CN110069606A - Man-machine conversation's method, apparatus, electronic equipment and storage medium - Google Patents
Man-machine conversation's method, apparatus, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN110069606A CN110069606A CN201711013017.0A CN201711013017A CN110069606A CN 110069606 A CN110069606 A CN 110069606A CN 201711013017 A CN201711013017 A CN 201711013017A CN 110069606 A CN110069606 A CN 110069606A
- Authority
- CN
- China
- Prior art keywords
- intention
- user
- man
- slot value
- machine conversation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3343—Query execution using phonetics
Abstract
The embodiment provides a kind of man-machine conversation's method, apparatus, electronic equipment and storage mediums, are related to mode identification technology.This method comprises: receiving the solicited message of user；The intention of the user is identified according to the solicited message；The intention that will identify that and the history of the user stored in database intention merge；Corresponding user response is generated according to fused intention.The technical solution of the embodiment of the present invention can be intended to determine the accurate intention that user is current in conjunction with the history of user, and then can lead to too small amount of session and can be obtained by accurate user response, improve the session experience of user.
Description
Technical field
The embodiment of the present invention is related to mode identification technology, in particular to a kind of man-machine conversation's method, man-machine
Conversational device, electronic equipment and computer readable storage medium.
Background technique
Man-machine conversation is a sub- direction of artificial intelligence field, and popular saying is exactly to allow people can be by the language of the mankind
(i.e. natural language) is interacted with computer such as man-machine conversation's system.It, can by the interaction of person to person's chance telephone system
It makes one chance telephone system and understands the intention and demand of people, to complete song search, shopping places an order, the tasks such as control of equipment.
Person to person's chance telephone system carries out communication, similar with spoken dialog interpersonal in daily life.By
In the diversity and complexity of chat script, often leading to man-machine conversation's system can not obtain only by the statement of user
It takes accurate user to be intended to, user is needed to change the intention of oneself in due course according to obtained feedback in subsequent session.So
Current man-machine conversation's system generally can complete the identification being integrally intended to user to take turns conversational mode more.When user statement not
When clear or unspecific, man-machine conversation's system can obtain more more specific information by clarifying and inquiring to user,
Effectively to be responded to user's request.
However, technical limitation of the prior art due to itself, in order to ensure in the solicited message of acquisition user
Accurately it is intended to, it has to carry out tediously long many and diverse more wheel conversation procedures, sacrifice the convenience of user experience.
It should be noted that information is only used for reinforcing the reason to background of the present invention disclosed in above-mentioned background technology part
Solution, therefore may include the information not constituted to the prior art known to persons of ordinary skill in the art.
Summary of the invention
The purpose of the present invention is to provide a kind of man-machine conversation's method, man-machine conversation's device, electronic equipment and computers
Readable storage medium storing program for executing, so overcome caused by the limitation and defect due to the relevant technologies at least to a certain extent one or
Multiple problems.
According to a first aspect of the embodiments of the present invention, a kind of man-machine conversation's method is provided, comprising: receive the request of user
Information；The intention of the user is identified according to the solicited message；The use stored in the intention and database that will identify that
The history intention at family is merged；Corresponding user response is generated according to fused intention.
In some embodiments of the invention, aforementioned schemes, the intention that will identify that and the institute stored in database are based on
The history intention for stating user merges, comprising: extracts the slot value of the intention identified；By the slot value extracted and the history
The slot value of intention merges, and is merged with the intention that will identify that with history intention.
In some embodiments of the invention, aforementioned schemes, man-machine conversation's method are based on further include: identify extracting
After the slot value of intention out, the intention and the slot value are stored in the database.
In some embodiments of the invention, aforementioned schemes, man-machine conversation's method further include: judge the number are based on
It whether is more than predetermined threshold according to the quantity that the history of the user stored in library is intended to；It is intended in the history for determining the user
Quantity when being more than the predetermined threshold, delete the history of the user stored at first and be intended to and corresponding slot value.
In some embodiments of the invention, aforementioned schemes, man-machine conversation's method further include: in the data are based on
When the deposit duration that any history of the user stored in library is intended to reaches scheduled duration, deletes any history and be intended to
And corresponding slot value.
In some embodiments of the invention, aforementioned schemes are based on, corresponding user is being generated according to fused intention
Before response, further includes: judge whether the slot value information of fused intention is complete；In the slot for determining the fused intention
When value information is complete, corresponding user response is generated according to the slot value information of the fused intention.
In some embodiments of the invention, aforementioned schemes, man-machine conversation's method further include: described in judgement are based on
When the slot value information of fused intention is imperfect, Xiang Suoshu user feedback prompt information.
In some embodiments of the invention, aforementioned schemes are based on, corresponding user is generated according to fused intention and is rung
It answers, comprising: retrieved according to the slot value information of fused intention, corresponding user response is generated based on search result；Or
Person generates corresponding user response by production model according to the slot value information of fused intention.
In some embodiments of the invention, aforementioned schemes, man-machine conversation's method further include: do not generating are based on
When the user response of effect, one or more slot values in the slot value information of fused intention are deleted, updated slot value is obtained
Information；Corresponding user response is generated according to the updated slot value information.
In some embodiments of the invention, aforementioned schemes are based on, one in the slot value information of fused intention is deleted
A or multiple slot values, comprising: be intended to the sequencing of storage according to the history of the user, delete the slot value of fused intention
The one or more slot values stored at first in information.
In some embodiments of the invention, aforementioned schemes are based on, extract the slot value of the intention identified, comprising: be based on
The mode of template matching and/or neural network extracts the slot value of the intention identified.
According to a second aspect of the embodiments of the present invention, a kind of man-machine conversation's device is provided, comprising: receiving unit is used for
Receive the solicited message of user；Intention assessment unit, for identifying the intention of the user according to the solicited message；Fusion is single
The history intention of member, the user stored in the intention and database for will identify that merges；Generation unit is responded,
For generating corresponding user response according to fused intention.
According to a third aspect of the embodiments of the present invention, a kind of electronic equipment is provided, comprising: processor；And memory,
It is stored with computer-readable instruction on the memory, is realized when the computer-readable instruction is executed by the processor as above
State man-machine conversation's method described in first aspect.
According to a fourth aspect of the embodiments of the present invention, a kind of computer readable storage medium is provided, meter is stored thereon with
Calculation machine program realizes man-machine conversation's method as described in above-mentioned first aspect when the computer program is executed by processor.
In the technical solution provided by some embodiments of the present invention, the intention that will be identified from user request information
It is merged with the history of the user stored in database intention, corresponding user response is generated according to fused intention.One
Aspect merges the history intention of the user stored from the intention identified in user request information and database, energy
Accurate intention enough current in conjunction with the historical requests acquisition of information user of user；On the other hand, it is generated according to fused intention
Corresponding user response can substantially reduce the wheel number of session, and not needing more wheel sessions can obtain more effectively accurately
User response improves the session experience of user.
It should be understood that above general description and following detailed description be only it is exemplary and explanatory, not
It can the limitation present invention.
Detailed description of the invention
The drawings herein are incorporated into the specification and forms part of this specification, and shows and meets implementation of the invention
Example, and be used to explain the principle of the present invention together with specification.It should be evident that the accompanying drawings in the following description is only the present invention
Some embodiments for those of ordinary skill in the art without creative efforts, can also basis
These attached drawings obtain other attached drawings.In the accompanying drawings:
Fig. 1 shows a kind of schematic block diagram of man-machine conversation's system in technical solution；
Fig. 2 shows the flow diagrams of man-machine conversation's method of first exemplary embodiment according to the present invention；
The history that Fig. 3 shows the user of the storage of an exemplary embodiment of the present invention is intended to and the signal of slot value
Figure；
Fig. 4 shows the application scenarios schematic diagram of man-machine conversation's method of an exemplary embodiment according to the present invention；
Fig. 5 shows the application scenarios signal of man-machine conversation's method of another exemplary embodiment according to the present invention
Figure；
Fig. 6 shows the flow diagram of man-machine conversation's method of second exemplary embodiment according to the present invention；
Fig. 7 shows the schematic block diagram of man-machine conversation's device of an exemplary embodiment according to the present invention；
Fig. 8 shows the structural schematic diagram for being suitable for the computer system for the electronic equipment for being used to realize the embodiment of the present invention.
Specific embodiment
Example embodiment is described more fully with reference to the drawings.However, example embodiment can be real in a variety of forms
It applies, and is not understood as limited to embodiment set forth herein；On the contrary, thesing embodiments are provided so that the present invention will be comprehensively and complete
It is whole, and the design of example embodiment is comprehensively communicated to those skilled in the art.Identical appended drawing reference indicates in figure
Same or similar part, thus repetition thereof will be omitted.
In addition, described feature, structure or characteristic can be incorporated in one or more implementations in any suitable manner
In example.In the following description, many details are provided to provide and fully understand to the embodiment of the present invention.However,
It will be appreciated by persons skilled in the art that technical solution of the present invention can be practiced without one or more in specific detail,
Or it can be using other methods, constituent element, device, step etc..In other cases, it is not shown in detail or describes known side
Method, device, realization or operation are to avoid fuzzy each aspect of the present invention.
Block diagram shown in the drawings is only functional entity, not necessarily must be corresponding with physically separate entity.
I.e., it is possible to realize these functional entitys using software form, or realized in one or more hardware modules or integrated circuit
These functional entitys, or these functional entitys are realized in heterogeneous networks and/or processor device and/or microcontroller device.
Flow chart shown in the drawings is merely illustrative, it is not necessary to including all content and operation/step,
It is not required to execute by described sequence.For example, some operation/steps can also decompose, and some operation/steps can close
And or part merge, therefore the sequence actually executed is possible to change according to the actual situation.
Fig. 1 shows a kind of schematic block diagram of man-machine conversation's system in technical solution.It is described referring to Fig.1, the people's chance
Telephone system may include: voice recognition unit 110, semantic understanding unit 120, text-to-speech unit 130 and dialogue management
Unit 140.The voice request information of user is converted to text request information by voice recognition unit 110；Semantic understanding unit 120
The semantic information for extracting the user in text information request is for example intended to (i.e. Intent) and slot value (i.e. slot) information, and according to
Extracted semantic information generates corresponding user response；The use that text-to-speech unit 130 generates semantic understanding unit 120
Family response contents are converted into voice content and are exported, dialogue management unit 140 and voice recognition unit 110, semantic understanding list
Member 120 and text-to-speech unit 130 connect, and are managed to above three unit.
In one of man-machine conversation's system shown in Fig. 1 technical solution, user's input is received by voice recognition unit 110
Solicited message, semantic understanding unit 120 is constantly present an initial shape for the semantic understanding for the solicited message that user inputs
State, the original state can specifically be expressed as the 0th in some root node or some user's process flow of a semantic tree
Grade processing function, and then the acquisition for the attribute value (slot value) that certain user is intended to by means of subsequent more wheel sessions completions.Slot
The acquisition of value can specifically be expressed as being filled the leaf node of semantic tree, or turn to next stage user and handle letter
Number, can just execute corresponding task after obtaining complete slot value to complete the intention of user.In this technical solution,
When only can completely does not obtain the slot value that user is intended to, just start to save the historical requests information of user, and transfers
More wheel sessions are carried out, but if current statement meets complete intent information and attribute information, semantic understanding unit 120 will
Operation corresponding to current statement is executed, intention to user in this case and slot value can't be intended to caches.
But people can not narrowly think only when upper sentence language semantic is sufficiently complete when carrying out communication,
Lower sentence language is only the continuity and supplement of sentence language.People always think during constantly talking by change or correction oneself
Method, the limitation of this technical solution to have to isolate this context pass being originally present in communication of coming
System, then guarantee the accuracy understood as far as possible language fragments to the greatest extent.But this way is difficult to meet information extraction
During completeness and accuracy, user is forced to need to repeat the meaning of oneself again and again, to allow human-computer interaction system
System has apparent understanding to the intention of oneself.For example, " how is the weather of Beijing tomorrow for the in short inquiry of user the
Sample? ", next sentence inquires again " how is the weather in Shanghai? ".The user that semantic understanding unit 120 obtains in this technical solution is defeated
Enter be " how is the weather in Shanghai? ", corresponding to be intended that inquiry weather, attribute is Shanghai, in this way for inquiry weather this
User just has complete attribute information for being intended to, and man-machine conversation's system will return to same day Shanghai Weather information to user
?.Only from this in short from the point of view of, the processing of man-machine conversation's system is not wrong, but in this way processing be apparently not it is optimal, because
For upper word user inquiry be tomorrow weather conditions, according to the statistical analysis to existing spoken phenomenon, current the words is answered
It is then more proper to the inquiry of the situation of Shanghai weather tomorrow that this is interpreted as user.
Since people are in the communication process of natural dialogue, front and back sentence has certain correlation, example in maximum probability
It can be the further explanation or further supplement to the sentence of front such as subsequent sentence.So previous user is anticipated
Figure gets up with slot value information cache, when user inputs new solicited message, by the intention of current request information and slot value information
It is combined with the intention of caching with slot value information, will be better understood when the intention of user.
Based on above content, a kind of man-machine conversation's method first proposed in example embodiments of the present invention.Referring to Fig. 2 institute
Show, which may comprise steps of:
Step S210 receives the solicited message of user；
Step S220 identifies the intention of the user according to the solicited message；
The history intention of the user stored in step S230, the intention that will identify that and database merges；
Step S240 generates corresponding user response according to fused intention.
According to man-machine conversation's method in this example embodiment, on the one hand, the meaning that will be identified from user request information
Scheme to merge with the history of the user stored in database intention, can work as in conjunction with the historical requests acquisition of information user of user
Preceding accurate intention；On the other hand, corresponding user response is generated according to fused intention, the wheel of session can be substantially reduced
Number, more effective accurate user response can be obtained by not needing more wheel sessions, improve the session experience of user.
In the following, by man-machine conversation's method in this example embodiment is described in detail in conjunction with attached drawing.
Referring to shown in Fig. 2, in step S210, the solicited message of user is received.
In this exemplary embodiment, the solicited message of user can be one section of language and characters or text, for example, user
Have input a Duan Yuyin " how is Beijing weather tomorrow? ", then can will be used by language identification unit 110 shown in Fig. 1
The voice request information at family is converted to text request information.
Next, in step S220, the intention of the user is identified according to the solicited message referring to shown in Fig. 2.
In this exemplary embodiment, after receiving the solicited message of user, semantic understanding shown in Fig. 1 can be passed through
Unit 120 extracts the semantic information in the solicited message of user, and the semantic information in the solicited message of user may include user
It is intended to and slot value information.For example, user inputs " book 1 May 10 and remove Pekinese's ticket ", and user's is intended that when ordering air ticket
" ticket booking ", semantic information relevant to ticket booking task include as " 1 " of " ticket booking quantity ", as the " May of " departure time "
No. 10 ", and as " Beijing " of " destination ".Ticket booking quantity, the departure time and destination are known as slot, provided in sentence this
The occurrence of a little slots is known as slot value.
Therefore, in this exemplary embodiment, the meaning of user can be identified according to the semantic information in the solicited message of user
Figure, then can extract the slot value of the intention identified.For example, according to user input " how is Beijing weather tomorrow? "
The available user's of solicited message is intended to weather lookup, extracts first slot value such as place slot of weather lookup intention
Value is " Beijing ", and second slot value such as time slot value is " tomorrow ".Pass through the intention of the user identified and extracts intention
Slot value can remove the duplicate information of redundancy in the spoken voice of user's input, only retain the necessary letter useful to subsequent processing
Breath.
In this exemplary embodiment, the intention identified can be extracted based on the mode of template matching and/or neural network
Slot value, for example, Seq2Seq (Sequence to Sequence, sequence to sequence) mode and CRF can be used
(conditional random field, condition random field) mode carries out intention assessment and slot value and extracts, using Seq2Seq and
CRF mode can reach higher accuracy in practical applications.It should be noted that the intention in example embodiments of the present invention
The method that identification and slot value extract can also use other methods appropriate, and the present invention is to this without particular determination.
Next, referring to shown in Fig. 2, in step S230, the use that is stored in the intention and database that will identify that
The history intention at family is merged.
It in this exemplary embodiment, can be after the slot value for extracting the intention identified, in the database described in storage
It is intended to and the slot value.For example, the intention for extracting the historical requests of user can be stored in the data buffer area of database
With slot value.For example, the historical requests record of several users can be saved using Redis database, Redis database
Key assignments can choose the id (identifier, identifier) of user, and the value of storage can be using intention plus the representation of slot value.
Specifically, can store the intention and slot value of historical user's request in the following manner referring to shown in Fig. 3: " [be intended to: slot value；
Slot value] [it is intended to: slot value；Slot value] [it is intended to: slot value；Slot value] [...] [...] ... ", for example, " Beijing weather tomorrow is why for user's input
Sample? ", can store after being extracted through intention assessment and slot value as [weather lookup: Beijing；Tomorrow].
Further, in this exemplary embodiment, several history can be saved and asked according to the demand of different business product
Seek information.For example, in order to remove intention and slot with the user of the lesser history of solicited message relevance currently entered request
Value can delete the intention and slot of user's request of a part of history stored in database according to storage time or storage quantity
Value.For example, it can be determined that whether the quantity that the history of the user stored in database is intended to is more than predetermined threshold；Determining
When the quantity that the history of user is intended to is more than predetermined threshold, the history for deleting the user stored at first is intended to and corresponding slot value.
Further, it is also possible to delete institute when the deposit duration that any history of the user stored in the database is intended to reaches scheduled duration
State any history intention and corresponding slot value.
It should be noted that in this exemplary embodiment, storage user is intended to and the database of slot value can be Redis number
Database according to library, but in example embodiments of the present invention is without being limited thereto, such as can also be MySQL database, Oracle
Database or SQL server database etc..
It further, can be by the slot value of the slot value extracted and history intention in this example embodiment example
It merges, is merged with the intention that will identify that with history intention.Active user is intended to and is cached in the queue
History intention merged after, current intention and the history of user are intended to unified consideration, will according to specific service logic
It enough integrates to the slot value under an intention and merges, more can accurately understand that user is intended in this way.In this way by being intended to
Fused result is likely to be obtained a slot value list more more abundant than active user request content.
Next, in step S240, generating corresponding user with continued reference to shown in Fig. 2 according to fused intention and ringing
It answers.
In this exemplary embodiment, it can be retrieved according to the slot value information of fused intention, be based on search result
Generate corresponding user response；Or the slot value information according to fused intention, corresponding use is generated by production model
Family response.Fig. 4 shows the schematic diagram of man-machine conversation's method using an exemplary embodiment of the present invention.Referring to institute in Fig. 4
Show, first solicited message of user's input is " playing song red bean ", carries out retrieval generation according to first solicited message
First user response is " red bean of Wang Fei is played for you ".Next, second solicited message of user's input is " broadcasting side
The song that Datong District is sung " is only analyzed from current statement and is only capable of obtaining that user wants to listen the song of Khalil Fong.It is asked having merged first
Ask " play song red bean " in information into, can be intended to obtain later according to fused user second user response " for
You play the red bean of Khalil Fong ".Therefore, man-machine conversation's method in this example embodiment can combine user's history solicited message
Judge that user is actually intended to listen the red bean of Khalil Fong, the user response of generation is " red bean of Khalil Fong is played for you ", from
And make man-machine conversation's system more accurate for the understanding of user and efficient.
Fig. 5 shows the schematic diagram of man-machine conversation's method using another example embodiment of the invention.Referring in Fig. 5
Shown, first solicited message of user's input is " play Liu Dehua sing song ", first user response of generation be " for
You play the date of Liu Dehua passed by together ".Second solicited message of user's input is " play love song ", the second of generation
A user response is " love song antiphonal singing song Liu Dehua is that love is thin primary ".Conversely, in the prior art, if inputted in user
Under the scene of second solicited message, if it is desired to listening the love song of Liu Dehua, then the request must completely be described in a word, such as
" playing the love song that Liu De China sings ".Though same purpose can be reached in this way, the dialogue habit of people is not met, is caused man-machine
Interactive redundancy and complexity.Compared with prior art, the man-machine interaction method in example embodiments of the present invention can utilize use
The historical requests information at family more fully understands human language by means of machine memory, to improve the easy-to-use of man-machine interactive system
Property.
Fig. 6 shows the flow diagram of man-machine conversation's method of second exemplary embodiment according to the present invention.Ginseng
According to shown in Fig. 6, in step S610, intention assessment is carried out to the solicited message of user's input and slot value extracts；In step S620
In, by the history intention and slot value information of the user stored in database and the intention and slot value information extracted in step S610
It is merged；The man-machine conversation side in treatment process and above-mentioned first exemplary embodiment in step S610 and step S620
The treatment process of method is similar, will not be described in great detail again.
Next, judging whether fused slot value information is complete in step S630.It is obtaining melting by historical information
The user closed is intended to that intent information completeness judgement can be carried out with after slot value information, checks whether the slot value letter of missing
Breath.Meet when slot value information in step S630 after decision fusion is complete and execute the necessary condition that user is intended to, in step
Corresponding user response is generated according to the slot value information of fused intention in S640.Slot in step S630 after decision fusion
It when value information is imperfect, initiates to inquire to user in step S650, triggers more wheel sessions of next round, obtained under with expectation
Remaining slot value information is supplied in wheel session.
Further, in step S640, because what is obtained has been complete intention slot value information, then it can use institute
It states and is intended to the effective user response of slot value information progress information retrieval generation, or used and generated using the intention slot value information
Formula model directly generates corresponding user response.In step S650, since active user requests to be intended to slot value and incomplete,
The operation that user's intention cannot be executed cannot generate effective user response, need to user feedback prompt information.Therefore, exist
In step S650, the slot value information that next round session is lacked to user feedback is initiated, it is available in this way in next round session
Remaining slot value information, to obtain executing necessary all slot values of the intention.
Further, in this exemplary embodiment, it is contemplated that the fused slot value of process is than the slot value that active user requests
Slot value information is increased, this semantic combination information is effective in linguistics, but can under more complicated constraint condition
It can not generate actually active user response.Therefore, it when not generating effective user response, can delete fused
One or more slot values in the slot value information of intention obtain updated slot value information；Believed according to the updated slot value
Breath generates corresponding user response.For example, failing effectively to be tied according to when carrying out information retrieval by fused slot value
When fruit, that is, suitable user response, the slot value information that successively removal is added to active user's request can choose, such as can be by
According to the precedence of historical requests information, the one or more slot value information stored at first are got rid of, the slot value letter after simplifying
Corresponding user response is generated by retrieval on the basis of breath, is intended to so as to match user to the maximum extent.
In addition, in an embodiment of the present invention, additionally providing a kind of man-machine conversation's device.Referring to shown in Fig. 7, the people's chance
Words device 700 may include: receiving unit 710, intention assessment unit 720, integrated unit 730 and response generation unit 730.
Wherein, receiving unit 710 is used to receive the solicited message of user；Intention assessment unit 720 is used to be known according to the solicited message
The intention of the not described user；The history of the user stored in the intention and database that integrated unit 730 is used to will identify that
Intention is merged；Generation unit 740 is responded to be used to generate corresponding user response according to fused intention.
In some embodiments of the invention, aforementioned schemes are based on, integrated unit 730 includes: extraction unit, for extracting
The slot value of the intention identified；Combining unit, the slot value for the slot value extracted to be intended to the history merge, with
The intention that will identify that is merged with history intention.
In some embodiments of the invention, aforementioned schemes, man-machine conversation's device 700 are based on further include: storage is single
Member, for after the extraction unit extracts the slot value of intention identified, store in the database the intention and
The slot value.
In some embodiments of the invention, aforementioned schemes, man-machine conversation's device 700 further include: first sentences are based on
Disconnected unit, whether the quantity for judging that the history of the user stored in the database is intended to is more than predetermined threshold；Institute
It states storage unit to be also used to, determines that the quantity that the history of the user is intended to is more than the predetermined threshold in first judging unit
When value, the history for deleting the user stored at first is intended to and corresponding slot value.
In some embodiments of the invention, aforementioned schemes are based on, the storage unit is also used to: in the database
The deposit duration that any history of the user of storage is intended to is when reaching scheduled duration, deletes any history and is intended to and right
The slot value answered.
In some embodiments of the invention, aforementioned schemes, man-machine conversation's device 700 further include: second sentences are based on
Disconnected unit, for before the response generation unit 740 generates corresponding user response according to fused intention, judgement to be melted
Whether the slot value information of the intention after conjunction is complete；The response generation unit 740 is configured that be determined in the second judgment unit
When the slot value information of the fused intention is complete, corresponding user is generated according to the slot value information of the fused intention
Response.
In some embodiments of the invention, aforementioned schemes, man-machine conversation's device 700 further include: feedback are based on
Member, when the slot value information for determining the fused intention in the second judgment unit is imperfect, Xiang Suoshu user is anti-
Present prompt information.
In some embodiments of the invention, aforementioned schemes are based on, response generation unit 740 is configured to: according to fusion
The slot value information of intention afterwards is retrieved, and generates corresponding user response based on search result；Or according to fused meaning
The slot value information of figure generates corresponding user response by production model.
In some embodiments of the invention, aforementioned schemes are based on, the integrated unit 730 is also used to: in the response
When generation unit 740 does not generate effective user response, one or more slots in the slot value information of fused intention are deleted
Value, obtains updated slot value information；The response generation unit 740 is also used to be generated according to the updated slot value information
Corresponding user response.
In some embodiments of the invention, aforementioned schemes are based on, the integrated unit 730 is configured to: according to described
The history of user is intended to the sequencing of storage, deletes the one or more stored at first in the slot value information of fused intention
Slot value.
In some embodiments of the invention, aforementioned schemes are based on, the extraction unit is configured to: based on template matching
And/or the mode of neural network extracts the slot value of the intention identified.
Each functional module and above-mentioned man-machine conversation side due to man-machine conversation's device 700 of example embodiments of the present invention
The step of example embodiment of method, is corresponding, therefore details are not described herein.
In an exemplary embodiment of the present invention, a kind of electronic equipment that can be realized the above method is additionally provided.
Below with reference to Fig. 8, it illustrates the computer systems 800 for the electronic equipment for being suitable for being used to realize the embodiment of the present invention
Structural schematic diagram.The computer system 800 of electronic equipment shown in Fig. 8 is only an example, should not be to the embodiment of the present invention
Function and use scope bring any restrictions.
As shown in figure 8, computer system 800 includes central processing unit (CPU) 801, it can be read-only according to being stored in
Program in memory (ROM) 802 or be loaded into the program in random access storage device (RAM) 803 from storage section 808 and
Execute various movements appropriate and processing.In RAM 803, it is also stored with various programs and data needed for system operatio.CPU
801, ROM 802 and RAM 803 is connected with each other by bus 804.Input/output (I/O) interface 805 is also connected to bus
804。
I/O interface 805 is connected to lower component: the importation 806 including keyboard, mouse etc.；It is penetrated including such as cathode
The output par, c 807 of spool (CRT), liquid crystal display (LCD) etc. and loudspeaker etc.；Storage section 808 including hard disk etc.；
And the communications portion 809 of the network interface card including LAN card, modem etc..Communications portion 809 via such as because
The network of spy's net executes communication process.Driver 810 is also connected to I/O interface 805 as needed.Detachable media 811, such as
Disk, CD, magneto-optic disk, semiconductor memory etc. are mounted on as needed on driver 810, in order to read from thereon
Computer program be mounted into storage section 808 as needed.
Particularly, according to an embodiment of the invention, may be implemented as computer above with reference to the process of flow chart description
Software program.For example, the embodiment of the present invention includes a kind of computer program product comprising be carried on computer-readable medium
On computer program, which includes the program code for method shown in execution flow chart.In such reality
It applies in example, which can be downloaded and installed from network by communications portion 809, and/or from detachable media
811 are mounted.When the computer program is executed by central processing unit (CPU) 801, executes and limited in the system of the application
Above-mentioned function.
It should be noted that computer-readable medium shown in the present invention can be computer-readable signal media or meter
Calculation machine readable storage medium storing program for executing either the two any combination.Computer readable storage medium for example can be --- but not
Be limited to --- electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor system, device or device, or any above combination.Meter
The more specific example of calculation machine readable storage medium storing program for executing can include but is not limited to: have the electrical connection, just of one or more conducting wires
Taking formula computer disk, hard disk, random access storage device (RAM), read-only memory (ROM), erasable type may be programmed read-only storage
Device (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), light storage device, magnetic memory device,
Or above-mentioned any appropriate combination.In the present invention, computer readable storage medium can be it is any include or storage journey
The tangible medium of sequence, the program can be commanded execution system, device or device use or in connection.And at this
In invention, computer-readable signal media may include in a base band or as carrier wave a part propagate data-signal,
Wherein carry computer-readable program code.The data-signal of this propagation can take various forms, including but unlimited
In electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be that computer can
Any computer-readable medium other than storage medium is read, which can send, propagates or transmit and be used for
By the use of instruction execution system, device or device or program in connection.Include on computer-readable medium
Program code can transmit with any suitable medium, including but not limited to: wireless, electric wire, optical cable, RF etc. are above-mentioned
Any appropriate combination.
Flow chart and block diagram in attached drawing are illustrated according to the system of various embodiments of the invention, method and computer journey
The architecture, function and operation in the cards of sequence product.In this regard, each box in flowchart or block diagram can generation
A part of one module, program segment or code of table, a part of above-mentioned module, program segment or code include one or more
Executable instruction for implementing the specified logical function.It should also be noted that in some implementations as replacements, institute in box
The function of mark can also occur in a different order than that indicated in the drawings.For example, two boxes succeedingly indicated are practical
On can be basically executed in parallel, they can also be executed in the opposite order sometimes, and this depends on the function involved.Also it wants
It is noted that the combination of each box in block diagram or flow chart and the box in block diagram or flow chart, can use and execute rule
The dedicated hardware based systems of fixed functions or operations is realized, or can use the group of specialized hardware and computer instruction
It closes to realize.
Being described in unit involved in the embodiment of the present invention can be realized by way of software, can also be by hard
The mode of part realizes that described unit also can be set in the processor.Wherein, the title of these units is in certain situation
Under do not constitute restriction to the unit itself.
As on the other hand, present invention also provides a kind of computer-readable medium, which be can be
Included in electronic equipment described in above-described embodiment；It is also possible to individualism, and without in the supplying electronic equipment.
Above-mentioned computer-readable medium carries one or more program, when the electronics is set by one for said one or multiple programs
When standby execution, so that the electronic equipment realizes such as above-mentioned man-machine conversation's method as described in the examples.
For example, the electronic equipment may be implemented as shown in Figure 2: step S210 receives the solicited message of user；Step
Rapid S220 identifies the intention of the user according to the solicited message；Step S230, the intention that will identify that and is deposited in database
The history intention of the user of storage merges；Step S240 generates corresponding user response according to fused intention.
It should be noted that although being referred to several modules for acting the device executed in the above detailed description
Or unit, but this division is not enforceable.In fact, embodiment according to the present invention, above-described two
Or more the feature and function of module or unit can be embodied in a module or unit.Conversely, above-described
One module or the feature and function of unit can be to be embodied by multiple modules or unit with further division.
Through the above description of the embodiments, those skilled in the art is it can be readily appreciated that example described herein is implemented
Mode can also be realized by software realization in such a way that software is in conjunction with necessary hardware.Therefore, according to the present invention
The technical solution of embodiment can be embodied in the form of software products, which can store non-volatile at one
Property storage medium (can be CD-ROM, USB flash disk, mobile hard disk etc.) in or network on, including some instructions are so that a calculating
Equipment (can be personal computer, server, touch control terminal or network equipment etc.) executes embodiment according to the present invention
Method.
Those skilled in the art after considering the specification and implementing the invention disclosed here, will readily occur to of the invention its
Its embodiment.This application is intended to cover any variations, uses, or adaptations of the invention, these modifications, purposes or
Person's adaptive change follows general principle of the invention and including the undocumented common knowledge in the art of the present invention
Or conventional techniques.The description and examples are only to be considered as illustrative, and true scope and spirit of the invention are by following
Claim is pointed out.
It should be understood that the present invention is not limited to the precise structure already described above and shown in the accompanying drawings, and
And various modifications and changes may be made without departing from the scope thereof.The scope of the present invention is limited only by the attached claims.
Claims (14)
1. a kind of man-machine conversation's method characterized by comprising
Receive the solicited message of user；
The intention of the user is identified according to the solicited message；
The intention that will identify that and the history of the user stored in database intention merge；
Corresponding user response is generated according to fused intention.
2. man-machine conversation's method according to claim 1, which is characterized in that stored in the intention and database that will identify that
The user history intention merge, comprising:
Extract the slot value of the intention identified；
The slot value that the slot value extracted is intended to the history is merged, is intended to the intention that will identify that and the history
It is merged.
3. man-machine conversation's method according to claim 2, which is characterized in that man-machine conversation's method further include:
After the slot value for extracting the intention identified, the intention and the slot value are stored in the database.
4. man-machine conversation's method according to claim 3, which is characterized in that man-machine conversation's method further include:
Whether the quantity for judging that the history of the user stored in the database is intended to is more than predetermined threshold；
When the quantity that the history for determining the user is intended to is more than the predetermined threshold, delete the user's stored at first
History is intended to and corresponding slot value.
5. man-machine conversation's method according to claim 3, which is characterized in that man-machine conversation's method further include:
When the deposit duration that any history of the user stored in the database is intended to reaches scheduled duration, institute is deleted
State any history intention and corresponding slot value.
6. man-machine conversation's method according to claim 2, which is characterized in that corresponding being generated according to fused intention
Before user response, further includes:
Judge whether the slot value information of fused intention is complete；
When the slot value information for determining the fused intention is complete, generated according to the slot value information of the fused intention
Corresponding user response.
7. man-machine conversation's method according to claim 6, which is characterized in that man-machine conversation's method further include:
When the slot value information for determining the fused intention is imperfect, Xiang Suoshu user feedback prompt information.
8. man-machine conversation's method according to claim 2, which is characterized in that generate corresponding use according to fused intention
Family response, comprising:
It is retrieved according to the slot value information of fused intention, corresponding user response is generated based on search result；Or
According to the slot value information of fused intention, corresponding user response is generated by production model.
9. man-machine conversation's method according to any one of claim 2 to 8, which is characterized in that man-machine conversation's method
Further include:
When not generating effective user response, one or more slot values in the slot value information of fused intention are deleted, are obtained
To updated slot value information；
Corresponding user response is generated according to the updated slot value information.
10. man-machine conversation's method according to claim 9, which is characterized in that delete the slot value information of fused intention
In one or more slot values, comprising:
It is intended to the sequencing of storage according to the history of the user, deletes and stored at first in the slot value information of fused intention
One or more slot values.
11. man-machine conversation's method according to any one of claim 2 to 8, which is characterized in that extract the intention identified
Slot value, comprising:
Mode based on template matching and/or neural network extracts the slot value of the intention identified.
12. a kind of man-machine conversation's device characterized by comprising
Receiving unit, for receiving the solicited message of user；
Intention assessment unit, for identifying the intention of the user according to the solicited message；
The history intention of integrated unit, the user stored in the intention and database for will identify that merges；
Generation unit is responded, for generating corresponding user response according to fused intention.
13. a kind of electronic equipment characterized by comprising
Processor；And
Memory is stored with computer-readable instruction on the memory, and the computer-readable instruction is held by the processor
Man-machine conversation's method as described in any one of claims 1 to 11 is realized when row.
14. a kind of computer readable storage medium, is stored thereon with computer program, the computer program is executed by processor
Man-machine conversation method of the Shi Shixian as described in any one of claims 1 to 11.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711013017.0A CN110069606A (en) | 2017-10-26 | 2017-10-26 | Man-machine conversation's method, apparatus, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711013017.0A CN110069606A (en) | 2017-10-26 | 2017-10-26 | Man-machine conversation's method, apparatus, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110069606A true CN110069606A (en) | 2019-07-30 |
Family
ID=67364592
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711013017.0A Pending CN110069606A (en) | 2017-10-26 | 2017-10-26 | Man-machine conversation's method, apparatus, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110069606A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110827831A (en) * | 2019-11-15 | 2020-02-21 | 广州洪荒智能科技有限公司 | Voice information processing method, device, equipment and medium based on man-machine interaction |
| CN110971754A (en) * | 2019-10-28 | 2020-04-07 | 深圳绿米联创科技有限公司 | Information processing method, information processing device, electronic equipment and storage medium |
| CN111209297A (en) * | 2019-12-31 | 2020-05-29 | 深圳云天励飞技术有限公司 | Data query method and device, electronic equipment and storage medium |
| CN111353035A (en) * | 2020-03-11 | 2020-06-30 | 镁佳(北京)科技有限公司 | Man-machine conversation method and device, readable storage medium and electronic equipment |
| CN111538818A (en) * | 2020-03-26 | 2020-08-14 | 深圳云天励飞技术有限公司 | Data query method and device, electronic equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105094315A (en) * | 2015-06-25 | 2015-11-25 | 百度在线网络技术（北京）有限公司 | Method and apparatus for smart man-machine chat based on artificial intelligence |
| CN105589848A (en) * | 2015-12-28 | 2016-05-18 | 百度在线网络技术（北京）有限公司 | Dialog management method and device |
| CN106448670A (en) * | 2016-10-21 | 2017-02-22 | 竹间智能科技（上海）有限公司 | Dialogue automatic reply system based on deep learning and reinforcement learning |
| CN107203265A (en) * | 2017-05-17 | 2017-09-26 | 广东美的制冷设备有限公司 | Information interacting method and device |
-
2017
- 2017-10-26 CN CN201711013017.0A patent/CN110069606A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105094315A (en) * | 2015-06-25 | 2015-11-25 | 百度在线网络技术（北京）有限公司 | Method and apparatus for smart man-machine chat based on artificial intelligence |
| CN105589848A (en) * | 2015-12-28 | 2016-05-18 | 百度在线网络技术（北京）有限公司 | Dialog management method and device |
| CN106448670A (en) * | 2016-10-21 | 2017-02-22 | 竹间智能科技（上海）有限公司 | Dialogue automatic reply system based on deep learning and reinforcement learning |
| CN107203265A (en) * | 2017-05-17 | 2017-09-26 | 广东美的制冷设备有限公司 | Information interacting method and device |
Non-Patent Citations (3)
| Title |
|---|
| XIAODONG ZHANG等: "A Joint Model of Intent Determination and Slot Filling for Spoken Language Understanding", 《TWENTY-FIFTH INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE》 * |
| 智言科技: "浅谈垂直领域的chatbot", 《HTTPS://ZHUANLAN.ZHIHU.COM/P/29757034》 * |
| 赵晗 等: "对话理解中的槽填充", 《中国科技论文在线》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110971754A (en) * | 2019-10-28 | 2020-04-07 | 深圳绿米联创科技有限公司 | Information processing method, information processing device, electronic equipment and storage medium |
| CN110971754B (en) * | 2019-10-28 | 2022-09-27 | 深圳绿米联创科技有限公司 | Information processing method, device, electronic equipment and storage medium |
| CN110827831A (en) * | 2019-11-15 | 2020-02-21 | 广州洪荒智能科技有限公司 | Voice information processing method, device, equipment and medium based on man-machine interaction |
| CN111209297A (en) * | 2019-12-31 | 2020-05-29 | 深圳云天励飞技术有限公司 | Data query method and device, electronic equipment and storage medium |
| CN111353035A (en) * | 2020-03-11 | 2020-06-30 | 镁佳(北京)科技有限公司 | Man-machine conversation method and device, readable storage medium and electronic equipment |
| CN111538818A (en) * | 2020-03-26 | 2020-08-14 | 深圳云天励飞技术有限公司 | Data query method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11030412B2 (en) | System and method for chatbot conversation construction and management | |
| CN110069606A (en) | Man-machine conversation's method, apparatus, electronic equipment and storage medium | |
| CN112100349B (en) | Multi-round dialogue method and device, electronic equipment and storage medium | |
| CN106407178B (en) | A kind of session abstraction generating method, device, server apparatus and terminal device | |
| US9865264B2 (en) | Selective speech recognition for chat and digital personal assistant systems | |
| CN109145104B (en) | Method and device for dialogue interaction | |
| CN101297355B (en) | Systems and methods for responding to natural language speech utterance | |
| CN108986805B (en) | Method and apparatus for sending information | |
| US20170337261A1 (en) | Decision Making and Planning/Prediction System for Human Intention Resolution | |
| CN108255934A (en) | A kind of sound control method and device | |
| CN107833574A (en) | Method and apparatus for providing voice service | |
| CN107623614A (en) | Method and apparatus for pushed information | |
| CN111081280B (en) | Text-independent speech emotion recognition method and device and emotion recognition algorithm model generation method | |
| CN108345692A (en) | A kind of automatic question-answering method and system | |
| US20210150398A1 (en) | Conversational interchange optimization | |
| WO2008119270A1 (en) | A chatting robot system and a method, a device for chatting automatically | |
| CN107943914A (en) | Voice information processing method and device | |
| CN111831795B (en) | Multi-round dialogue processing method and device, electronic equipment and storage medium | |
| CN109325091A (en) | Update method, device, equipment and the medium of points of interest attribute information | |
| KR102076793B1 (en) | Method for providing electric document using voice, apparatus and method for writing electric document using voice | |
| CN112131885A (en) | Semantic recognition method and device, electronic equipment and storage medium | |
| CN111813900B (en) | Multi-round dialogue processing method and device, electronic equipment and storage medium | |
| CN110457689A (en) | Semantic processes method and relevant apparatus | |
| CN108933730A (en) | Information-pushing method and device | |
| CN108924218A (en) | Method and apparatus for pushed information |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210525 Address after: 100176 room 1004, 10th floor, building 1, 18 Kechuang 11th Street, Beijing Economic and Technological Development Zone, Daxing District, Beijing Applicant after: Beijing Huijun Technology Co.,Ltd. Address before: 100195 Beijing Haidian Xingshikou Road 65 West Cedar Creative Garden 4 District 11 Building East 1-4 Floor West 1-4 Floor Applicant before: BEIJING JINGDONG SHANGKE INFORMATION TECHNOLOGY Co.,Ltd. Applicant before: BEIJING JINGDONG CENTURY TRADING Co.,Ltd. |
|
| TA01 | Transfer of patent application right |