding for a sentence and a phrase is simply and efficiently realized, and the situation that a computer automatically and completely understands a short text is realized. Automatic deep understanding and dialogue interaction of the computer for the sentence or the phrase of a natural language are realized and the purpose that a user needs a machine to understand the interaction language use automatically and accurately is achieved.
Description
Technical field
The present invention relates to a kind of semantic understanding of displaying with dialogue generation method, belong to natural language analysis, process with
And semantic understanding field, especially for the semantics comprehension on natural language method under special scenes.
Background technology
Artificial intelligence technology development is long-standing, and as the natural language in one of most important direction in artificial intelligence field
The focus that understanding technology is also always studied, it may be said that allow computer understanding natural language to be mankind's dream all the time.Mesh
Before, natural language understanding technology is mainly adopted rule-based and is based on two big class theory and technologies of statistics.Rule-based natural language
Speech understands that system constructing is relative complex, the main understanding problem in the face of restricted domain.And Statistics-Based Method progressively becomes certainly
So the main stream approach in Language Processing field, is typically suitable for shallow semantic understanding is carried out on the basis of mass data, and for depth
Layer matter of semantics cannot often be processed.And specific application is directed to, the commonly used semantic reason for being also based on rule
Solution.
The text input given for one, traditional its handling process of rule-based natural language understanding are generally included
Three below step.
(1) meaning of a word analysis:Main purpose is the meaning of a word that vocabulary is obtained to the word retrieval linguistic information in sentence.Relate to
And participle, part-of-speech tagging, name body identification etc. technology, while propping up for the knowledge base of meaning of a word correlation is needed in meaning of a word process
Hold, great efforts are had been made to this language specialist, such as existing knowledge base " Hownet ",《Chinese thesaurus》Etc..
(2) syntactic analysis:Main purpose is that the structure to sentence or phrase is analyzed, and obtains vocabulary, phrase in sentence
In grammatical function and correlation, conventional syntactic analysis method include context-free grammar analysis and dependency grammar analyze
Etc..
(3) semantic analysis:Main purpose is that sentence practical significance to be expressed is obtained, and the part is natural language understanding
The core of system.Deep layer natural language understanding system generally needs to build the rule of related semantic understanding in the stage, these rule
The linguistry for then relying primarily on linguistic expertise is closely related with domain knowledge.On the one hand semantic item, the opposing party to be defined
Face will define the semantic conversion rule base from the information such as the meaning of a word, syntactic structure to semantic item.The text input given for one
After meaning of a word analysis, syntactic analysis, finally realize obtaining semantic purpose using semantic conversion rule base.
Traditional rule-based natural language understanding system relates generally to meaning of a word analysis, syntactic analysis and semantic analysis three
Part, each several part are directed to expertise rule, the utilization of different knowledge bases, although can realize that the deep layer to natural language is managed
Solution, but still haves the shortcomings that to overcome, mainly have it is following some:
(1) system complex, three part of the above be related to participle, part-of-speech tagging, name body identification, syntactic analysis, word sense disambiguation,
Semantic role analysis etc. numerous technology points, the performance deficiency of each technology be likely to cause the performance of whole system compared with
Big to affect, system is realized and maintenance difficulties are all larger.
(2) regular structure depends critically upon linguistic expertise knowledge, and the knowledge resource of this aspect is often limited
Even it is difficult to what is obtained.Different scenes or field, its semantic understanding rule are also different, and most semantic understanding draws
Hold up, such as:Baidu's voice, University of Science and Technology interrogate winged etc., are required to field was limited before understanding.
(3) in short text, the finiteness of voice data increased difficulty for semantic understanding.
The problems referred to above be natural language analysis, process and semantic understanding during should pay attention to and asking of solving
Topic.
The content of the invention
It is an object of the invention to provide a kind of semantic understanding of displaying and dialogue generation method, can be different with flexible adaptation
The short text semantic understanding of scene and dialogue, solve system complex present in prior art, depend critically upon linguistic expertise
In knowledge, and short text the problems such as the finiteness of voice data.
The present invention technical solution be:
A kind of semantic understanding of displaying and dialogue generation method, comprise the following steps,
Step 1, user's model of place is set up, including user's scene dictionary model, Scene Semantics analytic modell analytical model and scene pair
Words model;
Step 2, the scene that determination current sessions are selected according to the result of Words partition system;
The Scene Semantics resolver that step 3, employing are selected, understands to the interaction content of epicycle;
Step 4, call corresponding scene dialog generator, with reference to dialogue management intermediateness engage in the dialogue it is comprehensive and raw
Dialogue result into after epicycle interaction;
Step 5, receive again new text input after, repeat step 2-4 carries out semantic understanding, until conversational system tie
Beam.
Further, step 1 is specially:
Step 1-1, user's scene dictionary model is set up, i.e., domain features are described by composite key;Set up scene
Semantic understanding rule under semantic analytic modell analytical model, the i.e. special scenes;
User's guiding is used by step 1-2, the scene dialog model for setting up user, the conversation content being related under concrete scene
The dialogue interaction at family.
Further, step 2 is specially:
Step 2-1, call Words partition system participle to be carried out herein to current input, obtain the bag of words of text;
Step 2-2, word segmentation result and user-defined scene dictionary model are matched, should with determine current session
Use scene;
The state of step 2-3, the scene talked with according to this wheel and conversational system Scene Semantics final to determine dialogue
Resolver and scene dialog generator start new dialogue.
Further, in step 2-2, model of place dictionary includes field layer and domain features layer, and layer correspondence in field is different
Application scenarios, domain features layer to should be under application scenarios characteristic vector.
Further, the application scenarios of current session in step 2-2, are determined, specially:
Step 2-2-1, for the characteristic vector of certain sceneThe bag of words of input text, calculate the word pair in bag of words
Matching degree m of each component in characteristic vectori, wherein 1≤i≤n;
Step 2-2-2, the degree of correlation that the bag of words and the scene that are input into text are calculated by the method for weighting:
Wherein, WiFor certain vectorial weight;
Step 2-2-3, matching degree M for comparing each scenei, obtain maximum, be exactly determined by scene.
Further, in step 2-3, when the intention of user cannot be expressly understood that, user's request is obtained by talking with,
By introduce displaying model automatically realize content of text scene determine, and based on model of place come initiate dialogue.
During user session, user view is performed the step of merge, specially:
It is assumed that before epicycle is talked with, the result that user obtains isAnd in epicycle dialogue, by calculating
The result that method is obtained isNewest result output is obtained according to following process
In resulting user vector, all of vector is neither space-time, end-of-dialogue.
Further, step 3 is specially:
Step 3-1, third-party Chinese grammar parser is called to carry out syntax parsing to this conversation content;
Step 3-2, semantic solution is carried out to syntax parsing result with reference to user's scene dictionary model, Scene Semantics analytic modell analytical model
Analysis, obtains the semantic understanding result of epicycle.
Further, step 4 is specially:
Step 4-1, the result based on step 2 select the dialog generator under corresponding scene;
Step 4-2, the output result based on the current round obtained by step 3, and the dialogue shape in dialog manager
State engages in the dialogue synthesis, so as to obtain ending the semantic understanding result after epicycle interaction.
A kind of semantic understanding for realizing above-mentioned displaying and the system for talking with generation method, including:
User's model of place, there is provided user's scene dictionary model, semantic understanding model and talk with generation model accordingly;
Scene selector, it is for the natural language of receiving user's input and corresponding to select according to user's model of place
User's scene;Pre-build and store scene dictionary model defined in user, after the natural language of receiving user's input, will
The corresponding content of text of the natural language is matched with institute dictionary, initiates dialogue to have obtained when semantic input is imperfect
Whole semantic input;
Dialog manager:For the state of the intermediate result and correlation of management of dialogs system;
The semantic parser of displaying:Implement semantic understanding for the result after syntax parsing is combined user's model of place
Obtain the semantic understanding result of displaying;
Dialogue synthesizer:For the understanding output of the intermediate result and epicycle of comprehensive conversational system, finally obtain what is completed
Text output.
The invention has the beneficial effects as follows:The semantic understanding of this kind of displaying and dialogue generation method and system, it is simple efficient
The natural language understanding to sentence, phrase etc. is realized on ground, realizes fully automated understanding of the computer automatically to short text.By profit
The calculating become stronger day by day with current computer, storage capacity, pre-build different models of place, including scene dictionary, scene
Semantic analytic modell analytical model and the dialog model of displaying, after the natural language of receiving user's input, by natural language correspondence
Content of text matched with the scene dictionary model;And using semantic understanding model and dialog model guiding user to hand over
Mutually.The present invention program realizes computer to the automatic deep understanding of the sentence or phrase of natural language etc. and dialogue interaction, full
Foot user needs machine accurately to understand automatically the purpose of interaction pragmatic.
Description of the drawings
Fig. 1 is the explanation block diagram of the semantic understanding with dialog generation system of embodiment of the present invention displaying.
Fig. 2 is the explanation schematic diagram of user's scene dictionary model in embodiment.
Fig. 3 is the explanation schematic diagram of grammer dependency tree in embodiment.
Specific embodiment
In order that those skilled in the art more fully understand the scheme of the embodiment of the present invention, below in conjunction with the accompanying drawings and implement
Mode is described in further detail to the embodiment of the present invention.
For natural language understanding system in prior art realize it is complicated, and depend critically upon linguistic expertise knowledge
Problem, the present invention provide a kind of natural language understanding of displaying and dialogue generation method and system, and it is right simply and efficiently to realize
The natural language understanding of sentence, phrase etc., realizes fully automated understanding of the computer automatically to short text.
The semantic understanding of the displaying of embodiment and dialogue generation method, specifically include following steps:
Step 1, user's model of place is set up, user's model of place is related to three partial contents, Part I is scene word
Allusion quotation model, Part II are Scene Semantics analytic modell analytical model and scene dialog model.
Step 1-1, user's scene dictionary model is set up, i.e., domain features are described by composite key;Set up scene
Semantic understanding rule under semantic analytic modell analytical model, the i.e. special scenes.
Step 1-2, the scene dialog model for setting up user, relate generally to the conversation content under concrete scene, user are drawn
Lead the conversational operation of user.
Step 2, the scene that determination current sessions are selected according to the result of Words partition system.
Step 2-1, Words partition system is called to carry out participle herein to current input.
Step 2-2, word segmentation result and user-defined scene dictionary model are matched, should with determine current session
Use scene.
The semantic parsing final to determine dialogue of the state of step 2-3, the scene talked with according to this wheel and conversational system
Device and dialog generator start new dialogue.
The Scene Semantics resolver that step 3, employing are selected, based in interaction of the corresponding displaying user model to epicycle
Appearance is understood.
Step 3-1, third-party Chinese grammar parser is called to carry out syntax parsing to this conversation content.
Step 3-2, semantic solution is carried out to syntax parsing result with reference to user's scene dictionary model, Scene Semantics analytic modell analytical model
Analysis, obtains the semantic understanding result of epicycle.
Step 4, call corresponding scene dialog generator, with reference to dialogue management intermediateness engage in the dialogue it is comprehensive and raw
Dialogue result into after epicycle interaction.
Step 4-1, the result based on step 2 select the dialog generator under corresponding scene.
Step 4-2, the output result based on the current round obtained by step 3, and the dialogue shape in dialog manager
State engages in the dialogue synthesis, so as to obtain ending the semantic understanding result after epicycle interaction.
Step 5, receive again new text input after, repeat step 2-4 carries out semantic understanding, until conversational system tie
Beam.
As shown in figure 1, being a kind of structured flowchart of the semantic understanding with dialogue generation of embodiment of the present invention displaying.At this
In embodiment, system includes:
User's model of place, there is provided user is generated to special scenes drag, semantic understanding model and corresponding dialogue
Model;
Scene selector, it is for the natural language of receiving user's input and corresponding to select according to user's model of place
User's scene;
Dialog manager, for the state of the intermediate result and correlation of management of dialogs system;
The semantic parser of displaying, implements semantic understanding for the result after syntax parsing is combined user's model of place
Obtain the semantic understanding result of displaying;
Dialogue synthesizer, for the understanding output of the intermediate result and epicycle of comprehensive conversational system, finally obtains what is completed
Text output.
Embodiment provide displaying semantic understanding with dialogue generate, pre-build store it is specific defined in user
Scene dictionary, after the natural language of receiving user's input, the natural language corresponding content of text and institute dictionary is carried out
Match and initiate to talk with to obtain complete semantic input.The present invention can automatically realize text by introducing the model of displaying
The scene of this content determines, and dialogue can be initiated based on scene dialog model.
As shown in Fig. 2 being scene dictionary constructed during the semantic understanding of embodiment of the present invention displaying is generated with dialogue
The schematic diagram of model.The scene dictionary model is mainly field layer and domain features layer including two levels.Field layer is corresponded to
Different application scenarios, and domain features layer has then corresponded to the characteristic vector under the scene, domain model vector can be expressed as
VD=(v1,v2,...,vn).For input text D, the bag of words (Word of Bag) of document are obtained after participle, can be with
With being expressed as set D={ w1,w2,...,wm}.The present invention based on traditional based on bag of words, based on set D and field
Vector set { D1,D2,...Dk, each field DiThere is characteristic vectorIn order to determine which scene input text D is subordinate to,
The present invention performs the steps:
For the characteristic vector of certain sceneThe bag of words of input text, calculate the word in bag of words in characteristic vector
Matching degree m of each componenti(1≤i≤n);
The bag of words of input text and the degree of correlation of the scene are calculated by the method for weighting:
Wherein WiFor certain vectorial weight;
Compare matching degree M of each scenei, maximum is obtained, is exactly scene determined by invention.
For example, the text message of natural language input, such as:" with the blue and white porcelain of KuGoo music Zhou Jielun ", Jing Guofen
" with KuGoo music Zhou Jielun blue and white porcelain " is obtained after word system participle, by implementing to match with dictionary model library, can be with portion
The verb " broadcasting " divided under Music fields and player " KuGoo music ", match object " blue and white porcelain " in the case completely,
Can determine that the content designed by the text is under the jurisdiction of Music fields.Obtain the characteristic vector in Music fields for one 4 tie up to
Amount Music=(music1, music2, music3, music4), player, action, singer and song title are corresponded to respectively.Cause
This algorithm will call the semantic parser in Music fields, and while initialize the dialog generator in Music fields.
After field resolver is determined, algorithm calls corresponding field resolver to carry out semantic understanding to text, and language
The premise of reason and good sense solution is to need to carry out text corresponding grammer participle, and the present invention is mainly using StanfordParser come to text
Originally carry out the syntactic analysis based on grammer dependency tree.In the present invention, just on the basis of participle is done to carrying out language to text
Method parsing obtains corresponding grammer dependency tree, obtains grammer dependency tree as shown in Figure 3.Result and phase based on grammer dependency tree
The domain lexicon model of pass, embodiment will obtain the field structure information after semantic understandingWhich walks substantially
It is rapid as follows:
1) grammer dependency tree is traveled through, corresponding text input vocabulary w is extracted according to grammer dependent Rulei;
2) Integrated Understanding result obtains structurized output.
The text input for continuing the above, the result of syntax parsing is:
1 use _ P P_3prep__
2 KuGoo music _ NR NR_1pobj__
3 broadcasting _ VV VV_0root__
4 Zhou Jielun _ NN NN_5nn__
5 blue and white porcelains _ NN NN_3dobj__
Can determine that title of the song is according to syntactic analysis result:Blue and white porcelain (dobj), Ge Zhewei:Zhou Jielun (nn), now
To user vector U=(" KuGoo music ", " broadcasting ", " Zhou Jielun ", " blue and white porcelain "), all properties of Music dictionaries are with full
Foot event is without the need for implementing dialogue.
In actual application, situation about being frequently encountered is, user will not disposable its intention of expressed intact, such as " I
Want to listen music ".In the case, the present invention guides user to complete complete intention expression by corresponding dialogue is initiated.
For " I wants to listen music ", result " I wants to listen music " is obtained after participle, grammer solution is carried out according to above step
After analysis, following analysis result is obtained:
1 I _ PN PN_2nsubj__
2 want to listen _ VV VV_0root__
3 music _ NN NN_2dobj__
After syntax parsing, corresponding semantic understanding result is obtained, and U=as follows (0, " listening ", 0,0), now use
The intention at family cannot be expressly understood that needs further obtain user's request by talking with.During with user session, need
The step of execution user view merges.
It is assumed that the domain features vector of user is VD=(v1,v2,...,vn), after interaction, the result for obtaining isWherein ui∈ ∪ D ∪ { 0 }, ∪ D represent the input vocabulary of up to the present user mutual, and { 0 } represents
The vectorial user is not input at present, needs to further confirm that by user mutual.
After user engages in the dialogue interaction, the dialogue result for merging current result and epicycle is finally obtained by the present invention
Final result.It is assumed that before epicycle is talked with, the result that user obtains isAnd in epicycle dialogue,
The result obtained by algorithm isIt is defeated that the present invention will obtain newest result according to following process
Go out
In resulting user vector, all of vector neither for sky is, end-of-dialogue.
Continue example above, by understanding " I will listen music ", it may be determined that for Music fields, by semantic understanding
Afterwards, can with user vector U=(0, " listening ", 0,0), according to rule above, need to initiate dialogue, user continues input " I
Blue and white porcelain to be listened ", after user understands, the user output U'=that obtains thinking (0, " listening ", 0, " blue and white porcelain "), it is intended that merging is by base
In the new result of the two results:
U'=(0, " listening ", 0, " blue and white porcelain ")
The apparent intention does not still know, therefore, user input " blue and white porcelain of Zhou Jielun " is used after semantic parsing
Family input U'=(0,0, " Zhou Jielun ", " blue and white porcelain "), exported after further merging:
U'=(0, " listening ", " Zhou Jielun ", " blue and white porcelain ")
Further, user input " using Netease's cloud music ", semantic parsing have obtained result U'=(" Netease cloud sound
It is happy ", " broadcasting ", 0,0), further fusion obtains result:
U'=(" Netease's cloud music ", " broadcasting ", " Zhou Jielun ", " blue and white porcelain ")
So far, the explicit requirement of user, semantic understanding are completed.
Claims (10)
1. a kind of semantic understanding of displaying with dialogue generation method, it is characterised in that:Comprise the following steps,
Step 1, user's model of place is set up, including user's scene dictionary model, Scene Semantics analytic modell analytical model and user's scene pair
Words model;
Step 2, the field that determination current sessions are selected according to the result of Words partition system;
The Scene Semantics resolver that step 3, employing are selected, based on corresponding user's scene dictionary model, Scene Semantics parsing mould
Type is understood to the interaction content of epicycle;
Step 4, based on user's scene dialog model, create the dialog generator under corresponding scene, with reference to the centre of dialogue management
State engage in the dialogue synthesis and generate epicycle interaction after dialogue result;
Step 5, receive again new text input after, repeat step 2-4 carries out semantic understanding, until conversational system terminates.
2. the semantic understanding of displaying as claimed in claim 1 with dialogue generation method, it is characterised in that:Step 1 is specially:
Step 1-1, user's scene dictionary model is set up, domain features are described by composite key, and sets up scene language
Semantic understanding rule under adopted analytic modell analytical model, the i.e. special scenes;
Step 1-2, user's scene dialog model is set up, the dialogue interactive mode being related under concrete scene, for guiding user's
Dialogue interaction.
3. the semantic understanding of displaying as claimed in claim 1 with dialogue generation method, it is characterised in that:Step 2 is specially:
Step 2-1, call Words partition system participle to be carried out herein to current input, obtain the bag of words of text;
Step 2-2, word segmentation result and user-defined scene universal model dictionary are matched, should with determine current session
Use scene;
Step 2-3, according to the state of the scene and conversational system of this wheel dialogue come determine the final semantic parser of dialogue and
Dialog generator starts new dialogue.
4. the semantic understanding of displaying as claimed in claim 3 with dialogue generation method, it is characterised in that:In step 2-2, neck
Domain model dictionary includes field layer and domain features layer, and field layer corresponds to different application scenarios, and domain features layer is to answering
With the characteristic vector under scene.
5. the semantic understanding of displaying as claimed in claim 4 with dialogue generation method, it is characterised in that:In step 2-2, really
Determine the application scenarios of current session, specially:
Step 2-2-1, for the characteristic vector of certain sceneThe bag of words of input text, calculate the word in bag of words to feature
Matching degree m of each component in vectori, wherein 1≤i≤n;
Step 2-2-2, the degree of correlation that the bag of words and the scene that are input into text are calculated by the method for weighting:
Wherein, WiFor certain vectorial weight;
Step 2-2-3, matching degree M for comparing each scenej, obtain maximum, be exactly determined by application scenarios.
6. the semantic understanding of displaying as claimed in claim 3 with dialogue generation method, it is characterised in that:In step 2-3,
When the intention of user cannot be expressly understood that, user's request is obtained by talking with, it is automatically true by the model for introducing displaying
Determine the application scenarios of content of text, and based on scene dialog model initiating dialogue.
7. the semantic understanding of displaying as claimed in claim 6 with dialogue generation method, it is characterised in that:In step 2-3,
During user session, user view is performed the step of merge, specially:
It is assumed that before epicycle is talked with, the result that user obtains isAnd in epicycle dialogue, by algorithm
The result for obtaining isNewest result output is obtained according to following process
In resulting user vector, all of vector is neither space-time, end-of-dialogue.
8. the semantic understanding of the displaying as described in any one of claim 1-7 with dialogue generation method, it is characterised in that:Step
3 are specially:
Step 3-1, third-party Chinese grammar parser is called to carry out syntax parsing to this conversation content;
Step 3-2, semantic parsing is carried out to syntax parsing result with reference to user's scene dictionary model, Scene Semantics analytic modell analytical model,
Obtain the semantic understanding result of epicycle.
9. the semantic understanding of the displaying as described in any one of claim 1-7 with dialogue generation method, it is characterised in that:Step
4 are specially:
Step 4-1, the result based on step 2 select the dialog generator under corresponding scene;
Step 4-2, the output result based on the current round obtained by step 3, and the dialogue state in dialog manager enters
Row dialogue is comprehensive, so as to obtain ending the semantic understanding result after epicycle interaction.
10. a kind of semantic understanding for realizing the displaying described in any one of claim 1-9 with dialogue generation method system, its
It is characterised by, including:
User's model of place, there is provided general dictionary model, semantic understanding model and dialogue generation model under special scenes;
Scene selector, for the natural language of receiving user's input, and according to user's model of place selecting corresponding user
Scene;Pre-build and store special scenes dictionary defined in user, after the natural language of receiving user's input, will be described
The corresponding content of text of natural language is matched with institute dictionary, initiates dialogue when semantic input is imperfect complete to obtain
Semantic input;
Dialog manager:For the state of the intermediate result and correlation of management of dialogs system;
The semantic parser of displaying:Obtain for the result after syntax parsing is implemented semantic understanding with reference to user's model of place
The semantic understanding result of displaying;
Dialogue synthesizer:For the understanding output of the intermediate result and epicycle of comprehensive conversational system, the text for completing is finally obtained
Output.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610747188.5A CN106528522A (en) | 2016-08-26 | 2016-08-26 | Scenarized semantic comprehension and dialogue generation method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610747188.5A CN106528522A (en) | 2016-08-26 | 2016-08-26 | Scenarized semantic comprehension and dialogue generation method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106528522A true CN106528522A (en) | 2017-03-22 |
Family
ID=58343593
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610747188.5A Pending CN106528522A (en) | 2016-08-26 | 2016-08-26 | Scenarized semantic comprehension and dialogue generation method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106528522A (en) |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107273406A (en) * | 2017-04-27 | 2017-10-20 | 上海奔影网络科技有限公司 | Dialog process method and device in task dialogue system |
CN107357787A (en) * | 2017-07-26 | 2017-11-17 | 微鲸科技有限公司 | Semantic interaction method, apparatus and electronic equipment |
CN107644641A (en) * | 2017-07-28 | 2018-01-30 | 深圳前海微众银行股份有限公司 | Session operational scenarios recognition methods, terminal and computer-readable recording medium |
CN107704521A (en) * | 2017-09-07 | 2018-02-16 | 北京零秒科技有限公司 | A kind of question and answer processing server, client and implementation method |
CN108114469A (en) * | 2018-01-29 | 2018-06-05 | 北京神州泰岳软件股份有限公司 | Game interaction method, apparatus, terminal and game interaction model based on dialogue |
CN108170764A (en) * | 2017-12-25 | 2018-06-15 | 上海大学 | A kind of man-machine more wheel dialog model construction methods based on scene context |
CN108304561A (en) * | 2018-02-08 | 2018-07-20 | 北京信息职业技术学院 | A kind of semantic understanding method, equipment and robot based on finite data |
CN109032731A (en) * | 2018-07-09 | 2018-12-18 | 西安电子科技大学 | A kind of voice interface method and system based on semantic understanding of oriented manipulation system |
CN109325155A (en) * | 2018-07-25 | 2019-02-12 | 南京瓦尔基里网络科技有限公司 | A kind of novel dialogue state storage method and system |
CN109543182A (en) * | 2018-11-15 | 2019-03-29 | 广东电网有限责任公司信息中心 | A kind of electric power enterprise based on solr engine takes turns interactive semantic analysis method more |
CN110290258A (en) * | 2018-03-19 | 2019-09-27 | 深圳奥斯比特科技有限公司 | A kind of method that chat robots are chatted using displaying |
CN111046146A (en) * | 2019-12-27 | 2020-04-21 | 北京百度网讯科技有限公司 | Method and apparatus for generating information |
CN111190522A (en) * | 2018-11-13 | 2020-05-22 | 奥多比公司 | Generating three-dimensional digital content from natural language requests |
CN111261150A (en) * | 2018-11-30 | 2020-06-09 | 阿里巴巴集团控股有限公司 | Method and device for realizing man-machine conversation |
CN111368549A (en) * | 2018-12-25 | 2020-07-03 | 深圳市优必选科技有限公司 | Natural language processing method, device and system supporting multiple services |
WO2020147609A1 (en) * | 2019-01-18 | 2020-07-23 | 阿里巴巴集团控股有限公司 | Speech recognition method and apparatus |
CN112204656A (en) * | 2018-05-29 | 2021-01-08 | 简单对话合同会社 | Efficient dialog configuration |
CN112823332A (en) * | 2018-10-10 | 2021-05-18 | N3有限责任公司 | Semantic industry terminology |
CN113407795A (en) * | 2021-06-22 | 2021-09-17 | 海南嗨宝科技有限公司 | Chat robot based on natural language real-time scene generation |
CN113515952A (en) * | 2021-08-18 | 2021-10-19 | 内蒙古工业大学 | Mongolian dialogue model combined modeling method, system and equipment |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102855874A (en) * | 2012-09-07 | 2013-01-02 | 深圳市信利康电子有限公司 | Method and system for controlling household appliance on basis of voice interaction of internet |
CN102868695A (en) * | 2012-09-18 | 2013-01-09 | 天格科技(杭州)有限公司 | Conversation tree-based intelligent online customer service method and system |
CN103000173A (en) * | 2012-12-11 | 2013-03-27 | 优视科技有限公司 | Voice interaction method and device |
CN104464733A (en) * | 2014-10-28 | 2015-03-25 | 百度在线网络技术(北京)有限公司 | Multi-scene managing method and device of voice conversation |
CN105096138A (en) * | 2015-06-30 | 2015-11-25 | 百度在线网络技术(北京)有限公司 | Method and device for realizing O2O dialogue interaction |
CN105512228A (en) * | 2015-11-30 | 2016-04-20 | 北京光年无限科技有限公司 | Bidirectional question-answer data processing method and system based on intelligent robot |
CN105786798A (en) * | 2016-02-25 | 2016-07-20 | 上海交通大学 | Natural language intention understanding method in man-machine interaction |
-
2016
- 2016-08-26 CN CN201610747188.5A patent/CN106528522A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102855874A (en) * | 2012-09-07 | 2013-01-02 | 深圳市信利康电子有限公司 | Method and system for controlling household appliance on basis of voice interaction of internet |
CN102868695A (en) * | 2012-09-18 | 2013-01-09 | 天格科技(杭州)有限公司 | Conversation tree-based intelligent online customer service method and system |
CN103000173A (en) * | 2012-12-11 | 2013-03-27 | 优视科技有限公司 | Voice interaction method and device |
CN104464733A (en) * | 2014-10-28 | 2015-03-25 | 百度在线网络技术(北京)有限公司 | Multi-scene managing method and device of voice conversation |
CN105096138A (en) * | 2015-06-30 | 2015-11-25 | 百度在线网络技术(北京)有限公司 | Method and device for realizing O2O dialogue interaction |
CN105512228A (en) * | 2015-11-30 | 2016-04-20 | 北京光年无限科技有限公司 | Bidirectional question-answer data processing method and system based on intelligent robot |
CN105786798A (en) * | 2016-02-25 | 2016-07-20 | 上海交通大学 | Natural language intention understanding method in man-machine interaction |
Cited By (25)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107273406A (en) * | 2017-04-27 | 2017-10-20 | 上海奔影网络科技有限公司 | Dialog process method and device in task dialogue system |
CN107357787B (en) * | 2017-07-26 | 2021-04-02 | 微鲸科技有限公司 | Semantic interaction method and device and electronic equipment |
CN107357787A (en) * | 2017-07-26 | 2017-11-17 | 微鲸科技有限公司 | Semantic interaction method, apparatus and electronic equipment |
CN107644641A (en) * | 2017-07-28 | 2018-01-30 | 深圳前海微众银行股份有限公司 | Session operational scenarios recognition methods, terminal and computer-readable recording medium |
CN107644641B (en) * | 2017-07-28 | 2021-04-13 | 深圳前海微众银行股份有限公司 | Dialog scene recognition method, terminal and computer-readable storage medium |
CN107704521A (en) * | 2017-09-07 | 2018-02-16 | 北京零秒科技有限公司 | A kind of question and answer processing server, client and implementation method |
CN108170764A (en) * | 2017-12-25 | 2018-06-15 | 上海大学 | A kind of man-machine more wheel dialog model construction methods based on scene context |
CN108170764B (en) * | 2017-12-25 | 2022-06-14 | 上海大学 | Man-machine multi-turn dialogue model construction method based on scene context |
CN108114469A (en) * | 2018-01-29 | 2018-06-05 | 北京神州泰岳软件股份有限公司 | Game interaction method, apparatus, terminal and game interaction model based on dialogue |
CN108304561A (en) * | 2018-02-08 | 2018-07-20 | 北京信息职业技术学院 | A kind of semantic understanding method, equipment and robot based on finite data |
CN108304561B (en) * | 2018-02-08 | 2019-03-29 | 北京信息职业技术学院 | A kind of semantic understanding method, equipment and robot based on finite data |
CN110290258A (en) * | 2018-03-19 | 2019-09-27 | 深圳奥斯比特科技有限公司 | A kind of method that chat robots are chatted using displaying |
CN112204656A (en) * | 2018-05-29 | 2021-01-08 | 简单对话合同会社 | Efficient dialog configuration |
CN109032731A (en) * | 2018-07-09 | 2018-12-18 | 西安电子科技大学 | A kind of voice interface method and system based on semantic understanding of oriented manipulation system |
CN109325155A (en) * | 2018-07-25 | 2019-02-12 | 南京瓦尔基里网络科技有限公司 | A kind of novel dialogue state storage method and system |
CN112823332A (en) * | 2018-10-10 | 2021-05-18 | N3有限责任公司 | Semantic industry terminology |
CN111190522A (en) * | 2018-11-13 | 2020-05-22 | 奥多比公司 | Generating three-dimensional digital content from natural language requests |
CN109543182A (en) * | 2018-11-15 | 2019-03-29 | 广东电网有限责任公司信息中心 | A kind of electric power enterprise based on solr engine takes turns interactive semantic analysis method more |
CN111261150A (en) * | 2018-11-30 | 2020-06-09 | 阿里巴巴集团控股有限公司 | Method and device for realizing man-machine conversation |
CN111368549A (en) * | 2018-12-25 | 2020-07-03 | 深圳市优必选科技有限公司 | Natural language processing method, device and system supporting multiple services |
WO2020147609A1 (en) * | 2019-01-18 | 2020-07-23 | 阿里巴巴集团控股有限公司 | Speech recognition method and apparatus |
CN111046146A (en) * | 2019-12-27 | 2020-04-21 | 北京百度网讯科技有限公司 | Method and apparatus for generating information |
CN113407795A (en) * | 2021-06-22 | 2021-09-17 | 海南嗨宝科技有限公司 | Chat robot based on natural language real-time scene generation |
CN113515952A (en) * | 2021-08-18 | 2021-10-19 | 内蒙古工业大学 | Mongolian dialogue model combined modeling method, system and equipment |
CN113515952B (en) * | 2021-08-18 | 2023-09-12 | 内蒙古工业大学 | Combined modeling method, system and equipment for Mongolian dialogue model |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106528522A (en) | Scenarized semantic comprehension and dialogue generation method and system | |
US11816435B1 (en) | Applied artificial intelligence technology for contextualizing words to a knowledge base using natural language processing | |
US11645547B2 (en) | Human-machine interactive method and device based on artificial intelligence | |
CN106649825B (en) | Voice interaction system and creation method and device thereof | |
TW201833903A (en) | Conversation generating method, device, and storage medium | |
CN103440234B (en) | Natural language understanding system and method | |
WO2017196784A1 (en) | Ontology discovery based on distributional similarity and lexico-semantic relations | |
KR102445519B1 (en) | System and method for manufacturing conversational intelligence service providing chatbot | |
CN108538294B (en) | Voice interaction method and device | |
JP2009036999A (en) | Interactive method using computer, interactive system, computer program and computer-readable storage medium | |
US11830482B2 (en) | Method and apparatus for speech interaction, and computer storage medium | |
WO2021179910A1 (en) | Text voice front-end conversion method and apparatus, and device and storage medium | |
WO2021073179A1 (en) | Named entity identification method and device, and computer-readable storage medium | |
CN108345612A (en) | A kind of question processing method and device, a kind of device for issue handling | |
CN108763195A (en) | A kind of non-limiting type relation excavation method based on interdependent syntax and pattern rules | |
WO2024011813A1 (en) | Text expansion method and apparatus, device, and medium | |
CN109933773A (en) | A kind of multiple semantic sentence analysis system and method | |
WO2022142823A1 (en) | Human-machine conversation method and apparatus, computer device, and readable storage medium | |
CN112559760B (en) | CPS (cyber physical system) resource capacity knowledge graph construction method for text description | |
CN109885835B (en) | Method and system for acquiring association relation between words in user corpus | |
CN110931002B (en) | Man-machine interaction method, device, computer equipment and storage medium | |
CN113779062A (en) | SQL statement generation method and device, storage medium and electronic equipment | |
Galley et al. | Hybrid natural language generation for spoken dialogue systems | |
CN108763355A (en) | A kind of intelligent robot interaction data processing system and method based on user | |
CN109656952B (en) | Query processing method and device and electronic equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170322 |
|
RJ01 | Rejection of invention patent application after publication |