CN106599998B - The method and system answered based on affective characteristics adjustment robot - Google Patents

The method and system answered based on affective characteristics adjustment robot Download PDF

Info

Publication number
CN106599998B
CN106599998B CN201611093206.9A CN201611093206A CN106599998B CN 106599998 B CN106599998 B CN 106599998B CN 201611093206 A CN201611093206 A CN 201611093206A CN 106599998 B CN106599998 B CN 106599998B
Authority
CN
China
Prior art keywords
information
user
text information
past
current
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.)
Active
Application number
CN201611093206.9A
Other languages
Chinese (zh)
Other versions
CN106599998A (en
Inventor
简仁贤
林志豪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Intelligent Technology (shanghai) Co Ltd
Original Assignee
Intelligent Technology (shanghai) Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Intelligent Technology (shanghai) Co Ltd filed Critical Intelligent Technology (shanghai) Co Ltd
Priority to CN201611093206.9A priority Critical patent/CN106599998B/en
Publication of CN106599998A publication Critical patent/CN106599998A/en
Application granted granted Critical
Publication of CN106599998B publication Critical patent/CN106599998B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00Computing arrangements using knowledge-based models
    • G06N5/02Knowledge representation; Symbolic representation
    • G06N5/022Knowledge engineering; Knowledge acquisition
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00Handling natural language data
    • G06F40/30Semantic analysis

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Computational Linguistics (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • General Health & Medical Sciences (AREA)
  • Manipulator (AREA)
  • Toys (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The present invention provides a kind of method and system answered based on affective characteristics adjustment robot, methods are as follows: obtains the current text information and text information in the past of user's input;According to current text information and past text information, the characteristic information of user is extracted, characteristic information includes user feeling characteristic information and user's meaning of one's words characteristic information;Robot affective characteristics information is obtained, in conjunction with user feeling characteristic information, information integration is carried out, obtains both sides' affective characteristics value;The corresponding answer of current text information is provided in conjunction with both sides' affective characteristics value according to user's meaning of one's words characteristic information.The present invention is based on the current text information and past text information during human-computer dialogue, extract user feeling characteristic information and robot affective characteristics information, in conjunction with user's meaning of one's words characteristic information, the answer for providing robot has the emotion of robot, human-computer dialogue is more intelligent, further improves user experience.

Description

The method and system answered based on affective characteristics adjustment robot
Technical field
The present invention relates to artificial intelligence field more particularly to interactive system fields.
Background technique
In existing artificial intelligence conversational system, judge that user talks using corpus and template, conversational system Suitable answer is found usually using the mode of search.But there is no adjusted back using the emotion of user as clue Strategy is answered, other than considering user emotion, robot should also select the strategy responded based on the mood of itself.In person to person Natural dialogue on, emotion is a key factor, the mankind can according to expressed by other side emotion and itself existing emotion come The answer strategy for adjusting oneself, achievees the effect that communication.
Therefore, defect in the prior art is that in artificial intelligence dialog procedure, robot is unable to base to the answer of user Emotion in the text information of user's input, and the emotion of robot itself is combined to make intelligent response, it can not provide It accurately answers, keeps user experience low.
Summary of the invention
In view of the above technical problems, the present invention provide it is a kind of based on affective characteristics adjustment robot answer method and be System extracts user feeling characteristic information and machine based on the current text information and past text information during human-computer dialogue Human feelings sense characteristic information, in conjunction with user's meaning of one's words characteristic information, the answer for providing robot has the emotion of robot, man-machine Dialogue is more intelligent, further improves user experience.
In order to solve the above technical problems, present invention provide the technical scheme that
In a first aspect, the present invention provides a kind of method answered based on affective characteristics adjustment robot, comprising:
Step S1 obtains the current text information and past text information of user's input;
Step S2 extracts the characteristic information of the user, institute according to the current text information and past text information Stating characteristic information includes user feeling characteristic information and user's meaning of one's words characteristic information;
Step S3 obtains robot affective characteristics information, in conjunction with the user feeling characteristic information, carries out information integration, Both sides' affective characteristics value is obtained, the robot affective characteristics information is that the information interacted by the past with the user is computed It obtains;
Step S4, according to user's meaning of one's words characteristic information, in conjunction with both sides' affective characteristics value, provide it is described ought be above The corresponding answer of this information.
The present invention provides a kind of method answered based on affective characteristics adjustment robot, technical solutions are as follows: first obtain The current text information of user's input and in the past text information;Then, according to the current text information and in the past text information, The characteristic information of the user is extracted, the characteristic information includes user feeling characteristic information and user's meaning of one's words characteristic information; Then, robot affective characteristics information is obtained, in conjunction with the user feeling characteristic information, information integration is carried out, obtains both sides' feelings Feel characteristic value, the robot affective characteristics information is that the information interacted by the past with the user is computed acquisition;Most Afterwards, it is corresponding to be provided in conjunction with both sides' affective characteristics value according to user's meaning of one's words characteristic information for the current text information It answers.
The method answered based on affective characteristics adjustment robot of the invention, based on the current text during human-computer dialogue Information and in the past text information, extract user feeling characteristic information and robot affective characteristics information, special in conjunction with user's meaning of one's words Reference breath, the answer for providing robot have the emotion of robot, and human-computer dialogue is more intelligent, further improves user Experience.
Further, the step S2, specifically:
According to the current text information and past text information, the current text information and in the past text information are obtained In emotional state information, the emotional state information include current emotional states information, user go over main emotional information and At least one of past secondary emotional information;
According to the current text information and past text information, the current text information and in the past text information are obtained In text message, the text message includes the current text information and the semantics information in text information, key in the past At least one of word information, proper noun information and verb information;
According to the current text information and past text information, the current text information and in the past text information are obtained In topic information, the topic information includes that current staple of conversation information, current secondary topic information, the past staple of conversation are believed Breath, past secondary topic information like topic information and at present at least one of hot topic information;
According to the current text information and past text information, the current text information and in the past text information are obtained In sentence pattern, tone information and language performance information, the language performance information include current dominant language behavioural information and time Want language performance information;
According to the emotional state information, text message, topic information, sentence pattern, tone information and language performance information, mention The characteristic information of the user is taken out, the characteristic information includes user feeling characteristic information and user's meaning of one's words characteristic information.
Further, in the step S4, specifically:
Data bank is obtained, multiple answers are stored in advance in the data bank;
According to user's meaning of one's words characteristic information, in conjunction with both sides' affective characteristics value, by machine learning algorithm, in institute It states and selects one to respond as the corresponding answer of the current text information in data bank.
Further, in the step S4, specifically:
According to user's meaning of one's words characteristic information, in conjunction with both sides' affective characteristics value, by the method for logic rules, certainly It is dynamic to generate the corresponding answer of the current text information:
According to the user semantic characteristic information, emotional characteristics in the semantic feature information are analyzed;
A kind of logical combination is formed in conjunction with the mood of the robot according to emotional characteristics in the semantic feature information;
According to the logical combination, automatically generated from the answer strategy that the method by logic rules pre-establishes described The corresponding answer of current text information.
Further, in the step S4, specifically:
According to user's meaning of one's words characteristic information, in conjunction with both sides' affective characteristics value, by the method for machine learning, certainly It is dynamic to generate the corresponding answer of the current text information:
According to the user semantic characteristic information, emotional characteristics in the semantic feature information are analyzed;
A kind of Response Policy is formed in conjunction with the mood of the robot according to emotional characteristics in the semantic feature information;
According to the Response Policy, automatically generates or select from the answer strategy that the method by machine learning pre-establishes It selects and generates the corresponding answer of the current text information.
Second aspect, the present invention provides a kind of systems answered based on affective characteristics adjustment robot, comprising:
Text information obtains module, for obtaining the current text information and past text information of user's input;
Talk with identification module, for extracting the user's according to the current text information and past text information Characteristic information, the characteristic information include user feeling characteristic information and user's meaning of one's words characteristic information;
Robot emotion module, for obtaining robot affective characteristics information, in conjunction with the user feeling characteristic information, into Row information integration, obtains both sides' affective characteristics value, the robot affective characteristics information is to interact by the past with the user Information be computed acquisition;
Talk with response means, for providing according to user's meaning of one's words characteristic information in conjunction with both sides' affective characteristics value The corresponding answer of the current text information.
A kind of system answered based on affective characteristics adjustment robot provided by the invention, technical solution are as follows: first pass through Text information obtains module, for obtaining the current text information and past text information of user's input;Then known by dialogue Other module, for extracting the characteristic information of the user, the spy according to the current text information and past text information Reference breath includes user feeling characteristic information and user's meaning of one's words characteristic information;Then by robot emotion module, for obtaining Robot affective characteristics information carries out information integration, obtains both sides' affective characteristics value in conjunction with the user feeling characteristic information, The robot affective characteristics information is that the information interacted by the past with the user is computed acquisition;Finally by dialogue Response means, for providing the current text in conjunction with both sides' affective characteristics value according to user's meaning of one's words characteristic information The corresponding answer of information.
The system answered based on affective characteristics adjustment robot of the invention, based on the current text during human-computer dialogue Information and in the past text information, extract user feeling characteristic information and robot affective characteristics information, special in conjunction with user's meaning of one's words Reference breath, the answer for providing robot have the emotion of robot, and human-computer dialogue is more intelligent, further improves user Experience.
Further, the dialogue identification module, is specifically used for:
According to the current text information and past text information, the current text information and in the past text information are obtained In emotional state information, the emotional state information include current emotional states information, user go over main emotional information and At least one of past secondary emotional information;
According to the current text information and past text information, the current text information and in the past text information are obtained In text message, the text message includes the current text information and the semantics information in text information, key in the past At least one of word information, proper noun information and verb information;
According to the current text information and past text information, the current text information and in the past text information are obtained In topic information, the topic information includes that current staple of conversation information, current secondary topic information, the past staple of conversation are believed Breath, past secondary topic information like topic information and at present at least one of hot topic information;
According to the current text information and past text information, the current text information and in the past text information are obtained In sentence pattern, tone information and language performance information, the language performance information include current dominant language behavioural information and time Want language performance information;
According to the emotional state information, text message, topic information, sentence pattern, tone information and language performance information, mention The characteristic information of the user is taken out, the characteristic information includes user feeling characteristic information and user's meaning of one's words characteristic information.
Further, the dialogue response means, are specifically used for:
Data bank is obtained, multiple answers are stored in advance in the data bank;
According to user's meaning of one's words characteristic information, in conjunction with both sides' affective characteristics value, by machine learning algorithm, in institute It states and selects one to respond as the corresponding answer of the current text information in data bank.
Further, the dialogue response means, are specifically used for:
According to user's meaning of one's words characteristic information, in conjunction with both sides' affective characteristics value, by the method for logic rules, certainly It is dynamic to generate the corresponding answer of the current text information:
According to the user semantic characteristic information, emotional characteristics in the semantic feature information are analyzed;
A kind of logical combination is formed in conjunction with the mood of the robot according to emotional characteristics in the semantic feature information;
According to the logical combination, automatically generated from the answer strategy that the method by logic rules pre-establishes described The corresponding answer of current text information.
Further, the dialogue response means, are specifically used for:
According to user's meaning of one's words characteristic information, in conjunction with both sides' affective characteristics value, by the method for machine learning, certainly It is dynamic to generate the corresponding answer of the current text information:
According to the user semantic characteristic information, emotional characteristics in the semantic feature information are analyzed;
A kind of Response Policy is formed in conjunction with the mood of the robot according to emotional characteristics in the semantic feature information;
According to the Response Policy, automatically generates or select from the answer strategy that the method by machine learning pre-establishes It selects and generates the corresponding answer of the current text information.
Detailed description of the invention
It, below will be to specific in order to illustrate more clearly of the specific embodiment of the invention or technical solution in the prior art Embodiment or attached drawing needed to be used in the description of the prior art are briefly described.
Fig. 1 shows a kind of method answered based on affective characteristics adjustment robot provided by first embodiment of the invention Flow chart;
Fig. 2 shows a kind of systems answered based on affective characteristics adjustment robot provided by second embodiment of the invention Schematic diagram.
Specific embodiment
It is described in detail below in conjunction with embodiment of the attached drawing to technical solution of the present invention.Following embodiment is only used for Clearly illustrate technical solution of the present invention, therefore be intended only as example, and cannot be used as a limitation and limit protection of the invention Range.
Embodiment one
Fig. 1 shows a kind of method answered based on affective characteristics adjustment robot provided by first embodiment of the invention Flow chart;As shown in Figure 1, the embodiment of the present invention one provides a kind of method answered based on affective characteristics adjustment robot, Include:
Step S1 obtains the current text information and past text information of user's input;
Step S2 extracts the characteristic information of user, characteristic information packet according to current text information and past text information Include user feeling characteristic information and user's meaning of one's words characteristic information;
Step S3 obtains robot affective characteristics information, in conjunction with user feeling characteristic information, carries out information integration, obtains Both sides' affective characteristics value, robot affective characteristics information are that the information interacted by the past with user is computed acquisition;
It is corresponding to provide current text information in conjunction with both sides' affective characteristics value according to user's meaning of one's words characteristic information by step S4 It answers.
The present invention provides a kind of method answered based on affective characteristics adjustment robot, technical solutions are as follows: first obtain The current text information of user's input and in the past text information;Then, it according to current text information and past text information, extracts The characteristic information of user out, characteristic information include user feeling characteristic information and user's meaning of one's words characteristic information;Then, machine is obtained Human feelings sense characteristic information carries out information integration, obtains both sides' affective characteristics value, machine human feelings in conjunction with user feeling characteristic information Feeling characteristic information is that the information interacted by the past with user is computed acquisition;Finally, according to user's meaning of one's words characteristic information, knot Both sides' affective characteristics value is closed, the corresponding answer of current text information is provided.
The method answered based on affective characteristics adjustment robot of the invention, based on the current text during human-computer dialogue Information and in the past text information, extract user feeling characteristic information and robot affective characteristics information, special in conjunction with user's meaning of one's words Reference breath, the answer for providing robot have the emotion of robot, and human-computer dialogue is more intelligent, further improves user Experience.
Specifically, step S2, specifically:
According to current text information and past text information, the mood in current text information and in the past text information is obtained Status information, emotional state information include current emotional states information, user's past main emotional information and past secondary mood At least one of information;
According to current text information and past text information, the text in current text information and in the past text information is obtained Message, text message includes current text information and the semantics information in text information, key word information, proper noun are believed in the past At least one of breath and verb information;
According to current text information and past text information, the topic in current text information and in the past text information is obtained Information, topic information includes current staple of conversation information, currently secondary topic information, past staple of conversation information, past are secondary Topic information likes topic information and at present at least one of hot topic information;
According to current text information and past text information, the sentence in current text information and in the past text information is obtained Type, tone information and language performance information, language performance information include current dominant language behavioural information and secondary language performance Information;
According to emotional state information, text message, topic information, sentence pattern, tone information and language performance information, extract The characteristic information of user, characteristic information include user feeling characteristic information and user's meaning of one's words characteristic information.
Current text information combination contextual information can also be rewritten the information of the rewriting sentence obtained after original sentence Be added to extract characteristic information source information in, with emotional state information, text message, topic information, sentence pattern, tone information and Language performance information extracts the characteristic information of user together.
Wherein, the information for rewriting sentence is phenomenon in order to solve some sentence information deficiencies.
Such as:
Do A: you like watching movie?
B: like.
But B merely enters " liking ", information content is insufficient at this time." like seeing if just will become according to rewriting above Film ", computer can make more accurate judgement at this time.
Original sentence: retain the appearance of original sentence, computer capacity is allowed to access original sentence.Sentence is rewritten in cooperation, more It is fully understood by user's input.
Such as:
A: we will go to the cinema tomorrow
B: En En.
At this point, " En En " can be rewritten into " I will go to the cinema tomorrow ".If retaining former sentence " En En ", input computer becomes At " En En+I go to the cinema tomorrow ";Computer is it can be understood that at+response problem of watching movie tomorrow.In this manner into And obtain more complete information.
Sufficiently the current session text information of analysis identification user and past text conversation information, analyze the factor of processing It is more, it can make to extract obtained user feeling characteristic information and user's meaning of one's words characteristic information is more acurrate, and then provide robot Answer it is more acurrate, it is more intelligent.
Specifically, in step S4, specifically:
Data bank is obtained, multiple answers are stored in advance in data bank;
According to user's meaning of one's words characteristic information, selected in data bank in conjunction with both sides' affective characteristics value by machine learning algorithm One is selected to respond as the corresponding answer of current text information.
The answer that robot provides can select a correspondence from data bank by way of selecting and reasonably answer, and lead to The related algorithm for crossing machine learning, the answer for providing robot are more intelligent.
Specifically, in step S4, specifically:
According to user's meaning of one's words characteristic information, automatically generates and work as by the method for logic rules in conjunction with both sides' affective characteristics value The corresponding answer of preceding text information:
According to user semantic characteristic information, emotional characteristics in semantic characteristic information are analyzed;
A kind of logical combination is formed in conjunction with the mood of robot according to emotional characteristics in semantic feature information;
According to logical combination, current text is automatically generated from the answer strategy that the method by logic rules pre-establishes The corresponding answer of information.
Citing is illustrated:
User's input: " I and boyfriend say good-bye ", the meaning of one's words is characterized in: " saying good-bye, boyfriend ", user emotion feature meeting It is biased to " sad ", it is assumed that the mood of robot at this time is " neutrality ", then having higher chance to answer, " I can understand your present sense Feel, I can accompany you chat chat " etc. comfort type reply.If robot emotion is " cold and detached ", most probably respond: " En En refuels ~".
Logic rules among these refer to: various groups of " meaning of one's words feature " and " both sides' mood " that we have pre-established It closes, if the logical combination that " meaning of one's words feature " and " both sides' mood " in current session content is formed, which drops into, passes through logic rules Certain combination of the answer strategy that pre-establishes of method when, one response will be selected in certain answer scope.By patrolling The method for collecting rule carries out human-computer dialogue, to the answer that user is quick, intelligent, while considering the mood of both sides, makes to answer more intelligence Can, it is more humane.
Specifically, in step S4, specifically:
According to user's meaning of one's words characteristic information, automatically generates and work as by the method for machine learning in conjunction with both sides' affective characteristics value The corresponding answer of preceding text information:
According to user semantic characteristic information, emotional characteristics in semantic characteristic information are analyzed;
A kind of Response Policy is formed in conjunction with the mood of robot according to emotional characteristics in semantic feature information;
According to Response Policy, life is automatically generated or selected from the answer strategy that the method by machine learning pre-establishes At the corresponding answer of current text information.
For example:
User's input: " I and boyfriend say good-bye ", the meaning of one's words is characterized in: " saying good-bye, boyfriend ", user emotion feature meeting It is biased to " sad ", it is assumed that the mood of robot at this time is " neutrality ", then having higher chance to answer, " I can understand your present sense Feel, I can accompany you chat chat " etc. comfort type reply.If robot emotion is " cold and detached ", most probably respond: " En En refuels ~".
Machine learning algorithm among these is to allow machine learning " meaning of one's words feature " and " both sides through prior data collection Collocation of the mood " with " automatically generating response " or " how selection is responded ".
It responds feedback according to the practical response situation of every user and experiment in training data to compare, if comparing result It is inconsistent, then update the parameter of machine learning, and repeat iteration, when in training data the response that generates with it is anti-with experiment It is fed back to when answering consistency up to highest, then stops iteration, and training is completed.In the stage of test, when user inputs one When this paper sentence, we can through passing data train come prediction module come responded (can automatically generate or Person, which is that selection is suitable, to be responded).
Answer strategy is pre-established by the method for machine learning, if " meaning of one's words feature " that includes in current text information The answer strategy that " both sides' mood " is formed is matched with some in the answer strategy pre-established, then can be automatically generated current The corresponding answer of text information.Answer strategy if the answer strategy formed does not rain in the answer strategy pre-established Match, then the corresponding answer of similar answer strategy generating current text information may be selected.By the method for machine learning, can be obtained A large amount of and accurate answer strategy provides the answer of accurate intelligence for user.
Embodiment two
Fig. 2 shows a kind of systems answered based on affective characteristics adjustment robot provided by second embodiment of the invention Schematic diagram.As shown in Fig. 2, second embodiment of the present invention provides a kind of systems answered based on affective characteristics adjustment robot 10, comprising:
Text information obtains module 101, for obtaining the current text information and past text information of user's input;
Talk with identification module 102, for extracting the feature letter of user according to current text information and past text information Breath, characteristic information includes user feeling characteristic information and user's meaning of one's words characteristic information;
Robot emotion module 103, for obtaining robot affective characteristics information, in conjunction with user feeling characteristic information, into Row information integration, obtains both sides' affective characteristics value, and robot affective characteristics information is the information warp interacted by the past with user Calculate acquisition;
Talk with response means 104, in conjunction with both sides' affective characteristics value, providing current according to user's meaning of one's words characteristic information The corresponding answer of text information.
A kind of system 10 answered based on affective characteristics adjustment robot provided by the invention, technical solution are as follows: first lead to It crosses text information and obtains module 101, for obtaining the current text information and past text information of user's input;Then by pair Identification module 102 is talked about, for extracting the characteristic information of user, feature letter according to current text information and past text information Breath includes user feeling characteristic information and user's meaning of one's words characteristic information;Then by robot emotion module 103, for obtaining machine Device human feelings sense characteristic information carries out information integration, obtains both sides' affective characteristics value, robot in conjunction with user feeling characteristic information Affective characteristics information is that the information interacted by the past with user is computed acquisition;Finally by dialogue response means 104, use In providing the corresponding answer of current text information in conjunction with both sides' affective characteristics value according to user's meaning of one's words characteristic information.
The of the invention system 10 answered based on affective characteristics adjustment robot, based on during human-computer dialogue ought be above This information and in the past text information, extract user feeling characteristic information and robot affective characteristics information, in conjunction with user's meaning of one's words Characteristic information, the answer for providing robot have the emotion of robot, and human-computer dialogue is more intelligent, further improves use Family experience.
Specifically, talk with identification module 102, be specifically used for:
According to current text information and past text information, the mood in current text information and in the past text information is obtained Status information, emotional state information include current emotional states information, user's past main emotional information and past secondary mood At least one of information;
According to current text information and past text information, the text in current text information and in the past text information is obtained Message, text message includes current text information and the semantics information in text information, key word information, proper noun are believed in the past At least one of breath and verb information;
According to current text information and past text information, the topic in current text information and in the past text information is obtained Information, topic information includes current staple of conversation information, currently secondary topic information, past staple of conversation information, past are secondary Topic information likes topic information and at present at least one of hot topic information;
According to current text information and past text information, the sentence in current text information and in the past text information is obtained Type, tone information and language performance information, language performance information include current dominant language behavioural information and secondary language performance Information;
According to emotional state information, text message, topic information, sentence pattern, tone information and language performance information, extract The characteristic information of user, characteristic information include user feeling characteristic information and user's meaning of one's words characteristic information.
Current text information combination contextual information can also be rewritten the information of the rewriting sentence obtained after original sentence Be added to extract characteristic information source information in, with emotional state information, text message, topic information, sentence pattern, tone information and Language performance information extracts the characteristic information of user together.
Wherein, the information for rewriting sentence is phenomenon in order to solve some sentence information deficiencies.
Such as:
Do A: you like watching movie?
B: like.
But B merely enters " liking ", information content is insufficient at this time." like seeing if just will become according to rewriting above Film ", computer can make more accurate judgement at this time.
Original sentence: retain the appearance of original sentence, computer capacity is allowed to access original sentence.Sentence is rewritten in cooperation, more It is fully understood by user's input.
Such as:
A: we will go to the cinema tomorrow
B: En En.
At this point, " En En " can be rewritten into " I will go to the cinema tomorrow ".If retaining former sentence " En En ", input computer becomes At " En En+I go to the cinema tomorrow ";Computer is it can be understood that at+response problem of watching movie tomorrow.In this manner into And obtain more complete information.
Sufficiently the current session text information of analysis identification user and past text conversation information, analyze the factor of processing It is more, it can make to extract obtained user feeling characteristic information and user's meaning of one's words characteristic information is more acurrate, and then provide robot Answer it is more acurrate, it is more intelligent.
Specifically, talk with response means 104, be specifically used for:
Data bank is obtained, multiple answers are stored in advance in data bank;
According to user's meaning of one's words characteristic information, selected in data bank in conjunction with both sides' affective characteristics value by machine learning algorithm One is selected to respond as the corresponding answer of current text information.
The answer that robot provides can select a correspondence from data bank by way of selecting and reasonably answer, and lead to The related algorithm for crossing machine learning, the answer for providing robot are more intelligent.
Specifically, talk with response means 104, be specifically used for:
According to user's meaning of one's words characteristic information, automatically generates and work as by the method for logic rules in conjunction with both sides' affective characteristics value The corresponding answer of preceding text information:
According to user semantic characteristic information, emotional characteristics in semantic characteristic information are analyzed;
A kind of logical combination is formed in conjunction with the mood of robot according to emotional characteristics in semantic feature information;
According to logical combination, current text is automatically generated from the answer strategy that the method by logic rules pre-establishes The corresponding answer of information.
Citing is illustrated:
User's input: " I and boyfriend say good-bye ", the meaning of one's words is characterized in: " saying good-bye, boyfriend ", user emotion feature meeting It is biased to " sad ", it is assumed that the mood of robot at this time is " neutrality ", then having higher chance to answer, " I can understand your present sense Feel, I can accompany you chat chat " etc. comfort type reply.If robot emotion is " cold and detached ", most probably respond: " En En refuels ~".
Logic rules among these refer to: various groups of " meaning of one's words feature " and " both sides' mood " that we have pre-established It closes, if the logical combination that " meaning of one's words feature " and " both sides' mood " in current session content is formed, which drops into, passes through logic rules Certain combination of the answer strategy that pre-establishes of method when, one response will be selected in certain answer scope.By patrolling The method for collecting rule carries out human-computer dialogue, to the answer that user is quick, intelligent, while considering the mood of both sides, makes to answer more intelligence Can, it is more humane.
Specifically, talk with response means 104, be specifically used for:
According to user's meaning of one's words characteristic information, automatically generates and work as by the method for machine learning in conjunction with both sides' affective characteristics value The corresponding answer of preceding text information:
According to user semantic characteristic information, emotional characteristics in semantic characteristic information are analyzed;
A kind of Response Policy is formed in conjunction with the mood of robot according to emotional characteristics in semantic feature information;
According to Response Policy, life is automatically generated or selected from the answer strategy that the method by machine learning pre-establishes At the corresponding answer of current text information.
Equally for example:
User's input: " I and boyfriend say good-bye ", the meaning of one's words is characterized in: " saying good-bye, boyfriend ", user emotion feature meeting It is biased to " sad ", it is assumed that the mood of robot at this time is " neutrality ", then having higher chance to answer, " I can understand your present sense Feel, I can accompany you chat chat " etc. comfort type reply.If robot emotion is " cold and detached ", most probably respond: " En En refuels ~".
Machine learning algorithm among these is to allow machine learning " meaning of one's words feature " and " both sides through prior data collection Collocation of the mood " with " automatically generating response " or " how selection is responded ".
It responds feedback according to the practical response situation of every user and experiment in training data to compare, if comparing result It is inconsistent, then update the parameter of machine learning, and repeat iteration, when in training data the response that generates with it is anti-with experiment It is fed back to when answering consistency up to highest, then stops iteration, and training is completed.In the stage of test, when user inputs one When this paper sentence, we can through passing data train come prediction module come responded (can automatically generate or Person, which is that selection is suitable, to be responded).
Answer strategy is pre-established by the method for machine learning, if " meaning of one's words feature " that includes in current text information The answer strategy that " both sides' mood " is formed is matched with some in the answer strategy pre-established, then can be automatically generated current The corresponding answer of text information.Answer strategy if the answer strategy formed does not rain in the answer strategy pre-established Match, then the corresponding answer of similar answer strategy generating current text information may be selected.By the method for machine learning, can be obtained A large amount of and accurate answer strategy provides the answer of accurate intelligence for user;The answer of robot, can be according to current " meaning of one's words, use Family emotion, robot emotion " responds.
Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention., rather than its limitations;To the greatest extent Pipe present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: its according to So be possible to modify the technical solutions described in the foregoing embodiments, or to some or all of the technical features into Row equivalent replacement;And these are modified or replaceed, various embodiments of the present invention technology that it does not separate the essence of the corresponding technical solution The range of scheme should all cover within the scope of the claims and the description of the invention.

Claims (8)

1. the method answered based on affective characteristics adjustment robot characterized by comprising
Step S1 obtains the current text information and past text information of user's input;
Step S2 extracts the characteristic information of the user, the spy according to the current text information and past text information Reference breath includes user feeling characteristic information and user's meaning of one's words characteristic information;
Step S3 obtains robot affective characteristics information, in conjunction with the user feeling characteristic information, carries out information integration, obtains Both sides' affective characteristics value, the robot affective characteristics information are that the information interacted by the past with the user is computed acquisition 's;
Step S4 provides the current text letter in conjunction with both sides' affective characteristics value according to user's meaning of one's words characteristic information Cease corresponding answer;
The step S2, specifically:
According to the current text information and past text information, obtain in the current text information and in the past text information Emotional state information, the emotional state information include current emotional states information, user's past main emotional information and past At least one of secondary emotional information;
According to the current text information and past text information, obtain in the current text information and in the past text information Text message, the text message includes the current text information and the semantics information in text information, keyword are believed in the past At least one of breath, proper noun information and verb information;
According to the current text information and past text information, obtain in the current text information and in the past text information Topic information, the topic information include current staple of conversation information, current secondary topic information, past staple of conversation information, Past secondary topic information likes topic information and at present at least one of hot topic information;
According to the current text information and past text information, obtain in the current text information and in the past text information Sentence pattern, tone information and language performance information, the language performance information include current dominant language behavioural information and secondary language Say behavioural information;
According to the emotional state information, text message, topic information, sentence pattern, tone information and language performance information, extract The characteristic information of the user, the characteristic information include user feeling characteristic information and user's meaning of one's words characteristic information.
2. the method according to claim 1 answered based on affective characteristics adjustment robot, which is characterized in that
In the step S4, specifically:
Data bank is obtained, multiple answers are stored in advance in the data bank;
According to user's meaning of one's words characteristic information, in conjunction with both sides' affective characteristics value, by machine learning algorithm, in the money One is selected to respond as the corresponding answer of the current text information in material library.
3. the method according to claim 1 answered based on affective characteristics adjustment robot, which is characterized in that
In the step S4, specifically:
It is automatic raw by the method for logic rules in conjunction with both sides' affective characteristics value according to user's meaning of one's words characteristic information At the corresponding answer of the current text information:
According to user's meaning of one's words characteristic information, emotional characteristics in the meaning of one's words characteristic information are analyzed;
A kind of logical combination is formed in conjunction with the mood of the robot according to emotional characteristics in the meaning of one's words characteristic information;
According to the logical combination, automatically generated from the answer strategy that the method by logic rules pre-establishes described current The corresponding answer of text information.
4. the method according to claim 1 answered based on affective characteristics adjustment robot, which is characterized in that
In the step S4, specifically:
It is automatic raw by the method for machine learning in conjunction with both sides' affective characteristics value according to user's meaning of one's words characteristic information At the corresponding answer of the current text information:
According to user's meaning of one's words characteristic information, emotional characteristics in the meaning of one's words characteristic information are analyzed;
A kind of Response Policy is formed in conjunction with the mood of the robot according to emotional characteristics in the meaning of one's words characteristic information;
According to the Response Policy, life is automatically generated or selected from the answer strategy that the method by machine learning pre-establishes At the corresponding answer of the current text information.
5. the system answered based on affective characteristics adjustment robot characterized by comprising
Text information obtains module, for obtaining the current text information and past text information of user's input;
Talk with identification module, for extracting the feature of the user according to the current text information and past text information Information, the characteristic information include user feeling characteristic information and user's meaning of one's words characteristic information;
Robot emotion module, in conjunction with the user feeling characteristic information, carries out letter for obtaining robot affective characteristics information Breath integration, obtains both sides' affective characteristics value, the robot affective characteristics information is the letter interacted by the past with the user Breath is computed acquisition;
Talk with response means, in conjunction with both sides' affective characteristics value, providing described according to user's meaning of one's words characteristic information The corresponding answer of current text information;
The dialogue identification module, is specifically used for:
According to the current text information and past text information, obtain in the current text information and in the past text information Emotional state information, the emotional state information include current emotional states information, user's past main emotional information and past At least one of secondary emotional information;
According to the current text information and past text information, obtain in the current text information and in the past text information Text message, the text message includes the current text information and the semantics information in text information, keyword are believed in the past At least one of breath, proper noun information and verb information;
According to the current text information and past text information, obtain in the current text information and in the past text information Topic information, the topic information include current staple of conversation information, current secondary topic information, past staple of conversation information, Past secondary topic information likes topic information and at present at least one of hot topic information;
According to the current text information and past text information, obtain in the current text information and in the past text information Sentence pattern, tone information and language performance information, the language performance information include current dominant language behavioural information and secondary language Say behavioural information;
According to the emotional state information, text message, topic information, sentence pattern, tone information and language performance information, extract The characteristic information of the user, the characteristic information include user feeling characteristic information and user's meaning of one's words characteristic information.
6. the system according to claim 5 answered based on affective characteristics adjustment robot, which is characterized in that
The dialogue response means, are specifically used for:
Data bank is obtained, multiple answers are stored in advance in the data bank;
According to user's meaning of one's words characteristic information, in conjunction with both sides' affective characteristics value, by machine learning algorithm, in the money One is selected to respond as the corresponding answer of the current text information in material library.
7. the system according to claim 5 answered based on affective characteristics adjustment robot, which is characterized in that
The dialogue response means, are specifically used for:
It is automatic raw by the method for logic rules in conjunction with both sides' affective characteristics value according to user's meaning of one's words characteristic information At the corresponding answer of the current text information:
According to user's meaning of one's words characteristic information, emotional characteristics in the meaning of one's words characteristic information are analyzed;
A kind of logical combination is formed in conjunction with the mood of the robot according to emotional characteristics in the meaning of one's words characteristic information;
According to the logical combination, automatically generated from the answer strategy that the method by logic rules pre-establishes described current The corresponding answer of text information.
8. the system according to claim 5 answered based on affective characteristics adjustment robot, which is characterized in that
The dialogue response means, are specifically used for:
It is automatic raw by the method for machine learning in conjunction with both sides' affective characteristics value according to user's meaning of one's words characteristic information At the corresponding answer of the current text information:
According to user's meaning of one's words characteristic information, emotional characteristics in the meaning of one's words characteristic information are analyzed;
A kind of Response Policy is formed in conjunction with the mood of the robot according to emotional characteristics in the meaning of one's words characteristic information;
According to the Response Policy, life is automatically generated or selected from the answer strategy that the method by machine learning pre-establishes At the corresponding answer of the current text information.
CN201611093206.9A 2016-12-01 2016-12-01 The method and system answered based on affective characteristics adjustment robot Active CN106599998B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201611093206.9A CN106599998B (en) 2016-12-01 2016-12-01 The method and system answered based on affective characteristics adjustment robot

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201611093206.9A CN106599998B (en) 2016-12-01 2016-12-01 The method and system answered based on affective characteristics adjustment robot

Publications (2)

Publication Number Publication Date
CN106599998A CN106599998A (en) 2017-04-26
CN106599998B true CN106599998B (en) 2019-02-01

Family

ID=58595972

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201611093206.9A Active CN106599998B (en) 2016-12-01 2016-12-01 The method and system answered based on affective characteristics adjustment robot

Country Status (1)

Country Link
CN (1) CN106599998B (en)

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107329986A (en) * 2017-06-01 2017-11-07 竹间智能科技(上海)有限公司 The interactive method and device recognized based on language performance
CN107301168A (en) * 2017-06-01 2017-10-27 深圳市朗空亿科科技有限公司 Intelligent robot and its mood exchange method, system
CN107423364B (en) * 2017-06-22 2024-01-26 百度在线网络技术(北京)有限公司 Method, device and storage medium for answering operation broadcasting based on artificial intelligence
CN109202922B (en) * 2017-07-03 2021-01-22 北京光年无限科技有限公司 Emotion-based man-machine interaction method and device for robot
WO2019098185A1 (en) * 2017-11-16 2019-05-23 株式会社Nttドコモ Dialog text generation system and dialog text generation program
CN108053826B (en) * 2017-12-04 2021-01-15 泰康保险集团股份有限公司 Method and device for man-machine interaction, electronic equipment and storage medium
US10522143B2 (en) * 2018-02-27 2019-12-31 Microsoft Technology Licensing, Llc Empathetic personal virtual digital assistant
CN109129467B (en) * 2018-07-27 2022-03-25 南京阿凡达机器人科技有限公司 Robot interaction method and system based on cognition
CN111199732B (en) * 2018-11-16 2022-11-15 深圳Tcl新技术有限公司 Emotion-based voice interaction method, storage medium and terminal equipment
CN111310460B (en) * 2018-12-12 2022-03-01 Tcl科技集团股份有限公司 Statement adjusting method and device
CN110675853B (en) * 2019-09-10 2022-07-05 苏宁云计算有限公司 Emotion voice synthesis method and device based on deep learning
CN112000781B (en) * 2020-07-20 2023-07-21 北京百度网讯科技有限公司 Information processing method and device in user dialogue, electronic equipment and storage medium
CN113076407B (en) * 2021-03-22 2023-07-21 联想(北京)有限公司 Information processing method and device
CN117808011B (en) * 2024-03-01 2024-06-04 青岛网信信息科技有限公司 Chat robot method, medium and system with simulated emotion

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103593054A (en) * 2013-11-25 2014-02-19 北京光年无限科技有限公司 Question-answering system combining emotion recognition and output
CN104809103A (en) * 2015-04-29 2015-07-29 北京京东尚科信息技术有限公司 Man-machine interactive semantic analysis method and system

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103593054A (en) * 2013-11-25 2014-02-19 北京光年无限科技有限公司 Question-answering system combining emotion recognition and output
CN104809103A (en) * 2015-04-29 2015-07-29 北京京东尚科信息技术有限公司 Man-machine interactive semantic analysis method and system

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
"基于中文文本情感分类的情感宣泄系统的研究与实现";罗广清;《中国优秀硕士学位论文全文数据库 信息科技辑》;20151215;第1-5章

Also Published As

Publication number Publication date
CN106599998A (en) 2017-04-26

Similar Documents

Publication Publication Date Title
CN106599998B (en) The method and system answered based on affective characteristics adjustment robot
Lubis et al. Eliciting positive emotion through affect-sensitive dialogue response generation: A neural network approach
Williams et al. A broad-coverage challenge corpus for sentence understanding through inference
Kamaruddin et al. Cultural dependency analysis for understanding speech emotion
KR101359718B1 (en) Conversation Managemnt System and Method Thereof
Sutoyo et al. Designing an emotionally realistic chatbot framework to enhance its believability with AIML and information states
CN106776926A (en) Improve the method and system of responsibility when robot talks with
CN105991847A (en) Call communication method and electronic device
Steffan Managing menopause at work: The contradictory nature of identity talk
Hung et al. Towards a method for evaluating naturalness in conversational dialog systems
CN106855879A (en) The robot that artificial intelligence psychology is seeked advice from music
CN106815321A (en) Chat method and device based on intelligent chat robots
CN107066568A (en) The interactive method and device predicted based on user view
CN106775665A (en) The acquisition methods and device of the emotional state change information based on sentiment indicator
CN108595609A (en) Generation method, system, medium and equipment are replied by robot based on personage IP
CN116561265A (en) Personalized dialogue generation method, model training method and device
Baktir Speech Act Theory: Austin, Searl Derrida’s Response and Deleuze’s Theory of Order-word
Needle et al. Gendered associations of English morphology
CN107025278A (en) Based on interactive user portrait extraction method and device
Endrass et al. Culture-related topic selection in small talk conversations across Germany and Japan
CN109948139A (en) A kind of semantic tendency analysis method and system
Miehle et al. Estimating user communication styles for spoken dialogue systems
D’Andrea et al. EMAG: an extended multimodal attribute grammar for behavioural features
Lugrin et al. Modeling and evaluating a bayesian network of culture-dependent behaviors
Altarawneh et al. Predicting Evoked Emotions in Conversations

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant