CN109670030A - Question and answer exchange method and device - Google Patents
Question and answer exchange method and device Download PDFInfo
- Publication number
- CN109670030A CN109670030A CN201811644570.9A CN201811644570A CN109670030A CN 109670030 A CN109670030 A CN 109670030A CN 201811644570 A CN201811644570 A CN 201811644570A CN 109670030 A CN109670030 A CN 109670030A
- Authority
- CN
- China
- Prior art keywords
- information
- user
- inputs
- feeling
- increment
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- User Interface Of Digital Computer (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the present application discloses a kind of question and answer exchange method and device, the every input primary information of user, information is inputted according to user and history mutual information judges whether the reduction of user feeling state meets preset condition, if the determination result is YES, illustrate that user emotion declines, then multiple feedback informations are provided and are selected for user, answer hit rate is improved, and then extend the dialogue service life, improve the probability that intelligent conversational system solves the problems, such as.
Description
Technical field
This application involves field of artificial intelligence, more specifically to a kind of question and answer exchange method and device.
Background technique
In intelligent customer service system, in order to solve the problem of that user proposes to propose to understand user according to user
It is intended to, is intended to provide solution according to user.However, inventor the study found that the prior art answer selection and statement
Relatively single fixation in mode, i.e. the output answer of medelling easily lead to user and generate disappointed emotion, so that the dialogue service life is short,
The problem of customer service system, fix-rate was lower.
Summary of the invention
The purpose of the application is to provide a kind of question and answer exchange method and device, with it is at least part of overcome deposit in the prior art
The technical issues of.
To achieve the above object, this application provides following technical solutions:
A kind of question and answer exchange method is applied in intelligent conversational system, and the intelligence conversational system can be to receiving
Input information is responded and provides feedback information, comprising:
It obtains user and inputs information, the user inputs information representation user and expects described in the intelligent conversational system feedback
Feedback information；
Information and history mutual information are inputted according to the user, judges that the user inputs user's feelings of information representation
Whether the reduction of sense state meets the first preset condition；
If judging result, which characterizes the user feeling state and reduces, meets the first preset condition, output is defeated for the user
Enter M in N number of feedback information of acquisition of information, described M, N are the positive integer greater than 1, and M is less than or equal to N.
The above method, it is preferred that further include:
If the judging result, which characterizes the user feeling state and promoted, meets the second preset condition, output is directed to the use
Family inputs 1 in N number of feedback information of acquisition of information.
The above method, it is preferred that the output inputs M in N number of feedback information of acquisition of information for the user,
Include:
The information for expressing apology is exported, and for the M in N number of feedback information of user input acquisition of information
It is a；
The information for expressing apology is located at before the M feedback information.
The above method, it is preferred that it is described that information and history mutual information are inputted according to the user, judge the user
Whether the user feeling state reduction of input information representation meets the first preset condition, comprising:
Information is inputted to the user to analyze, and obtains the first affective characteristics that the user inputs information, described the
The user feeling that the user inputs information representation is included at least in one affective characteristics；
According to each user in the corresponding first scoring increment of first affective characteristics and the history mutual information
The corresponding second scoring increment of the second affective characteristics for inputting information determines that the user inputs the corresponding third scoring of information and increases
Amount；Wherein, scoring increment is used to characterize the change direction and variation degree of user feeling；
If the third scoring increment list is levied user feeling and changed to negative sense emotion, and third scoring increment is absolute
Value is greater than preset value, determines that the user inputs user feeling state reduction the first preset condition of satisfaction of information representation.
The above method, it is preferred that it includes: the first information and the second information that the user, which inputs in information,；
The first information is characterized in front of obtaining user's input information, the feedback that the intelligence conversational system provides
Whether information is accurate；The second information representation user expects that the intelligent conversational system feeds back the feedback information；
In first affective characteristics further include: the first information.
The above method, it is preferred that information is inputted to the user and is analyzed, the user is obtained and inputs information representation
User feeling, comprising:
It inputs information to the user using preparatory trained Information Extraction Model to handle, to extract the user
The entity information in information is inputted, the entity information is the word of predefined characterization user feeling；
The user is inputted at information and the entity information using preparatory trained textual classification model
Reason obtains the user feeling that the user inputs information representation.
The above method, it is preferred that described according to the corresponding first scoring increment of first affective characteristics and described to go through
Each user in history interactive information inputs the corresponding second scoring increment of the second affective characteristics of information, determines user's input
The corresponding third scoring increment of information, comprising:
It obtains and is determined according to first user input information in the portrait information and/or the history mutual information of user
User feeling keynote；
By the corresponding first scoring increment of first affective characteristics, the user feeling keynote the corresponding 4th, which scores, to be increased
Each user inputs the second emotion of information in addition to first user inputs information in amount and the history mutual information
The corresponding second scoring increment summation of feature, obtains the third scoring increment；Alternatively,
By the corresponding first scoring increment of first affective characteristics, the user feeling keynote the corresponding 4th, which scores, to be increased
The corresponding second scoring increment of the second affective characteristics that each user inputs information in amount and the history mutual information is asked
With obtain third scoring increment.
The above method, it is preferred that the output inputs M in N number of feedback information of acquisition of information for the user,
Include:
If the first information is characterized in front of obtaining user's input information, the intelligence conversational system provides anti-
Feedforward information is accurate, determines target clause；
According to M in the target clause output N number of feedback information for inputting acquisition of information for the user.
A kind of question and answer interactive device is applied in intelligent conversational system, and the intelligence conversational system can be to receiving
Input information is responded and provides feedback information, comprising:
Module is obtained, inputs information for obtaining user, the user inputs information representation user and expects the intelligence meeting
Telephone system feeds back the feedback information；
Judgment module judges that the user inputs letter for inputting information and history mutual information according to the user
Whether the user feeling state reduction of breath characterization meets the first preset condition；
Output module meets the first preset condition if characterizing the user feeling state for judging result and reducing, output
M in N number of feedback information of acquisition of information are inputted for the user, described M, N are the positive integer greater than 1, and M is less than or waits
In N.
A kind of question and answer interactive device is applied in intelligent conversational system, and the intelligence conversational system can be to receiving
Input information is responded and provides feedback information, comprising:
Memory, at least storing one group of instruction set；
Processor is carried out for calling and executing the described instruction collection in the memory by executing described instruction collection
It operates below:
It obtains user and inputs information, the user inputs information representation user and expects described in the intelligent conversational system feedback
Feedback information；
Information and history mutual information are inputted according to the user, judges that the user inputs user's feelings of information representation
Whether the reduction of sense state meets the first preset condition；
If judging result, which characterizes the user feeling state and reduces, meets the first preset condition, output is defeated for the user
Enter M in N number of feedback information of acquisition of information, described M, N are the positive integer greater than 1, and M is less than or equal to N.
By above scheme it is found that a kind of question and answer exchange method provided by the present application and device, the primary letter of the every input of user
Breath inputs information according to user and history mutual information judges whether the reduction of user feeling state meets preset condition, if judgement
As a result be it is yes, illustrate user emotion decline, then provide multiple feedback informations for user select, improve answer hit rate, Jin Eryan
The long dialogue service life improves the probability that intelligent conversational system solves the problems, such as.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is a kind of implementation flow chart of question and answer exchange method provided by the embodiments of the present application；
Fig. 2 is provided by the embodiments of the present application according to user's input information and history mutual information, judges that user inputs
The user feeling state of information representation reduces a kind of implementation flow chart for whether meeting the first preset condition；
Fig. 3 analyzes user's input information to be provided by the embodiments of the present application, obtains user and inputs information representation
A kind of implementation flow chart of user feeling；
Fig. 4 is a kind of structural schematic diagram of question and answer interactive device provided by the embodiments of the present application；
Fig. 5 is another structural schematic diagram of question and answer interactive device provided by the embodiments of the present application.
Specification and claims and term " first " in above-mentioned attached drawing, " second ", " third " " the 4th " etc. (if
In the presence of) it is part for distinguishing similar, without being used to describe a particular order or precedence order.It should be understood that using in this way
Data be interchangeable under appropriate circumstances, so that embodiments herein described herein can be in addition to illustrating herein
Sequence in addition is implemented.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, those of ordinary skill in the art are obtained every other under that premise of not paying creative labor
Embodiment shall fall within the protection scope of the present invention.
Question and answer exchange method and device provided by the present application are applied in intelligent conversational system, which can
The input information received is responded and feedback information is provided.The intelligence conversational system can be the system of customer service after sale
System, alternatively, can be pre-sales customer service system or other systems etc. for providing service for user.
Referring to Fig. 1, Fig. 1 is a kind of implementation flow chart of question and answer exchange method provided by the embodiments of the present application, specifically may be used
To include:
Step S11: it obtains user and inputs information, which inputs information representation user and expect in intelligent conversational system feedback
State feedback information.
User can input information with text mode, information can also be inputted with voice mode, i.e., user's input information can
To be text, it is also possible to voice.
Step S12: inputting information and history mutual information according to user, judges that user inputs user's feelings of information representation
Whether the reduction of sense state meets the first preset condition.
After intelligent conversational system inputs information offer feedback information for user, which is possible to not reach
The expection of user, then user is more likely to produce negative sense emotion.For example, in customer service system, if intelligent conversational system
Enquirement for user gives incoherent answer, then user is likely to generate negative emotions.
In the embodiment of the present application, user feeling can be divided into positive emotion, neutral emotion or negative sense emotion.Positive emotion can
To be sub-divided into different degree, such as low positive (for example, satisfaction is lower), high positive (for example, very satisfied) etc.,
Negative sense emotion can also be sub-divided into different degree, for example, slight negative sense (for example, disappointed), serious negative sense (for example, indignation)
Deng.
In the embodiment of the present application, the reduction of user feeling state can refer to that user feeling develops to negative sense emotion direction.
Step S13: if judging result characterization user feeling state, which reduces, meets the first preset condition, output is defeated for user
Enter M in N number of feedback information of acquisition of information, M, N are the positive integer greater than 1, and M is less than or equal to N.
Intelligent conversational system can identify that user is intended to after obtaining user and inputting information, and be intended to filter out according to user
Scheme to be selected.In the embodiment of the present application, when user feeling state reduces and meets the first preset condition, filter out defeated with user
Enter information correlation maximum several (being denoted as N number of) feedback information to be selected, and according to user input information correlation by
High to Low sequence selects at least two (being denoted as M) feedback information outputs from N number of feedback information to be selected.
Question and answer exchange method provided by the embodiments of the present application, in question and answer interactive process, the every input primary information of user, root
Information is inputted according to user and history mutual information judges whether the reduction of user feeling state meets preset condition, if judging result is
It is to illustrate that user emotion declines, i.e., to negative sense affective development, then multiple feedback informations is provided and are selected for user, answer is improved
Hit rate, and then extend the dialogue service life, improve the probability that intelligent conversational system solves the problems, such as.
In an optional embodiment, when output is directed to M in N number of feedback information that user inputs acquisition of information,
The information for expressing apology can be exported, simultaneously with the M feedback information to mitigate the mood of user.This is used to express apology
The information of meaning is located at before above-mentioned M feedback information.
For example, sorry, I am wrong, you mean XXX either**?
In an optional embodiment, after obtaining user and inputting information, if inputting information and history friendship according to user
Mutual information judges that user inputs user feeling state promotion the second preset condition of satisfaction of information representation, can export for use
Family input acquisition of information N number of feedback information in 1,1 feedback information be with user input information it is maximally related that
Feedback information.
Wherein, user feeling state promotion refers to that user feeling develops to positive emotion direction.When user feeling state mentions
When rising the second preset condition of satisfaction, illustrate correctly to have recognized user's intention, inputs the feedback that information provides for user before
Information has reached user's expectation.
It is above-mentioned that information and history mutual information are inputted according to user in an optional embodiment, judge that user inputs
The user feeling state of information representation reduces a kind of implementation flow chart for whether meeting the first preset condition as shown in Fig. 2, can be with
Include:
Step S21: inputting information to user and analyze, and obtains the first affective characteristics that user inputs information, this first
The user feeling that user inputs information representation is included at least in affective characteristics.
Step S22: according to each user in the corresponding first scoring increment of the first affective characteristics and history mutual information
The corresponding second scoring increment of the second affective characteristics for inputting information determines that user inputs the corresponding third scoring increment of information；
Wherein, scoring increment is used to characterize the change direction and variation degree of user feeling.The value of scoring increment has positive and negative point,
Positive value characterization user feeling develops to positive emotion direction, and negative value characterization user feeling develops to negative sense emotion direction, and scoring increases
The degree that the absolute value characterization user feeling of amount develops to certain emotion direction, the absolute value for the increment that scores is bigger, characterizes emotion
The degree developed to certain direction is bigger.
When affective characteristics only include an information, for example, when only including user feeling, the corresponding scoring increment of affective characteristics
It is exactly the corresponding scoring increment of an information, if affective characteristics include multinomial information, the corresponding scoring of affective characteristics increases
Amount be exactly the multinomial information it is corresponding score increment and value, alternatively, weighted sum.The corresponding scoring increment of different information can be with
It is identical, it can also be different.
The change direction and variation journey for the user feeling that first scoring increment list requisition family information currently entered is characterized
Degree.The second emotion that some user in history mutual information inputs information (to be denoted as history input information convenient for narration) is special
Levy change direction and variation degree that corresponding second scoring increment characterizes the user feeling that history input information is characterized.The
Three scoring increments characterize the whole change direction and variation degree of user feeling in up to the present entire dialog procedure.
The corresponding relationship of affective characteristics and scoring increment can be stored in advance in intelligent conversational system.In general, negative sense feelings
Feeling corresponding scoring increment is negative value, and the corresponding scoring increment of positive emotion is positive value, and the corresponding scoring increment of neutral emotion is
0。
Step S23: if third scoring increment list levies user feeling to negative sense emotion direction change, and third scoring increment
Absolute value is greater than preset value, determines that user inputs user feeling state reduction the first preset condition of satisfaction of information representation.
For ease of calculation, each user can be inputted to the corresponding affective characteristics of information to be added in feature pool, emotion
Feature sorts according to the sequencing that user inputs information.When needing to calculate scoring increment using affective characteristics, directly from spy
It levies and extracts affective characteristics in pond.
Optionally, if third scoring increment list sign user feeling is to positive emotion direction change, although alternatively, third scores
Increment list levies user feeling to negative sense emotion direction change, but the absolute value of third scoring increment is less than or equal to preset value, then
It can determine that user inputs user feeling state promotion the second preset condition of satisfaction of information representation.
In an optional embodiment, it includes two parts that user, which inputs information, is denoted as the first information and the second information,
In,
The first information is characterized in front of obtaining user's input information, and whether the feedback information that intelligent conversational system provides is quasi-
Really；Second information representation user expects that intelligent conversational system feeds back feedback information；
Correspondingly, in the first affective characteristics further include: the above-mentioned first information.Based on this, when needing to calculate affective characteristics
When scoring increment, if the feedback information that the first information characterizes intelligent conversational system offer is accurate, the corresponding scoring of the first information
Increment is positive value, if the first information characterizes the feedback information inaccuracy that intelligent conversational system provides, the first information is corresponding to be commented
Dividing increment is negative value.
In the embodiment of the present application, intelligent conversational system inputs information for user each time, other than feedback information is provided,
The characterization whether correct option of feedback information is given, whether the feedback information for selecting intelligent conversational system to provide by user is quasi-
Really.Selection for answer next time provides foundation.
It is provided by the present application that user's input information is analyzed in an optional embodiment, it obtains user and inputs letter
A kind of implementation flow chart of the user feeling of characterization is ceased as shown in figure 3, may include:
Step S31: information is inputted to user using preparatory trained Information Extraction Model and is handled, to extract user
The entity information in information is inputted, which is the word of predefined characterization user feeling.
Entity information refers to the information with certain sense, for example, name, place name, mechanism name, proper noun, word of swearing at people
Deng.In the embodiment of the present application, entity information at least may include word of swearing at people.
The sample that Information Extraction Model can use entity information label is trained to obtain.Information Extraction Model can be with
Select CRF (Conditional Random Field, condition random field) model, LSTM (Long Short-Term Memory,
Shot and long term memory network)+CRF model etc..
Step S32: user is inputted at information and entity information using preparatory trained textual classification model
Reason obtains the user feeling that user inputs information representation.Textual classification model can select convolutional neural networks, alternatively, length
Phase memory network etc..
In the embodiment of the present application, textual classification model inputs information to user and carries out word segmentation processing, from word segmentation processing result
The keyword of extraction and entity information group are combined into the feature that user inputs information, are identified using this feature by middle extraction keyword
The user feeling of user's input information representation.
It should be noted that may then be input to textual classification model there is no entity information in the information of user's input
Entity information be sky.
User provided by the present application inputs the method that information carries out sentiment analysis, and sentiment analysis process has been incorporated with specific
The entity information of emotive meaning, sentiment analysis result are more nearly user's real feelings.
It is above-mentioned to be handed over according to the corresponding first scoring increment of the first affective characteristics and history in an optional embodiment
Each user in mutual information inputs the corresponding second scoring increment of the second affective characteristics of information, and it is corresponding to determine that user inputs information
Third score increment, comprising:
It obtains and the use that information determines is inputted according to first user in the portrait information and/or history mutual information of user
Family emotion keynote.
The portrait information of user refers to some behavior characteristic informations of user, dialogue duration, right for example, frequency of making a phone call
Words wheel number, problem fix rate, word of whether swearing at people in dialog procedure, the frequency etc. that word of swearing at people occurs.User's portrait can be logical
Crossing multiple session of the trained portrait model to user and intelligent conversational system in advance, (the multiple session is usually current current
Session several times before session, wherein intelligent conversational system start with user session to end-of-dialogue be a session) number
According to being analyzed to obtain.
In the application, the emotion keynote of user can be determined only by the portrait of user, can also only pass through history interaction
First input information in information determines, alternatively, in conjunction with user portrait and history mutual information in first input believe
Breath determines.
After establishing user's portrait, user portrait can be analyzed according to certain rule, determine that the portrait is corresponding
User feeling, which is exactly the emotion keynote of user.For example, there are multiple self-depreciatory expressions, mild and roundabout and courtesy expression meanings
Figure is considered as positive emotion keynote.The word that occurs swearing at people can be considered the emotion keynote of negative sense.Feelings are determined when needing to draw a portrait by user
When feeling keynote, the corresponding user feeling of user portrait is directly read.
The user feeling of first input information representation can also be used as user feeling keynote in history mutual information.
If in conjunction with user portrait and history mutual information in first input information determine user feeling keynote, can
The first emotion keynote is determined to be based respectively on user's portrait, determines second according to first input information in history mutual information
Emotion keynote determines third emotion keynote according to the first emotion keynote and the second emotion keynote, using the third emotion keynote as
User feeling keynote.
The third emotion keynote can be the characterization higher feelings of negative sense degree in the first emotion keynote and the second emotion keynote
Feel keynote；For example, the first emotion keynote is positive emotion, the second emotion keynote is negative sense emotion, then third emotion keynote can be with
It is negative sense emotion.For another example it is serious negative sense emotion that the first emotion keynote, which is slight negative sense emotion, the second emotion keynote, then
Three emotion keynotes can be serious negative sense emotion.
Alternatively,
By the first emotion keynote to the emotion direction that the second emotion keynote is characterized emotion keynote adjusted.For example, the
One emotion keynote is positive emotion, and the second emotion keynote is negative sense emotion, then third emotion keynote can be neutral emotion.Example again
Such as, the first emotion keynote is high positive emotion, and the second emotion keynote is negative sense emotion, then third emotion keynote can be minuent
Positive emotion.
After obtaining user feeling keynote, by the corresponding first scoring increment of the first affective characteristics, user feeling keynote pair
Each user inputs the of information in addition to first user inputs information in the 4th scoring increment and history mutual information answered
The corresponding second scoring increment summation of two affective characteristics, obtains third scoring increment.Since user feeling keynote is based on
First user inputs what information determined in history mutual information, therefore, when calculating third scoring increment, can not consider further that
First user inputs the emotion of information in history mutual information.Certainly, it when calculating third scoring increment, can continue to examine
Consider the emotion that first user in history mutual information inputs information.
Alternatively,
After obtaining user feeling keynote, by the corresponding first scoring increment of the first affective characteristics, user feeling keynote pair
The second affective characteristics corresponding second that each user inputs information in the 4th scoring increment and history mutual information answered are commented
Divide increment summation, obtains third scoring increment.
Optionally, if user feeling keynote is positive emotion keynote, the value of the corresponding scoring increment of user feeling keynote
Can scoring increment more corresponding than information in affective characteristics it is big, for example, if the corresponding scoring increment of positive user feeling is+1,
Then the corresponding scoring increment of positive user feeling keynote can be+5.And the corresponding scoring increment of negative sense user feeling keynote is then small
In the corresponding scoring increment of positive user feeling keynote, for example, the corresponding scoring increment of negative sense user feeling is 0.
In an optional embodiment, during calculating third scoring increment, if there is the case where continuous deduction, example
The corresponding scoring increment of affective characteristics as user twice in succession inputs information is all negative value, then additional deduction, i.e., according to the
Each user in the corresponding first scoring increment of one affective characteristics and history mutual information inputs the second affective characteristics of information
Corresponding second scoring increment subtracts third scoring increment after determining that user inputs the corresponding third scoring increment of information again
Go it is preset chase after deduction numerical value, obtain final third scoring increment.
If there is the case where continuous bonus point, can optionally not have to additional score because in dialog procedure (especially with
Inhuman dialogue), user's subconsciousness, which can tend to other side, can not understand and (hold the suspicious attitude), so more holding compared to positive emotion
It is also easy to produce negative sense emotion, multiple correct answers can may just retrieve a false answer of beginning.If from beginning of conversation feelings
Sense state is (continuous several times, which provide, is satisfied with answer) in rising trend always, can add bonus point in this case.
In an optional embodiment, before user inputs in information and is characterized in acquisition user's input information, intelligence
When the accurate first information of feedback information that conversational system provides, target clause is determined.That the target clause can be is milder and roundabout,
Mild clause.
When output is directed to M in N number of feedback information that user inputs acquisition of information, needle is exported according to the target clause
M in N number of feedback information of acquisition of information are inputted to user.
In the embodiment of the present application, when intelligent conversational system provides correct option, still user feeling but shows becoming for negative sense
When gesture, it is believed that be that intelligent conversational system provides the mode of feedback information and goes wrong.
For example, user inputs information: do you sell chargers that support mini USB ports? intelligence
Energy conversational system detects that user is intended to buy Cellphone Accessories, directly releases link answer: Where to find
Motorola products.And provide option Yes or NO.Here, intelligent conversational system does not ask the understanding that user is intended to
Topic, but it is excessively rude on the presentation mode of answer, if user selects YES, but found in the analysis of next user feeling
User feeling shows negative sense trend, then in next dialogue, clause optimization is carried out when releasing answer, for example, using
" Do you......? " clause.To improve user emotion.
In an optional embodiment, information processing method provided by the present application can also include:
Before including to be characterized in acquisition user's input information in continuous K user input information, intelligent conversational system
The first information of the feedback information mistake of offer, or the intelligent conversational system feedback information provided and the information of user's input without
Guan Shi, operator attendance of transferring.K is the positive integer greater than 1, for example, K can be 2 or 3 or other values.
Under specific circumstances, for example, in the initial stage of session, user is after inputting information for the first time, if intelligent session system
It unites unidentified to user's intention, the information of feedback may be for " Sorry, I don ' t understand ", in this case can
User is set to generate sense of defeat rapidly, if next round dialogue cannot still provide the user with satisfied answer, in order to mitigate user emotion
And solve the problems, such as, can actively transfer operator attendance, select operator attendance without waiting user.
The value of K can take different values according to the difference of session process.For example, if in the incipient stage of session, K
Value can be smaller, such as can be 2, if a period of time has been carried out in session, or has been carried out more wheel dialogues, then K
Value can be larger, for example, can be 3.
In another optional embodiment, if dialogue alreadys exceed a fixed wheel number, but there are no customer problem is solved, then answer
This actively helps user to turn operator attendance, because the purpose of user and intelligent conversational system dialogue is very strong, not chats, so
Even if current affective state value with higher (i.e. scoring increment), can only illustrate that user quality itself is higher, but right to this
It has been the affective state of negative sense for words.
Some differences of the application and the prior art are exemplified below.
In the prior art, the session context of user and traditional intelligence conversational system (hereinafter referred to as traditional customer service) may be as follows
It is shown:
User's input: Excuse me, would you mind helping me to repair my phone?
Traditional customer service: Please check our website to find the latest models.
User's input: What are you saying? I need to get my phone repaired.
Traditional customer service: Please check our website to find the latest models.
User's input: Don ' t waste my time, I want to talk to a live agent.
Obviously, the answer that customer service is fed back twice is all wrong, and regardless of what information user inputs, all with same side
Formula (identical clause, the tone or content etc.) provides answer.
And it is based on scheme provided by the present application, the intelligent conversational system (hereinafter referred to as emotion customer service) of user and the application
Session context may be as follows:
User's input: Excuse me, would you mind helping me to repair my phone?
Here, sentiment analysis is carried out to the information of user's input, determines that user feeling for positive emotion, can provide higher
User feeling keynote.Based on this, emotion customer service provides following feedback information:
Emotion customer service: I ' m sorry to hear that, Angela.Please kindly check our
website to find the latest models.
Based on the embodiment of the present application, customer service is replied in a manner of same courtesy.The relevant some information of user are provided simultaneously
(these information are pre-stored in intelligent conversational system), such as name, academic title etc..It is provided in this example name " Angela ", increases and use
Family experience.But intelligent conversational system is intended to misinterpretation to user, is proposed wrong answer.
User's input: Don ' t you understand what I meant? I need to get my phone
repaired.
Intelligent conversational system analyzes this user input, and obtaining user feeling is negative sense emotion, can use a mark
Label are marked, for example, negative label.That is, being once intended to misinterpretation to user due to preceding, user occurs
Negative emotions.
Intelligent conversational system calculates the decline of user feeling state according to emotion keynote and this user input and meets first
Preset condition.It when releasing answer, apologizes first, then abandons an answer, and attempt to release multiple related candidates
Answer increases hit rate, keeps client.Based on this, emotion customer service provides following answer:
Emotion customer service: Sorry, it ' s my fault, do you mean repair your phone or check
You repair status?
User's input: repair my phone.
Intelligent conversational system analyzes this user input, obtains user feeling as neutral emotion, intelligent session system
System provides normal label, and correct due to being intended to understand to user, user feeling is restored.
Intelligent conversational system calculates user feeling state and promoted according to emotion keynote and user's input twice meets second
Condition, it may be possible to hit answer, restore the normal tone and there is information to release an answer.Based on this, emotion customer service is given
Following answer out:
Emotion customer service: Ok, please visit this site www.XXXX.com to file a repair
form.
Problem is solved, conversation end.
Based on the embodiment of the present application, the process that user conversates with emotion customer service may also be as follows:
User's input: Excuse me, would you mind helping me to repair my phone?
Emotion customer service: I ' m sorry to hear that, Angela.Please kindly check our
Website to find the latest models.Yes or no?
User's input: No, Don ' t you understand what I meant? I need to get my phone
repaired.
Emotion customer service: Sorry, it ' s my fault, do you mean repair your phone or check
You repair status? Yes or no?
User's input: Yes, repair my phone.
Emotion customer service: Ok, please visit this site www.XXXX.com to file a repair
form.
Corresponding with embodiment of the method, the application also provides a kind of question and answer interactive device, provided by the embodiments of the present application to ask
A kind of structural schematic diagram of interactive device is answered as shown in figure 4, may include:
Obtain module 41, judgment module 42 and output module 43；Wherein,
It obtains module 41 and inputs information for obtaining user, which inputs information representation user and expect intelligent conversational system
Feed back the feedback information.
Judgment module 42 is used to input information and history mutual information according to user, judges that user inputs information representation
Whether the reduction of user feeling state meets the first preset condition.
If output module 43 is reduced for judging result characterization user feeling state meets the first preset condition, output is directed to
User inputs M in N number of feedback information of acquisition of information, and M, N are the positive integer greater than 1, and M is less than or equal to N.
Question and answer interactive device provided by the present application, in question and answer interactive process, the every input primary information of user, according to user
Input information and history mutual information judge whether the reduction of user feeling state meets preset condition, if the determination result is YES, say
Bright user emotion decline then provides multiple feedback informations and selects for user that is, to negative sense affective development, improves answer hit
Rate, and then extend the dialogue service life, improve the probability that intelligent conversational system solves the problems, such as.
In an optional embodiment, output module 43 be can be also used for:
If judging result, which characterizes user feeling state and promoted, meets the second preset condition, output inputs information for user and obtains
1 in N number of feedback information taken.
In an optional embodiment, output module 43 inputs N number of feedback information of acquisition of information in output for user
In M when, be specifically used for:
The information for expressing apology is exported, and for M in N number of feedback information of user's input acquisition of information；
Information for expressing apology is located at before M feedback information.
In an optional embodiment, judgment module 42 specifically can be used for:
Information is inputted to user to analyze, and the first affective characteristics that user inputs information is obtained, in the first affective characteristics
The user feeling of information representation is inputted including at least user；
Information is inputted according to each user in the corresponding first scoring increment of the first affective characteristics and history mutual information
The corresponding second scoring increment of the second affective characteristics, determine that user inputs information corresponding third scoring increment；Wherein, it scores
Increment is used to characterize the change direction and variation degree of user feeling；
If third scores, increment list is levied user feeling and is changed to negative sense emotion, and the absolute value of third scoring increment is greater than pre-
If value, determine that user inputs user feeling state reduction the first preset condition of satisfaction of information representation.
In an optional embodiment, it includes: the first information and the second information that user, which inputs in information,；
The first information is characterized in front of obtaining user's input information, and whether the feedback information that intelligent conversational system provides is quasi-
Really；Second information representation user expects that intelligent conversational system feeds back the feedback information；
In first affective characteristics further include: the first information.
In an optional embodiment, judgment module 42 inputs information to user and analyzes, when obtaining user feeling, tool
Body can be used for:
Information is inputted to user using preparatory trained Information Extraction Model to handle, and inputs information to extract user
In entity information, entity information be it is predefined characterization user feeling word；
Information is inputted to user using preparatory trained textual classification model and entity information is handled, is obtained
The user feeling of user's input information representation.
In an optional embodiment, judgment module 42 specifically can be used for when determining third scoring increment:
It obtains and the use that information determines is inputted according to first user in the portrait information and/or history mutual information of user
Family emotion keynote；
By the first affective characteristics it is corresponding first scoring increment, user feeling keynote it is corresponding 4th scoring increment, and
Each user inputs the second affective characteristics corresponding second of information in addition to first user inputs information in history mutual information
The increment that scores is summed, and third scoring increment is obtained；Alternatively,
By the first affective characteristics it is corresponding first scoring increment, user feeling keynote it is corresponding 4th scoring increment, and
Each user inputs the corresponding second scoring increment summation of the second affective characteristics of information in history mutual information, obtains third and comments
Divide increment.
In an optional embodiment, output module 43 inputs N number of feedback information of acquisition of information in output for user
In M when, specifically can be used for:
If the first information is characterized in front of obtaining user's input information, the feedback information that intelligent conversational system provides is accurate,
Determine target clause；
M in N number of feedback information of acquisition of information are inputted for the user according to the output of target clause.
Another structural schematic diagram of question and answer interactive device provided by the embodiments of the present application is as shown in figure 5, may include:
Memory 51 and processor 52；Wherein,
Memory 51 is at least storing one group of instruction set；
Processor 52 is used to calling and executing the instruction set in memory 51, is performed the following operation by executing instruction collection:
It obtains user and inputs information, user inputs information representation user and expects that the intelligent conversational system feeds back the feedback
Information；
Information and history mutual information are inputted according to user, judges that user inputs the user feeling state drop of information representation
It is low whether to meet the first preset condition；
If judging result, which characterizes user feeling state and reduces, meets the first preset condition, output inputs information for user and obtains
The M in N number of feedback information taken, described M, N are the positive integer greater than 1, and M is less than or equal to N.
Question and answer interactive device provided by the present application, in question and answer interactive process, the every input primary information of user, according to user
Input information and history mutual information judge whether the reduction of user feeling state meets preset condition, if the determination result is YES, say
Bright user emotion decline then provides multiple feedback informations and selects for user that is, to negative sense affective development, improves answer hit
Rate, and then extend the dialogue service life, improve the probability that intelligent conversational system solves the problems, such as.
In an optional embodiment, processor 52 be can be also used for:
If judging result, which characterizes user feeling state and promoted, meets the second preset condition, output inputs information for user and obtains
1 in N number of feedback information taken.
In an optional embodiment, processor 52 is in N number of feedback information that output inputs acquisition of information for user
M when, be specifically used for:
The information for expressing apology is exported, and for M in N number of feedback information of user's input acquisition of information；
Information for expressing apology is located at before M feedback information.
In an optional embodiment, processor 52 is reduced in the user feeling state for judging user's input information representation is
When the first preset condition of no satisfaction, specifically it can be used for:
Information is inputted to user to analyze, and the first affective characteristics that user inputs information is obtained, in the first affective characteristics
The user feeling of information representation is inputted including at least user；
Information is inputted according to each user in the corresponding first scoring increment of the first affective characteristics and history mutual information
The corresponding second scoring increment of the second affective characteristics, determine that user inputs information corresponding third scoring increment；Wherein, it scores
Increment is used to characterize the change direction and variation degree of user feeling；
If third scores, increment list is levied user feeling and is changed to negative sense emotion, and the absolute value of third scoring increment is greater than pre-
If value, determine that user inputs user feeling state reduction the first preset condition of satisfaction of information representation.
In an optional embodiment, it includes: the first information and the second information that user, which inputs in information,；
The first information is characterized in front of obtaining user's input information, and whether the feedback information that intelligent conversational system provides is quasi-
Really；Second information representation user expects that intelligent conversational system feeds back the feedback information；
In first affective characteristics further include: the first information.
In an optional embodiment, processor 52 inputs information to user and analyzes, when obtaining user feeling, specifically
It can be used for:
Information is inputted to user using preparatory trained Information Extraction Model to handle, and inputs information to extract user
In entity information, entity information be it is predefined characterization user feeling word；
Information is inputted to user using preparatory trained textual classification model and entity information is handled, is obtained
The user feeling of user's input information representation.
In an optional embodiment, processor 52 specifically can be used for when determining third scoring increment:
It obtains and the use that information determines is inputted according to first user in the portrait information and/or history mutual information of user
Family emotion keynote；
By the first affective characteristics it is corresponding first scoring increment, user feeling keynote it is corresponding 4th scoring increment, and
Each user inputs the second affective characteristics corresponding second of information in addition to first user inputs information in history mutual information
The increment that scores is summed, and third scoring increment is obtained；Alternatively,
By the first affective characteristics it is corresponding first scoring increment, user feeling keynote it is corresponding 4th scoring increment, and
Each user inputs the corresponding second scoring increment summation of the second affective characteristics of information in history mutual information, obtains third and comments
Divide increment.
In an optional embodiment, processor 52 is in N number of feedback information that output inputs acquisition of information for user
M when, specifically can be used for:
If the first information is characterized in front of obtaining user's input information, the feedback information that intelligent conversational system provides is accurate,
Determine target clause；
M in N number of feedback information of acquisition of information are inputted for the user according to the output of target clause.
Those of ordinary skill in the art may be aware that list described in conjunction with the examples disclosed in the embodiments of the present disclosure
Member and algorithm steps can be realized with the combination of electronic hardware or computer software and electronic hardware.These functions are actually
It is implemented in hardware or software, the specific application and design constraint depending on technical solution.Professional technician
Each specific application can be used different methods to achieve the described function, but this realization is it is not considered that exceed
The scope of the present invention.
In several embodiments provided herein, it should be understood that disclosed systems, devices and methods, it can be with
It realizes by another way.Another point, shown or discussed mutual coupling, direct-coupling or communication connection can
To be the indirect coupling or communication connection of device or unit through some interfaces, it can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.
It should be appreciated that can be combined with each other combination in the embodiment of the present application from power, each embodiment, feature, can realize
Solve aforementioned technical problem.
It, can be with if the function is realized in the form of SFU software functional unit and when sold or used as an independent product
It is stored in a computer readable storage medium.Based on this understanding, technical solution of the present invention is substantially in other words
The part of the part that contributes to existing technology or the technical solution can be embodied in the form of software products, the meter
Calculation machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be a
People's computer, server or network equipment etc.) it performs all or part of the steps of the method described in the various embodiments of the present invention.
And storage medium above-mentioned includes: that USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), arbitrary access are deposited
The various media that can store program code such as reservoir (RAM, Random Access Memory), magnetic or disk.
The foregoing description of the disclosed embodiments enables those skilled in the art to implement or use the present invention.
Various modifications to these embodiments will be readily apparent to those skilled in the art, as defined herein
General Principle can be realized in other embodiments without departing from the spirit or scope of the present invention.Therefore, of the invention
It is not intended to be limited to the embodiments shown herein, and is to fit to and the principles and novel features disclosed herein phase one
The widest scope of cause.
Claims (10)
1. a kind of question and answer exchange method is applied in intelligent conversational system, the intelligence conversational system can be defeated to what is received
Enter information to be responded and feedback information is provided characterized by comprising
It obtains user and inputs information, the user inputs information representation user and expects that the intelligent conversational system feeds back the feedback
Information；
Information and history mutual information are inputted according to the user, judges that the user inputs the user feeling shape of information representation
Whether state reduction meets the first preset condition；
If judging result, which characterizes the user feeling state and reduces, meets the first preset condition, output inputs letter for the user
M in the N number of feedback information obtained are ceased, described M, N are the positive integer greater than 1, and M is less than or equal to N.
2. the method according to claim 1, wherein further include:
If the judging result, which characterizes the user feeling state and promoted, meets the second preset condition, output is defeated for the user
Enter 1 in N number of feedback information of acquisition of information.
3. the method according to claim 1, wherein the output inputs the N of acquisition of information for the user
M in a feedback information, comprising:
The information for expressing apology is exported, and for M in N number of feedback information of user input acquisition of information；
The information for expressing apology is located at before the M feedback information.
4. the method according to claim 1, wherein described input information and history interaction according to the user
Information, judges whether the user feeling state reduction of user's input information representation meets the first preset condition, comprising:
Information is inputted to the user to analyze, and obtains the first affective characteristics that the user inputs information, first feelings
Feel and includes at least the user feeling that the user inputs information representation in feature；
According to each user input in the corresponding first scoring increment of first affective characteristics and the history mutual information
The corresponding second scoring increment of second affective characteristics of information determines that the user inputs the corresponding third scoring increment of information；
Wherein, scoring increment is used to characterize the change direction and variation degree of user feeling；
If the third scoring increment list is levied user feeling and changed to negative sense emotion, and the absolute value of third scoring increment is big
In preset value, determine that the user inputs user feeling state reduction the first preset condition of satisfaction of information representation.
5. according to the method described in claim 4, it is characterized in that, it includes: the first information and the in information that the user, which inputs,
Two information；
The first information is characterized in front of obtaining user's input information, the feedback information that the intelligence conversational system provides
It is whether accurate；The second information representation user expects that the intelligent conversational system feeds back the feedback information；
In first affective characteristics further include: the first information.
6. method according to claim 4 or 5, which is characterized in that input information to the user and analyze, obtain institute
State the user feeling that user inputs information representation, comprising:
It inputs information to the user using preparatory trained Information Extraction Model to handle, to extract user's input
Entity information in information, the entity information are the word of predefined characterization user feeling；
Information is inputted to the user using preparatory trained textual classification model and the entity information is handled,
Obtain the user feeling that the user inputs information representation.
7. method according to claim 4 or 5, which is characterized in that described according to first affective characteristics corresponding
Each user in one scoring increment and the history mutual information inputs corresponding second scoring of the second affective characteristics of information
Increment determines that the user inputs the corresponding third scoring increment of information, comprising:
It obtains and the use that information determines is inputted according to first user in the portrait information and/or the history mutual information of user
Family emotion keynote；
By the corresponding first scoring increment of first affective characteristics, the corresponding 4th scoring increment of the user feeling keynote,
And each user inputs the second emotion spy of information in addition to first user inputs information in the history mutual information
Corresponding second scoring increment summation is levied, the third scoring increment is obtained；Alternatively,
By the corresponding first scoring increment of first affective characteristics, the corresponding 4th scoring increment of the user feeling keynote,
And each user inputs the corresponding second scoring increment summation of the second affective characteristics of information in the history mutual information, obtains
To third scoring increment.
8. according to the method described in claim 5, it is characterized in that, the output inputs the N of acquisition of information for the user
M in a feedback information, comprising:
If the first information is characterized in front of obtaining user's input information, the feedback letter that the intelligence conversational system provides
Breath is accurate, determines target clause；
According to M in the target clause output N number of feedback information for inputting acquisition of information for the user.
9. a kind of question and answer interactive device is applied in intelligent conversational system, the intelligence conversational system can be defeated to what is received
Enter information to be responded and feedback information is provided characterized by comprising
Module is obtained, inputs information for obtaining user, the user inputs information representation user and expects the intelligent session system
System feeds back the feedback information；
Judgment module judges that the user inputs information table for inputting information and history mutual information according to the user
Whether the user feeling state reduction of sign meets the first preset condition；
Output module meets the first preset condition if characterizing the user feeling state for judging result and reducing, and output is directed to
The user inputs M in N number of feedback information of acquisition of information, and described M, N are the positive integer greater than 1, and M is less than or equal to N.
10. a kind of question and answer interactive device is applied in intelligent conversational system, the intelligence conversational system can be defeated to what is received
Enter information to be responded and feedback information is provided characterized by comprising
Memory, at least storing one group of instruction set；
Processor, it is following by executing the progress of described instruction collection for calling and executing the described instruction collection in the memory
Operation:
It obtains user and inputs information, the user inputs information representation user and expects that the intelligent conversational system feeds back the feedback
Information；
Information and history mutual information are inputted according to the user, judges that the user inputs the user feeling shape of information representation
Whether state reduction meets the first preset condition；
If judging result, which characterizes the user feeling state and reduces, meets the first preset condition, output inputs letter for the user
M in the N number of feedback information obtained are ceased, described M, N are the positive integer greater than 1, and M is less than or equal to N.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811644570.9A CN109670030B (en) | 2018-12-30 | 2018-12-30 | Question-answer interaction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811644570.9A CN109670030B (en) | 2018-12-30 | 2018-12-30 | Question-answer interaction method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109670030A true CN109670030A (en) | 2019-04-23 |
| CN109670030B CN109670030B (en) | 2022-06-28 |
Family
ID=66146978
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811644570.9A Active CN109670030B (en) | 2018-12-30 | 2018-12-30 | Question-answer interaction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109670030B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111985248A (en) * | 2020-06-30 | 2020-11-24 | 联想(北京)有限公司 | Information interaction method and device |
| WO2021225550A1 (en) * | 2020-05-06 | 2021-11-11 | Iren Yaser Deniz | Emotion recognition as feedback for reinforcement learning and as an indicator of the explanation need of users |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103678361A (en) * | 2012-09-13 | 2014-03-26 | 腾讯科技（深圳）有限公司 | Method and device for switching search results |
| CN104156359A (en) * | 2013-05-13 | 2014-11-19 | 腾讯科技（深圳）有限公司 | Linking information recommendation method and device |
| CN105183848A (en) * | 2015-09-07 | 2015-12-23 | 百度在线网络技术（北京）有限公司 | Human-computer chatting method and device based on artificial intelligence |
| JP2016071394A (en) * | 2014-09-26 | 2016-05-09 | 日本電信電話株式会社 | Emotional information providing device, emotional information providing method, and emotional information providing program |
| CN106126636A (en) * | 2016-06-23 | 2016-11-16 | 北京光年无限科技有限公司 | A kind of man-machine interaction method towards intelligent robot and device |
| CN108153169A (en) * | 2017-12-07 | 2018-06-12 | 北京康力优蓝机器人科技有限公司 | Guide to visitors mode switching method, system and guide to visitors robot |
| WO2018147193A1 (en) * | 2017-02-08 | 2018-08-16 | 日本電信電話株式会社 | Model learning device, estimation device, method therefor, and program |
| CN108491519A (en) * | 2018-03-26 | 2018-09-04 | 上海智臻智能网络科技股份有限公司 | Man-machine interaction method and device, storage medium, terminal |
| CN108553905A (en) * | 2018-03-30 | 2018-09-21 | 努比亚技术有限公司 | Data feedback method, terminal and computer storage media based on game application |
| CN109036405A (en) * | 2018-07-27 | 2018-12-18 | 百度在线网络技术（北京）有限公司 | Voice interactive method, device, equipment and storage medium |
-
2018
- 2018-12-30 CN CN201811644570.9A patent/CN109670030B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103678361A (en) * | 2012-09-13 | 2014-03-26 | 腾讯科技（深圳）有限公司 | Method and device for switching search results |
| CN104156359A (en) * | 2013-05-13 | 2014-11-19 | 腾讯科技（深圳）有限公司 | Linking information recommendation method and device |
| JP2016071394A (en) * | 2014-09-26 | 2016-05-09 | 日本電信電話株式会社 | Emotional information providing device, emotional information providing method, and emotional information providing program |
| CN105183848A (en) * | 2015-09-07 | 2015-12-23 | 百度在线网络技术（北京）有限公司 | Human-computer chatting method and device based on artificial intelligence |
| CN106126636A (en) * | 2016-06-23 | 2016-11-16 | 北京光年无限科技有限公司 | A kind of man-machine interaction method towards intelligent robot and device |
| WO2018147193A1 (en) * | 2017-02-08 | 2018-08-16 | 日本電信電話株式会社 | Model learning device, estimation device, method therefor, and program |
| CN108153169A (en) * | 2017-12-07 | 2018-06-12 | 北京康力优蓝机器人科技有限公司 | Guide to visitors mode switching method, system and guide to visitors robot |
| CN108491519A (en) * | 2018-03-26 | 2018-09-04 | 上海智臻智能网络科技股份有限公司 | Man-machine interaction method and device, storage medium, terminal |
| CN108553905A (en) * | 2018-03-30 | 2018-09-21 | 努比亚技术有限公司 | Data feedback method, terminal and computer storage media based on game application |
| CN109036405A (en) * | 2018-07-27 | 2018-12-18 | 百度在线网络技术（北京）有限公司 | Voice interactive method, device, equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 姚丹 等: "社会化问答平台中问题属性对答案域的影响", 《图书情报知识》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021225550A1 (en) * | 2020-05-06 | 2021-11-11 | Iren Yaser Deniz | Emotion recognition as feedback for reinforcement learning and as an indicator of the explanation need of users |
| CN111985248A (en) * | 2020-06-30 | 2020-11-24 | 联想(北京)有限公司 | Information interaction method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109670030B (en) | 2022-06-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110555095B (en) | Man-machine conversation method and device | |
| CN112365894B (en) | AI-based composite voice interaction method and device and computer equipment | |
| CN112346567B (en) | Virtual interaction model generation method and device based on AI (Artificial Intelligence) and computer equipment | |
| CN109981910B (en) | Service recommendation method and device | |
| CN110059182A (en) | Art recommended method and device towards customer service | |
| CN112988991B (en) | Method and system for performing anti-fraud intervention through man-machine conversation | |
| CN110610705B (en) | Voice interaction prompter based on artificial intelligence | |
| CN110046230B (en) | Method for generating recommended speaking collection, and recommended speaking method and device | |
| CN110209897A (en) | Intelligent dialogue method, apparatus, storage medium and equipment | |
| CN113360622B (en) | User dialogue information processing method and device and computer equipment | |
| CN110347863A (en) | Talk about art recommended method and device and storage medium | |
| CN111078856B (en) | Group chat conversation processing method and device and electronic equipment | |
| CN116303949B (en) | Dialogue processing method, dialogue processing system, storage medium and terminal | |
| CN110569344B (en) | Method and device for determining standard question corresponding to dialogue text | |
| CN109684459A (en) | A kind of information processing method and device | |
| CN111429157A (en) | Method, device and equipment for evaluating and processing complaint work order and storage medium | |
| CN112015875B (en) | Construction method of online customer service assistant | |
| CN109670030A (en) | Question and answer exchange method and device | |
| CN109949830A (en) | User's intension recognizing method and equipment | |
| CN110489519B (en) | Session method based on session prediction model and related products | |
| CN113486166B (en) | Construction method, device and equipment of intelligent customer service robot and storage medium | |
| CN111144906A (en) | Data processing method and device and electronic equipment | |
| CN115409585B (en) | Information recommendation method based on voice interaction system and related device | |
| CN112182189A (en) | Conversation processing method and device, electronic equipment and storage medium | |
| EP3908941A1 (en) | Artificial intelligence system for business processes |
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 |