CN109145104A - For talking with interactive method and apparatus - Google Patents

For talking with interactive method and apparatus Download PDF

Info

Publication number
CN109145104A
CN109145104A CN201811146511.9A CN201811146511A CN109145104A CN 109145104 A CN109145104 A CN 109145104A CN 201811146511 A CN201811146511 A CN 201811146511A CN 109145104 A CN109145104 A CN 109145104A
Authority
CN
China
Prior art keywords
conversation object
conversation
target
matching
candidate
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
Application number
CN201811146511.9A
Other languages
Chinese (zh)
Other versions
CN109145104B (en
Inventor
牛正雨
姜迪
吴华
王海峰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Baidu Netcom Science and Technology Co Ltd
Original Assignee
Beijing Baidu Netcom Science and Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Baidu Netcom Science and Technology Co Ltd filed Critical Beijing Baidu Netcom Science and Technology Co Ltd
Priority to CN201811146511.9A priority Critical patent/CN109145104B/en
Publication of CN109145104A publication Critical patent/CN109145104A/en
Application granted granted Critical
Publication of CN109145104B publication Critical patent/CN109145104B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities
    • G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking

Abstract

The embodiment of the present application is disclosed for talking with interactive method and apparatus.One specific embodiment of this method includes: the sentence in response to receiving user's input, determines and is intended to word；In response to there is no the conversation objects to match with the intention word in the preset conversation object set of determination, the conversation object to match with the intention word is obtained；Acquired conversation object is determined as Matching conversation target；And acquired conversation object is added to the conversation object set；According to Matching conversation target, preset return operation is executed.The embodiment is realized to engage in the dialogue interaction according to the conversation object in the conversation object set that can dynamically update.

Description

For talking with interactive method and apparatus
Technical field
The invention relates to field of computer technology, and in particular to for talking with interactive method and apparatus.
Background technique
With the rapid development of artificial intelligence technology, Intelligent dialogue system increasingly attracts people's attention.Relevant skill Art is usually to use depth learning technology, indicate and reply generation strategy come learning characteristic using a large amount of data, is mainly used for Reply to user's sentence.
Summary of the invention
The embodiment of the present application is proposed for talking with interactive method and apparatus.
In a first aspect, the embodiment of the present application provide it is a kind of for talking with interactive method, this method comprises: in response to connecing The sentence of user's input is received, determines and is intended to word；In response to being not present in the preset conversation object set of determination and being intended to word phase Matched conversation object obtains the conversation object to match with intention word；Acquired conversation object is determined as Matching conversation Target；And acquired conversation object is added to conversation object set；According to Matching conversation target, preset reply is executed Operation.
In some embodiments, before executing preset return operation, this method further include: talk with mesh in response to determining There is the conversation object to match with intention word in mark set, the conversation object to match with intention word is determined as Matching conversation Target.
In some embodiments, the conversation object in above-mentioned conversation object set includes attribute information, and attribute information includes At least one of below: priority is stored in the time of set；And this method further include: obtain dialogue associated with word is intended to Target is as guiding target；According to the attribute information of the conversation object in guiding target and conversation object set, dialogue mesh is determined With the presence or absence of candidate conversation object in mark set；There is candidate conversation object in response to determining in conversation object set, according to time Conversation object is selected, preset return operation is executed.
In some embodiments, before in above-mentioned determining conversation object set with the presence or absence of candidate conversation object, the party Method further include: there is no the conversation objects to match with guiding target in conversation object set in response to determining, by guiding target It is added to conversation object set.
In some embodiments, above-mentioned attribute information includes the time of priority and deposit set；And above-mentioned determination pair It talks about in target collection with the presence or absence of candidate conversation object, comprising: by the conversation object of the highest priority in conversation object set It is determined as the first candidate conversation object；In response to determining that the number of the first candidate conversation object is 1, by the first candidate conversation object It is determined as the second candidate conversation object；In response to determining that the number of the first candidate conversation object is greater than 1, according to the first candidate dialogue The the first candidate conversation object for being stored in set at first is determined as the second candidate conversation object by the time of the deposit set of target； In response to determining that the second candidate conversation object is not Matching conversation target, determine in conversation object set there is candidate dialogue mesh Mark；And the second candidate conversation object is determined as candidate conversation object.
In some embodiments, this method further include: in response to determining the guiding target being added in conversation object set It is not candidate conversation object, guiding target is deleted from conversation object set.
In some embodiments, above-mentioned determining intention word, comprising: by above-mentioned input by sentence to intention word point trained in advance Class model obtains being intended to word, wherein is intended to the corresponding relationship that word disaggregated model is used to characterize sentence and be intended between word.
In some embodiments, above-mentioned attribute information includes the time of deposit set；And this method further include: determining pair It talks about in target collection with the presence or absence of other conversation objects other than Matching conversation target；It is deposited in response to determining in conversation object set Other conversation objects other than Matching conversation target are right according to this for other conversation objects other than Matching conversation target The attribute information for talking about target, determines whether the conversation object is completed；In response to determining that the conversation object is completed, by the conversation object It is deleted from conversation object set.
In some embodiments, above-mentioned attribute information further includes task parameters information, and task parameters information includes task ginseng Several and parameter value；And whether the above-mentioned determination conversation object is completed, comprising: according to the task parameters and parameter of the conversation object Value, determines whether the conversation object is completed.
In some embodiments, this method further include: according to Matching conversation target, determine the task ginseng of Matching conversation target Number；According to the task parameters of above-mentioned sentence and identified Matching conversation target, the parameter value of Matching conversation target is generated.
In some embodiments, above-mentioned according to Matching conversation target, execute preset return operation, comprising: according to matching The task parameters and parameter value of conversation object, determine whether Matching conversation target is completed；In response to determining that Matching conversation target is complete At execution predetermined registration operation corresponding with Matching conversation target；In response to determining that Matching conversation target does not complete, according to matching pair The task parameters and parameter value of target are talked about, default return operation corresponding with Matching conversation target is executed.
In some embodiments, above-mentioned attribute information includes triggering mode information and priority；And this method further include: For the conversation object in conversation object set, in response to determining that the triggering mode information of the conversation object meets preset condition, It is higher level by the priority modification of the conversation object.
Second aspect, the embodiment of the present application provide it is a kind of for talking with interactive device, the device include: be intended to determine Unit is configured in response to receive the sentence of user's input, determines and be intended to word；First adding unit, is configured to respond to In determining in preset conversation object set there is no the conversation object to match with intention word, obtains and be intended to what word matched Conversation object；Acquired conversation object is determined as Matching conversation target；And acquired conversation object is added to pair Talk about target collection；First replys unit, is configured to execute preset return operation according to Matching conversation target.
In some embodiments, device further include: the first determination unit is configured in response to determine conversation object collection There is the conversation object to match with intention word in conjunction, the conversation object to match with intention word is determined as Matching conversation mesh Mark.
In some embodiments, the conversation object in above-mentioned conversation object set includes attribute information, and attribute information includes At least one of below: priority is stored in the time of set；And device further include: acquiring unit is configured to obtain and anticipate The associated conversation object of figure word is as guiding target；Second determination unit, is configured to according to guiding target and conversation object The attribute information of conversation object in set determines in conversation object set with the presence or absence of candidate conversation object；Second reply sheet Member is configured in response to determine in conversation object set there is candidate conversation object, according to candidate conversation object, executes default Return operation.
In some embodiments, device further include: the second adding unit is configured in response to determine conversation object collection There is no the conversation object to match with guiding target in conjunction, guiding target is added to conversation object set.
In some embodiments, above-mentioned attribute information includes the time of priority and deposit set；And above-mentioned second really Order member includes: the first determining module, is configured to for the conversation object of the highest priority in conversation object set being determined as First candidate conversation object；Second determining module is configured in response to determine that the number of the first candidate conversation object is 1, will First candidate conversation object is determined as the second candidate conversation object；Third determining module is configured in response to determine that first waits It selects the number of conversation object to be greater than 1, according to the time of the deposit set of the first candidate conversation object, the of set will be stored at first One candidate conversation object is determined as the second candidate conversation object；4th determining module is configured in response to determine that second is candidate Conversation object is not Matching conversation target, determines in conversation object set there is candidate conversation object；And it is candidate right by second Words target is determined as candidate conversation object.
In some embodiments, device further include: first deletes unit, is configured in response to determination and is added to dialogue Guiding target in target collection is not candidate conversation object, and guiding target is deleted from conversation object set.
In some embodiments, above-mentioned intent determination unit, is further configured to: by above-mentioned input by sentence to preparatory instruction Experienced intention word disaggregated model obtains being intended to word, wherein is intended to pair that word disaggregated model is used to characterize sentence and be intended between word It should be related to.
In some embodiments, above-mentioned attribute information includes the time of deposit set；And device further include: third is true Order member is configured to determine in conversation object set with the presence or absence of other conversation objects other than Matching conversation target；Second Unit is deleted, is configured in response to determine in conversation object set there are other conversation objects other than Matching conversation target, For other conversation objects other than Matching conversation target, according to the attribute information of the conversation object, determine that the conversation object is No completion；In response to determining that the conversation object is completed, which is deleted from conversation object set.
In some embodiments, above-mentioned attribute information further includes task parameters information, and task parameters information includes task ginseng Several and parameter value；And above-mentioned second deletion unit, it is further configured to: according to the task parameters and parameter of the conversation object Value, determines whether the conversation object is completed.
In some embodiments, device further include: the 4th determination unit is configured to according to Matching conversation target, really Determine the task parameters of Matching conversation target；Generation unit is configured to according to above-mentioned sentence and identified Matching conversation target Task parameters, generate Matching conversation target parameter value.
In some embodiments, above-mentioned first reply unit includes: the 5th determining module, is configured to according to Matching conversation The task parameters and parameter value of target, determine whether Matching conversation target is completed；First execution module is configured in response to really Determine the completion of Matching conversation target, executes predetermined registration operation corresponding with Matching conversation target；Second execution module is configured to ring It should not be completed in determining Matching conversation target, according to the task parameters and parameter value of Matching conversation target, execution and Matching conversation The corresponding default return operation of target.
In some embodiments, above-mentioned attribute information includes triggering mode information and priority；And the device further include: Unit is modified, is configured to for the conversation object in conversation object set, in response to determining the triggering mode of the conversation object Information meets preset condition, is higher level by the priority modification of the conversation object.
The third aspect, the embodiment of the present application provide a kind of electronic equipment, which includes: one or more processing Device；Storage device is stored thereon with one or more programs；When one or more programs are executed by one or more processors, So that one or more processors realize the method as described in implementation any in first aspect.
Fourth aspect, the embodiment of the present application provide a kind of computer-readable medium, are stored thereon with computer program, should The method as described in implementation any in first aspect is realized when program is executed by processor.
The language provided by the embodiments of the present application for talking with interactive method and apparatus, inputted in response to receiving user Sentence determines according to the sentence and is intended to word.Later, in response to being not present and above-mentioned intention word in the preset conversation object set of determination The conversation object to match obtains the conversation object to match with above-mentioned intention word.Next, acquired conversation object is true It is set to Matching conversation target, and acquired conversation object is added to above-mentioned conversation object set.Finally, according to matching pair Target is talked about, preset return operation is executed.To realize according to the dialogue mesh in the conversation object set that can dynamically update Mark the interaction that engages in the dialogue.
Detailed description of the invention
By reading a detailed description of non-restrictive embodiments in the light of the attached drawings below, the application's is other Feature, objects and advantages will become more apparent upon:
Fig. 1 is that one embodiment of the application can be applied to exemplary system architecture figure therein；
Fig. 2 is the flow chart according to one embodiment for talking with interactive method of the application；
Fig. 3 is the schematic diagram according to an application scenarios for talking with interactive method of the embodiment of the present application；
Fig. 4 is the flow chart according to another embodiment for talking with interactive method of the application；
Fig. 5 is the structural schematic diagram according to one embodiment for talking with interactive device of the application；
Fig. 6 is adapted for the structural schematic diagram for the computer system for realizing the electronic equipment of the embodiment of the present application.
Specific embodiment
The application is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched The specific embodiment stated is used only for explaining related invention, rather than the restriction to the invention.It also should be noted that in order to Convenient for description, part relevant to related invention is illustrated only in attached drawing.
It should be noted that in the absence of conflict, the features in the embodiments and the embodiments of the present application can phase Mutually combination.The application is described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.
Fig. 1 is shown can example using the application for talking with interactive method or for talking with interactive device Property framework 100.
As shown in Figure 1, system architecture 100 may include terminal device 101,102,103, network 104 and server 105. Network 104 between terminal device 101,102,103 and server 105 to provide the medium of communication link.Network 104 can be with Including various connection types, such as wired, wireless communication link or fiber optic cables etc..
Terminal device 101,102,103 is interacted by network 104 with server 105, to receive or send message etc..Terminal Various telecommunication customer end applications can be installed in equipment 101,102,103, for example, web browser applications, the application of shopping class, Searching class application, instant messaging tools, mailbox client, social platform software, the application of text editing class etc..
Terminal device 101,102,103 can be hardware, be also possible to software.When terminal device 101,102,103 is hard When part, it can be the various electronic equipments that can support dialogue interaction, including but not limited to smart phone, tablet computer, MP3 is broadcast Put device (Moving Picture Experts Group Audio Layer III, dynamic image expert's compression standard audio layer Face 3), MP4 (Moving Picture Experts Group Audio Layer IV, dynamic image expert's compression standard audio Level 4) player, pocket computer on knee and desktop computer etc..When terminal device 101,102,103 is software, It may be mounted in above-mentioned cited electronic equipment.Multiple softwares or software module may be implemented into (such as providing in it Distributed Services), single software or software module also may be implemented into.It is not specifically limited herein.
Server 105 can be to provide the server of various services, for example, carry out on terminal device 101,102,103 pair Words interaction provides the background server supported.Background server can carry out intention analysis to the user's sentence received, and raw Terminal device is fed back at analysis result or by analysis result (as being intended to word).
It should be noted that the processing step of above-mentioned intention analysis can also be executed by terminal device.At this point it is possible to not There are networks 104 and server 105.
It should be noted that server can be hardware, it is also possible to software.When server is hardware, may be implemented At the distributed server cluster that multiple servers form, individual server also may be implemented into.It, can when server is software To be implemented as multiple softwares or software module (such as providing Distributed Services), single software or software also may be implemented into Module.It is not specifically limited herein.
It should be noted that the embodiment of the present application provided by for talk with interactive method generally by terminal device 101, 102, it 103 executes, correspondingly, is generally positioned in terminal device 101,102,103 for talking with interactive device.
It should be noted that terminal device 101,102,103 can also be used for receiving the sentence of user's input, by what is received Sentence is sent to server 105, and server 105 is allowed to execute this for talking with interactive method.
It should be understood that the number of terminal device, network and server in Fig. 1 is only schematical.According to realization need It wants, can have any number of terminal device, network and server.
With continued reference to Fig. 2, the process of one embodiment for talking with interactive method according to the application is shown 200.This for talking with interactive method the following steps are included:
Step 201, the sentence inputted in response to receiving user, determines and is intended to word.
In the present embodiment, can lead to for talking with the executing subject (terminal device as shown in Figure 1) of interactive method It crosses various methods and determines and be intended to word.Wherein, above-mentioned sentence can be textual form.It can be user and is directly inputted by text Textual form sentence.It is also possible to user and voice input is carried out by microphone, language is then utilized by above-mentioned executing subject The sentence of speech form is converted to textual form by sound identification technology.Above-mentioned intention word can be for expressing sentence intention Word.It can be the keyword being present in sentence, be also possible to the phrase for characterizing intention expressed in sentence.It is above-mentioned Being intended to word may belong to different types.The above-mentioned type can include but is not limited at least one in following four type: not busy Merely type, knowledge chat-type, question and answer type, task type.Specifically, the intention word for belonging to chat type can be the " spare time Merely ", for characterizing the open field chat for being not limited to specific topics.The intention word for belonging to knowledge chat-type may include characterization The word of some theme, such as can be " historical knowledge chat ", " sports knowledge chat " etc..Belong to the intention word of question and answer type It may include characterizing the word of some particular problem, such as can be " inquiry hotel ", " inquiry weather " etc..Belong to task type Intention word may include the word for characterizing some specific tasks, such as can be " ordering air ticket ", " setting alarm clock " etc..
In the present embodiment, the sentence that above-mentioned executing subject can be inputted in response to receiving user, in accordance with the following steps It determines and is intended to word:
The first step extracts the keyword of the sentence of user's input.
Above-mentioned keyword can be used for characterizing the particular content of sentence.Above-mentioned executing subject can be extracted by various modes The keyword of the sentence of user's input.Such as TF-IDF (Term Frequency-Inverse Document can be passed through Frequency, word frequency-inverse document frequency), TextRank (a kind of algorithm by PageRank algorithm improvement), RAKE (Rapid Automatic Keyword Extraction, fast automatic keyword extraction) algorithm determines the sentence of user's input Keyword.
It should be noted that the above-mentioned keyword for determining sentence using TF-IDF, TextRank, RAKE algorithm is wide at present The well-known technique of general research and application, details are not described herein.
Second step determines from the mapping table constructed in advance according to extracted keyword and is intended to word.
It is preparatory based on the statistics to a large amount of user's sentence and intention word that above-mentioned mapping table can be technical staff Mapping table formulating, being stored with multiple keywords with the corresponding relationship for being intended to word.Wherein, corresponding relationship can be one Keyword corresponds to an intention word；Be also possible to by multiple crucial phrases at the corresponding intention word of entirety.Above-mentioned execution master Body can determine corresponding intention word according to extracted keyword from above-mentioned mapping table.For example, extracted Keyword can be " Bird's Nest " " how to get to ", corresponding intention word can be " enquiring route ".
In some optional implementations of the present embodiment, the sentence that above-mentioned executing subject can also input user is defeated Enter to intention word disaggregated model trained in advance, obtains being intended to word.Wherein, it is intended that word disaggregated model is for characterizing sentence and being intended to Corresponding relationship between word.
Above-mentioned intention word disaggregated model can be using machine learning method, based on a large amount of training sample to various initial The model that disaggregated model training obtains.As an example, above-mentioned intention word disaggregated model can be obtained by following steps training:
The first step obtains initial intention word disaggregated model.
The executing subject of training step can obtain initial intention word disaggregated model first.Wherein, initially it is intended to word classification Model can be the existing various models that can be used for classifying.For example, SVM (Support Vector Machine, support to Amount machine), CNN (Convolutional Neural Network, convolutional neural networks) etc..
Second step obtains training sample set.
Each training sample may include that sample sentence and sample corresponding with sample sentence are intended to word.Above-mentioned training sample It can obtain in several ways.As an example, can be arranged by technical staff according to a large amount of historical data after It arrives, the intention word marked by sentence and by hand is associated storage to obtain training sample.It as another example, can be from mutual Extract sentence and intention word corresponding with sentence in networking in existing knowledge base, then by sentence and corresponding intention word into Obtained from row associated storage.A large amount of training sample can form above-mentioned training sample set.
Third step is intended to the defeated of word disaggregated model using the sample sentence of the training sample in training sample set as initial Enter, sample corresponding with the sample sentence of input is intended to word as desired output, is obtained using machine learning method training State initial intention word disaggregated model.
Specifically, obtained intention word can be calculated first with preset loss function and sample is intended between word Difference.It is then possible to adjust the initial parameter for being intended to word disaggregated model, and default meeting based on obtained difference is calculated Training termination condition in the case where, terminate training.Herein, preset trained termination condition can include but is not limited to down toward One item missing: it calculates resulting difference and is less than default discrepancy threshold；Training duration is more than preset duration；Frequency of training is more than default time Number.
The initial intention word disaggregated model obtained after training is determined as being intended to word disaggregated model by the 4th step.
It should be noted that in the present embodiment, the executing subject of training step can with for talking with the method interacted Executing subject it is identical, can also be different.If identical, the executing subject of training step can will be intended to after the completion of training The structural information of word disaggregated model and the parameter value of parameter are stored in local.If it is different, then the executing subject of training step can The parameter value of the trained structural information for being intended to word disaggregated model and parameter to be sent to and is used to talk with after the completion of training The executing subject of interactive method.
Step 202, in response to being not present and being intended to word the conversation object to match in the preset conversation object set of determination, Obtain the conversation object to match with intention word；Acquired conversation object is determined as Matching conversation target；And it will be obtained The conversation object taken is added to conversation object set.
In the present embodiment, above-mentioned conversation object can be the information of structuring, may include being intended to word in conversation object. Above-mentioned preset conversation object set can be the set constructed in advance, can store in above-mentioned executing subject.It is above-mentioned to hold Row main body can operate the conversation object in above-mentioned conversation object set, such as conversation object is added to above-mentioned collection It closes, or conversation object is deleted from above-mentioned set.It should be noted that the conversation object collection involved by the present embodiment Conjunction refers to the conversation object set, only including element can change.That is, working as the conversation object After set updates, when executing subject again operates the conversation object in the conversation object set, the object of operation is this The current included conversation object of conversation object set, and it is not necessarily conversation object included when building set in advance.
In the present embodiment, above-mentioned executing subject can identify intentional word can come from the intention constructed in advance Word information library.Correspondingly, above-mentioned executing subject is able to carry out what matched all conversation objects can be from constructing in advance Conversation object information bank.The matching of above-mentioned intention word and conversation object can refer to that the two is same or similar.Above-mentioned executing subject can Whether matched with intention word with determining conversation object by various modes.As an example, if when constructing conversation object information bank Using information bank same as word information library is intended to, then conversation object can be corresponded with word is intended to.When executing subject carries out When matching, intention word included by conversation object can be compared and be intended to word determined by step 201, if they are the same, it is determined that Matching.It as another example, can be using the information bank different from word information library is intended to when constructing conversation object information bank.Example Such as, conversation object information bank can be what technical staff was voluntarily constructed by manual compiling mode, it is intended that word information library can be It is constructed using network existing resource.At this point it is possible to by determine conversation object and determined by step 201 intention word it Between whether meet preset condition to determine whether to match.Preset condition for example can be in preset conversation object-intention Inquired in word relation table conversation object and be intended to word have corresponding relationship, be also possible to intention word included by conversation object and Text similarity between the intention word determined by step 201 is more than preset threshold.It should be noted that above-mentioned text is similar The calculating of degree is the well-known technique studied and applied extensively at present, and details are not described herein.
In the present embodiment, above-mentioned executing subject can determine through the above way in preset conversation object set whether In the presence of the conversation object to match with intention word.In response to being not present in the preset conversation object set of determination and being intended to word phase The conversation object matched, above-mentioned executing subject can obtain pair to match with intention word from the conversation object information bank constructed in advance Talk about target.Optionally, the above-mentioned conversation object information bank constructed in advance can store in the local of above-mentioned executing subject, can also be with It is stored in the database server with the communication connection of above-mentioned executing subject.Correspondingly, above-mentioned executing subject can from it is local or from It communicates the conversation object information bank constructed in advance in the database server of connection and obtains pair to match with intention word Talk about target.Then, acquired conversation object can be determined as Matching conversation target by above-mentioned executing subject.Wherein, matching pair Words target is used to characterize the conversation object to match with identified intention word.Next, above-mentioned executing subject can will be obtained The conversation object taken is added to above-mentioned preset conversation object set.To by adding member into preset conversation object set Element realizes the update to above-mentioned set.
In some optional implementations in the present embodiment, conversation object can also include attribute information.Wherein, belong to Property information can be used for characterizing associated with conversation object parameter.Above-mentioned attribute information can include but is not limited to it is following at least One: priority, task parameters information are stored in the time of set, status information, triggering mode information.Wherein, priority can be with It is the priority of conversation object.In general, the priority of conversation object can be with the type phase of the intention word included by conversation object It is corresponding.In general, can be by the "high" priority for being intended to word and corresponding to conversation object of question and answer type and task type；Knowledge is chatted Its type be intended to word correspond to conversation object " in " priority；The word that is intended to for chatting type is corresponded into conversation object " low " priority.
Task parameters information is including can have different forms in the different types of conversation object for being intended to word.Task ginseng Number information may include task parameters and parameter value.It is including appointing in the conversation object of intention word for belonging to knowledge chat-type Business parameter can be intended to type of theme belonging to word, and parameter value can be included in sentence, is corresponding with conversation object Type of theme.Such as, it is intended that word can be " chat of sight spot knowledge ", and task parameters can be " knowledge chat theme ", parameter value It can be " Xiamen ".It is including in the conversation object of intention word for belonging to question and answer type, task parameters can be the specific of question and answer Content, parameter value can be included in sentence, particular content corresponding with conversation object.Such as, it is intended that word, which can be, " to be looked into Ask weather ", task parameters can be " question and answer particular content ", and parameter value can be " Beijing ".It is including belonging to task type It is intended in the conversation object of word, task parameters can be used for characterizing to complete the corresponding required by task of intention word of task type The information to be obtained.Parameter value can be used for characterizing to complete the corresponding task of the intention word letter obtained of task type Breath.As an example, being intended to word can be " ordering air ticket ", then task parameters can be " departure time ", " departure place ", " destination " Deng.Correspondingly, parameter value can be specifying information corresponding to above-mentioned each task parameter, such as " 10:00 tomorrow ", " Beijing ", " Shanghai ".
The task parameters of conversation object can be preset.For example, conversation object and task parameters can be preset Incidence relation.As an example, the task parameters of the conversation object including being intended to word " subscribing dining room " can be set to " when dining Between ", " dining room title ", " number " etc..The parameter value of conversation object can be determined according to the relevant information of user.Above-mentioned use The relevant information at family may include the information that user inputted in other scenes, also may include being installed by executing subject Information determined by other applications.For example, above-mentioned executing subject can pass through schedule in the conversation object of " ordering air ticket " Table class application program inquires the departure place information implied in the departure time, can also pass through GPS (Global Positioning System, global positioning system) determine current position.
The time of deposit set can be above-mentioned executing subject for conversation object from the conversation object information bank constructed in advance It is added to the time of above-mentioned preset conversation object set.In this case, only in preset conversation object set The time of the deposit set of conversation object just has practical significance.
The status information of conversation object can be used for characterizing whether dialogue corresponding to conversation object is carrying out.For example, Dialogue corresponding to can indicating conversation object with " activation " is carrying out.It can be indicated corresponding to conversation object with " hang-up " Talk with paused.In such cases, dialogue corresponding to conversation object is also not finished, but has been switched to other dialogue.? Relevant information (such as task parameters information of task class) under this dialogue is saved, in order to the backtracking of dialogue.In general, In preset conversation object set, can have multiple status informations is the conversation object of " hang-up ", to realize multiple conversation objects Coexist.In practice, flag bit field, which can be set, for characterizing the status information of conversation object, such as with " 1 " indicates " sharp It is living ", it is indicated " to hang up " with " 0 ".
The triggering mode of conversation object may include user's triggering and system trigger.User triggers for characterizing and talking with mesh Marking corresponding dialogue is initiated by user.In practice, above-mentioned executing subject is usually expressed as in response to receiving user's input Sentence determined by conversation object.System trigger for characterizing dialogue corresponding with conversation object is initiated by system.It is real In trampling, above-mentioned executing subject can be shown as according to preset scheduling logic, actively switch to new conversation object.
Optionally, for the conversation object in conversation object set, in response to determining that the triggering mode of the conversation object is believed Breath meets preset condition, and the priority modification of the conversation object can be higher level by executing subject.Above-mentioned triggering mode letter Breath can be the number of different triggering modes.Above-mentioned preset condition can be triggering mode be user triggering number preset Reach preset priority promotion threshold value in period, being also possible to triggering mode continuously is the number of user's triggering more than default Frequency threshold value.By the priority modification of conversation object be higher level for example can be by " in " priority modification is that "high" is excellent First grade.
In some optional implementations in the present embodiment, before executing preset return operation, the above method It can also include: in response to there is the dialogue mesh to match with identified intention word in the preset conversation object set of determination The conversation object to match with identified intention word can be determined as Matching conversation target by mark, above-mentioned executing subject.Wherein, It can in advance include conversation object in preset conversation object set.Above-mentioned executing subject can be determined default using aforesaid way Conversation object set in whether there is and identified be intended to the conversation object that matches of word.Optionally, above-mentioned executing subject Can also according to be intended to word corresponding to conversation object attribute information determine in preset conversation object set with the presence or absence of with The conversation object to match.As an example, when be intended to word it is identical and be intended to the attribute information of conversation object corresponding to word with When the attribute information difference of the conversation object in preset conversation object set, above-mentioned executing subject can determine preset dialogue Matched conversation object is not present in target collection.When the attribute letter for being intended to conversation object corresponding to word and intention word When ceasing with the intention word of the conversation object in conversation object set and identical attribute information, above-mentioned executing subject can determine pre- If conversation object set in there are matched conversation objects.As an example, meaning determined by the sentence of user's input Figure word be " chat of sight spot knowledge ", corresponding to conversation object parameter value be " Xiamen ".So, preset conversation object collection In conjunction, it is intended that word is " chat of sight spot knowledge ", the conversation object that parameter value is " Xi'an " is matched with above-mentioned intention word Conversation object.In response to there are matched conversation object, above-mentioned executing subjects in the preset conversation object set of determination The conversation object to match with identified intention word can be determined as Matching conversation target.
In some optional implementations in the present embodiment, the above method can also include: by above-mentioned Matching conversation The status information of target is revised as the ongoing information of dialogue corresponding to characterization conversation object, such as status information is modified For " activation ".
In some optional implementations in the present embodiment, the above method can also include: according to above-mentioned matching pair Target is talked about, determines the task parameters of above-mentioned Matching conversation target；The sentence and identified Matching conversation mesh inputted according to user Target task parameters generate the parameter value of above-mentioned Matching conversation target.It may include attribute information in above-mentioned Matching conversation target. Attribute information may include task parameters information.Task parameters information may include task parameters.Therefore, above-mentioned executing subject is aobvious So corresponding task parameters, the i.e. task parameters of Matching conversation target can be determined according to Matching conversation target.It is above-mentioned to hold Row main body can also utilize machine learning algorithm, and the input by sentence of user's input to argument sequence trained in advance is marked mould Type obtains included in above-mentioned sentence, the parameter value of above-mentioned Matching conversation target.Wherein, argument sequence marking model can be with For characterizing the corresponding relationship between sentence and parameter value.The initial model of above-mentioned parameter sequence labelling model can be and can answer For the various models of sequence labelling problem, including but not limited to MEMM (Maximum Entropy Markov Model, maximum Entropy Markov model), CRF (Conditional Random Field, condition random field) model, RNN (Recurrent Neural Networks, Recognition with Recurrent Neural Network).As an example, user input sentence for example can be " 5 points of tomorrow afternoon, The dining room XX is allowed to reserve the position of 4 people ".It is available by above-mentioned input by sentence to argument sequence marking model trained in advance Included in above-mentioned sentence, the parameter value of above-mentioned Matching conversation target, such as " 5 points of tomorrow afternoon ", " dining room XX ", " 4 people ". It should be noted that solving the problems, such as that sequence labelling is research and application extensively at present using machine learning model trained in advance Well-known technique, details are not described herein.
Step 203, according to Matching conversation target, preset return operation is executed.
In the present embodiment, above-mentioned executing subject can be according to the affiliated type of intention word included in conversation object not Together, different return operations is executed.In general, above-mentioned executing subject according to different type belonging to word is intended to, can be called corresponding Corpus replied.The mode of above-mentioned reply can be is replied in the form of text, is also possible to utilize speech synthesis Technology is replied in the form of broadcasting by the loudspeaker that above-mentioned executing subject communicates to connect.Above-mentioned corpus may include needle Corpus is talked with to the chat from Internet user for chatting type, also may include the particular text for knowledge chat-type The related text data of the themes such as resource, such as history, sight spot, film.Belong to the conversation object of question and answer type for being intended to word, Above-mentioned return operation can be problem included in inquiry session target in the knowledge mapping constructed in advance, by what is inquired As a result as reply；It is also possible to directly go retrieval answer according to problem included in conversation object, such as is drawn using search The other applications (for example, cuisines recommend class application, map class application) in above-mentioned executing subject are held up or are mounted on from interconnection It is inquired in net, and using the result inquired as reply.Belong to the conversation object of task type, above-mentioned reply behaviour for being intended to word It can be and putd question to for specific task definition, is also possible to through API (Application Programming Interface, application programming interface) relevant task (for example, ticket booking software is called to carry out plane ticket booking operation) is completed, Simple confirmation (such as " air ticket has been ordered ") after can also be completion task.
In some optional implementations of the present embodiment, the conversation object in conversation object set may include attribute Information, attribute information may include the time of deposit set.The above method further include: determine in above-mentioned conversation object set whether There are other conversation objects other than Matching conversation target；In response to determine conversation object set in there are Matching conversation target with Other outer conversation objects, for other conversation objects other than Matching conversation target, according to the attribute information of the conversation object, Determine whether the conversation object is completed；In response to determining that the conversation object is completed, by the conversation object from above-mentioned conversation object collection It is deleted in conjunction.Wherein, above-mentioned executing subject according to the attribute information of conversation object, can determine dialogue mesh by various modes Whether mark is completed.As an example, can determine whether conversation object is completed according to the time of the deposit set of conversation object.Tool Body, above-mentioned executing subject can by conversation object in set there are when it is long be greater than preset duration threshold value, determine dialogue Target has been completed.Then, executing subject can will determine that the conversation object completed is deleted from above-mentioned conversation object set, from And complete the update to above-mentioned conversation object set.Optionally, in response to determining that conversation object does not complete, above-mentioned executing subject is also It can be the paused information of dialogue corresponding to the characterization conversation object by the status modifier of unfinished conversation object, such as For " hang-up ".As another example, in conversation object in set there are when fail to grow up in the case where preset duration threshold value, Whether simple confirmation or evaluation can also be carried out according to user, to determine whether conversation object is completed.For example, conversation object For " inquiry hotel ", after above-mentioned executing subject has replied alternative hotel, user replys " ordering first ".At this point, above-mentioned hold Row main body can be confirmed according to user, determine that conversation object corresponding to " inquiry hotel " has been completed.
Optionally, attribute information can also include task parameters and parameter value.Above-mentioned executing subject can also be according to dialogue The task parameters and parameter value of target, determine whether conversation object is completed.Specifically, the dialogue of task type is belonged in intention word In target, whether above-mentioned executing subject can all have corresponding parameter value according to each task parameter of conversation object, determine Whether conversation object is completed.When each task parameter of conversation object all has corresponding parameter value, above-mentioned executing subject can To determine that conversation object is completed.
Optionally, above-mentioned executing subject can determine whether Matching conversation target is completed by the above method.In response to true Determine the completion of Matching conversation target, above-mentioned executing subject can execute predetermined registration operation corresponding with Matching conversation target.Wherein, in advance It calls corresponding API to complete task if operation can be, such as opens map class application and start to navigate, open " calendar " application and add Add backlog etc..Optionally, above-mentioned predetermined registration operation can also be that returning to corresponding task completes " to open as a result, for example replying Begin to navigate " or " new backlog has added ".In response to determining that Matching conversation target does not complete, above-mentioned executing subject can root According to the task parameters and parameter value of Matching conversation target, default return operation corresponding with Matching conversation target is executed.Wherein, Preset return operation can be the enquirement for the task parameters for not having parameter value.As an example, user's input " helps me to set Conversation object is determined as " setting alarm clock " by the alarm clock of a tomorrow ", executing subject, and task parameters are " time ", and parameter value can not be true It is fixed.In this case, preset return operation can be to user put question to " be set as several points? "
With continued reference to one that Fig. 3, Fig. 3 are the application scenarios according to the embodiment of the present application for talking with interactive method Schematic diagram 300.In the application scenarios of Fig. 3, the sentence 302 of the input of user 301 " me is helped to set 7 points of alarm clock on a tomorrow morning ".Terminal Equipment 303 determines that the intention word of sentence 302 is " setting alarm clock " 304 in response to receiving sentence 302.Preset conversation object collection It include " chat " conversation object and " ordering air ticket " conversation object in conjunction.Later, terminal device 303 is in response to the preset dialogue of determination There is no the conversation object to match with above-mentioned intention word 304 in target collection, acquisition " is set noisy with what above-mentioned intention word matched Clock " conversation object 305.Next, acquired conversation object 305 is determined as Matching conversation target by terminal device 303, and Conversation object 305 is added to preset conversation object set, as shown in figures 306.Later, terminal device 303 is according to matching Conversation object " sets alarm clock ", opens " alarm clock " application program being mounted in terminal device, and the addition time is the new noisy of " 7:00 " Clock.Finally, terminal device 303 can return to implementing result.Optionally, terminal device 303 can also export reply and " set for you Set the alarm clock of 7:00 tomorrow morning ", as shown in 307 in figure.
The method provided by the above embodiment of the application, first in response to the sentence for receiving user's input, according to the language Sentence, which determines, is intended to word.Later, in response to there is no pairs to match with above-mentioned intention word in the preset conversation object set of determination Target is talked about, the conversation object to match with above-mentioned intention word is obtained.Next, acquired conversation object is determined as matching pair Target is talked about, and acquired conversation object is added to above-mentioned conversation object set.Finally, being held according to Matching conversation target The preset return operation of row.It is engaged in the dialogue to realize according to the conversation object in the conversation object set that can dynamically update Interaction.
With further reference to Fig. 4, it illustrates the processes 400 of another embodiment for talking with interactive method.The use In the process 400 of the method for dialogue interaction, comprising the following steps:
Step 401, the sentence inputted in response to receiving user, determines and is intended to word.
Step 402, in response to being not present and being intended to word the conversation object to match in the preset conversation object set of determination, Obtain the conversation object to match with intention word；Acquired conversation object is determined as Matching conversation target；And it will be obtained The conversation object taken is added to conversation object set.
Step 403, according to Matching conversation target, preset return operation is executed.
Above-mentioned steps 401, step 402, step 403 respectively with step 201, step 202, the step in previous embodiment 203 is consistent, and the description above with respect to step 201, step 202 and step 203 is also applied for step 401, step 402 and step 403, details are not described herein again.
Step 404, conversation object associated with word is intended to is obtained as guiding target.
In the present embodiment, above-mentioned guiding target is determined for characterizing above-mentioned executing subject according to identified intention word Conversation object associated therewith.Above-mentioned executing subject can be obtained related to word is intended to by the intention map constructed in advance The conversation object of connection.Wherein, above-mentioned intention map is used to indicate the incidence relation between conversation object.Above-mentioned association be often referred to It is related and different.Incidence relation can be obtained by arranging to existing corpus resource a large amount of on internet, It can manually be marked to obtain by technical staff.For example, conversation object associated with " ordering air ticket " can be " inquiry Weather ".As an example, above-mentioned executing subject, which can export, " recommends the weather in the city XX for you after the completion of " ordering air ticket " conversation object Information: shower, please with good rain gear ".Wherein, the city XX is with can be the above-mentioned trip purpose made a reservation.Alternatively it is also possible to by Technical staff makes conversation object incidence relation table by carrying out analysis and arrangement to a large amount of dialog history data.Wherein, right Corresponding relationship between conversation object of the words target association relation table for characterizing conversation object and being associated.Above-mentioned execution master Body can obtain conversation object associated with word is intended to as guiding target by above-mentioned conversation object incidence relation table.
In some optional implementations of the present embodiment, the above method further include: in response to determining conversation object collection There is no the conversation object to match with above-mentioned guiding target in conjunction, above-mentioned guiding target can be added to by above-mentioned executing subject Preset conversation object set, to realize the update to preset conversation object set.
Step 405, according to the attribute information of the conversation object in guiding target and conversation object set, conversation object is determined With the presence or absence of candidate conversation object in set.
In the present embodiment, above-mentioned candidate conversation object, which can be used for characterizing, executes return operation to Matching conversation target Afterwards, conversation object corresponding to the dialogue (that is, next round dialogue) that will be carried out.In the present embodiment, conversation object includes belonging to Property information.Attribute information may include at least one of following: priority is stored in the time of set.Above-mentioned executing subject can lead to Various modes are crossed, according to the attribute information of the conversation object in guiding target and conversation object set, determine conversation object set In with the presence or absence of candidate conversation object.As an example, the rule of FIFO (First-In, First-Out, first in first out) can be used It then determines in conversation object set with the presence or absence of candidate conversation object.Specifically, the time gathered according to the deposit of conversation object, The conversation object for being stored in set at first is determined as candidate conversation object.If there is deposit in set gathers time earliest dialogue Target, then above-mentioned executing subject, which can determine in conversation object set, has candidate conversation object.As another example, it can adopt It is determined in conversation object set with priority rule with the presence or absence of candidate conversation object.Specifically, according to the preferential of conversation object Grade (such as being divided into high, medium and low three ranks from high to low), is determined as candidate dialogue for the conversation object of highest priority Target.If the conversation object of one highest priority of existence anduniquess, above-mentioned executing subject can determine conversation object in set There is candidate conversation object in set.If the conversation object of only one highest priority, above-mentioned execution are not present in set Main body can determine that there is no candidate conversation objects in conversation object set.
In some optional implementations of the present embodiment, attribute information may include priority and deposit set when Between.The time and determine dialogue with the relationship of Matching conversation target that above-mentioned executing subject can be gathered according to priority and deposit With the presence or absence of candidate conversation object in target collection.Specifically, it can carry out in accordance with the following steps:
The conversation object of highest priority in conversation object set is determined as the first candidate conversation object by the first step.
Above-mentioned executing subject by way of traversal, can determine the dialogue mesh of the highest priority in conversation object set Mark.Wherein, the conversation object of highest priority can be 1, be also possible to multiple conversation objects with equal priority.
First candidate conversation object is determined as the in response to determining that the number of the first candidate conversation object is 1 by second step Two candidate conversation objects.
Above-mentioned executing subject can determine the quantity of the above-mentioned first candidate conversation object.In response to determining conversation object set In the conversation object of highest priority there was only 1, above-mentioned executing subject can determine the conversation object of the highest priority For the second candidate conversation object.
Third step, in response to determining that the number of the first candidate conversation object is greater than 1, according to depositing for the first candidate conversation object The the first candidate conversation object for being stored in set at first is determined as the second candidate conversation object by the time for entering set.
In response to determining that it is multiple that the conversation object of the highest priority in conversation object set has, above-mentioned executing subject can be with The the first candidate conversation object for being stored in set at first is determined as the second candidate conversation object.
4th step determines in conversation object set in response to determining that the second candidate conversation object is not Matching conversation target There are candidate conversation objects；And the second candidate conversation object is determined as candidate conversation object.
In practice, if the identified second candidate conversation object is Matching conversation target, conversation object is simultaneously had not been changed, because And above-mentioned executing subject can determine that there is no candidate conversation objects in conversation object set.In response to determining the second candidate dialogue Target is not Matching conversation target, then above-mentioned executing subject, which can determine in conversation object set, has candidate conversation object；With And the second candidate conversation object is determined as candidate conversation object.
Step 406, in response to determining in conversation object set that there is candidate conversation object holds according to candidate conversation object The preset return operation of row.
In the present embodiment, in response to determining that the guiding target being added in conversation object set is not candidate dialogue mesh Mark, it is meant that candidate conversation object was present in corresponding to dialogue in preset conversation object set, carrying out before this Conversation object.In this case, the preset return operation of above-mentioned execution, inside the Pass with the phase of the step 203 in previous embodiment Hold consistent, the preset return operation of execution that the description above with respect to step 203 is also applied in this step, it is no longer superfluous herein It states.
In the present embodiment, in response to determining that the guiding target being added in conversation object set is candidate conversation object, Mean that candidate conversation object was not present in corresponding to dialogue in preset conversation object set, carrying out before this pair Words target, but the dialogue mesh new according to determined by Matching conversation target and pre-set incidence relation as executing subject Mark.In this case, preset reply words art can be used in the preset return operation of above-mentioned execution.For example, " being ordered in completion After the conversation object of air ticket ", as candidate conversation object, then above-mentioned executing subject " can will order hotel " according to above-mentioned steps Can be putd question to user: " needing predetermined hotel? "
In some optional implementations of the present embodiment, the above method further include: by the state of candidate conversation object Information is revised as the ongoing information of dialogue corresponding to characterization conversation object.
In some optional implementations of the present embodiment, the above method further include: be added to dialogue in response to determination Guiding target in target collection is not candidate conversation object, it is meant that above-mentioned guiding target loses associated to language Border, above-mentioned executing subject can delete guiding target from conversation object set.
Figure 4, it is seen that being used to talk with interactive method in the present embodiment compared with the corresponding embodiment of Fig. 2 Process 400 embody acquisition be intended to the step of associated conversation object of word is as guiding target, according to guiding target with The attribute information of conversation object in conversation object set determines in conversation object set with the presence or absence of the step of candidate conversation object Suddenly, and in response to determining in conversation object set there is candidate conversation object, according to candidate conversation object, execute preset time The step of operating again.The scheme of the present embodiment description can determine candidate conversation object and execute preset return operation as a result, The addition of the backtracking of original conversation object and guiding target in conversation object set is realized, to enrich the interior of dialogue interaction Hold.
With further reference to Fig. 5, as the realization to method shown in above-mentioned each figure, this application provides interactive for talking with One embodiment of device, the Installation practice is corresponding with embodiment of the method shown in Fig. 2, which specifically can be applied to In various electronic equipments.
As shown in figure 5, provided in this embodiment include intent determination unit 501, first for talking with interactive device 500 Adding unit 502 and first replys unit 503.Wherein, it is intended that determination unit 501 is configured in response to receive user defeated The sentence entered determines and is intended to word；First adding unit 502 is configured in response to determine in preset conversation object set not In the presence of the conversation object to match with intention word, the conversation object to match with intention word is obtained；By acquired conversation object It is determined as Matching conversation target；And acquired conversation object is added to conversation object set；First replys unit 503, It is configured to execute preset return operation according to Matching conversation target.
In the present embodiment, for talking in interactive device 500: intent determination unit 501, the first adding unit 502 The specific processing for replying unit 503 with first and its brought technical effect can be respectively with reference to the steps in Fig. 2 corresponding embodiment Rapid 201, the related description of step 202 and step 203, details are not described herein.
It is above-mentioned to include: for talking with interactive device 500 in some optional implementations of the present embodiment First determination unit (not shown) may be configured in response to there is and be intended to word phase in determining conversation object set The conversation object to match with intention word is determined as Matching conversation target by the conversation object matched.
In some optional implementations of the present embodiment, the conversation object in above-mentioned conversation object set may include Attribute information, attribute information may include at least one of following: priority is stored in the time of set；And it is above-mentioned for talking with Interactive device 500 can also include: acquiring unit (not shown), the second determination unit (not shown), and second time Multiple unit (not shown).Wherein, acquiring unit may be configured to obtain conversation object conduct associated with word is intended to Guiding target；Second determination unit may be configured to the category according to the conversation object in guiding target and conversation object set Property information, determine in conversation object set with the presence or absence of candidate conversation object；Second reply unit, may be configured in response to Determine in conversation object set that there is candidate conversation object executes preset return operation according to candidate conversation object.
It is above-mentioned to include: for talking with interactive device 500 in some optional implementations of the present embodiment Second adding unit (not shown) may be configured to be not present and guiding target in response to determining in conversation object set Guiding target is added to conversation object set by the conversation object to match.
In some optional implementations of the present embodiment, above-mentioned attribute information may include priority and deposit set Time；And above-mentioned second determination unit may include: the first determining module (not shown), the second determining module (figure In be not shown), third determining module (not shown) and the 4th determining module (not shown).Wherein, first mould is determined Block may be configured to for the conversation object of the highest priority in conversation object set to be determined as the first candidate conversation object； Second determining module may be configured in response to the number for determining the first candidate conversation object be 1, candidate talk with mesh for first Mark is determined as the second candidate conversation object；Third determining module may be configured in response to determining the first candidate conversation object Number be greater than 1, according to the time of the deposit set of the first candidate conversation object, the first candidate dialogue of set will be stored at first Target is determined as the second candidate conversation object；4th determining module may be configured in response to determining the second candidate dialogue mesh Mark is not Matching conversation target, determines in conversation object set there is candidate conversation object；And by the second candidate conversation object It is determined as candidate conversation object.
It is above-mentioned to include: for talking with interactive device 500 in some optional implementations of the present embodiment First deletes unit (not shown), may be configured in response to determining the guiding target being added in conversation object set It is not candidate conversation object, guiding target is deleted from conversation object set.
In some optional implementations of the present embodiment, above-mentioned intent determination unit 501 can be further configured At: above-mentioned input by sentence to intention word disaggregated model trained in advance obtains being intended to word, wherein be intended to word disaggregated model and use In characterization sentence and the corresponding relationship being intended between word.
In some optional implementations of the present embodiment, above-mentioned attribute information may include the time of deposit set； And for talk with interactive device 500 can also include: third determination unit (not shown) and second delete unit (figure In be not shown).Wherein, third determination unit may be configured to determine in conversation object set with the presence or absence of Matching conversation mesh It is marked with other outer conversation objects；Second delete unit, may be configured in response to determine conversation object set in exist With other conversation objects other than conversation object, for other conversation objects other than Matching conversation target, according to the dialogue mesh Target attribute information, determines whether the conversation object is completed；In response to determining that the conversation object is completed, by the conversation object from right It is deleted in words target collection.
In some optional implementations of the present embodiment, above-mentioned attribute information can also include task parameters information, Task parameters information may include task parameters and parameter value；And above-mentioned second deletion unit, it can be further configured to: According to the task parameters and parameter value of the conversation object, determine whether the conversation object is completed.
It can also include: the 4th for talking with interactive device 500 in some optional implementations of the present embodiment Determination unit (not shown) and generation unit (not shown).Wherein, the 4th determination unit may be configured to basis Matching conversation target determines the task parameters of Matching conversation target；Generation unit may be configured to according to above-mentioned sentence and institute The task parameters of determining Matching conversation target generate the parameter value of Matching conversation target.
In some optional implementations of the present embodiment, above-mentioned first reply unit 503 may include: the 5th determining Module (not shown), the first execution module (not shown) and the second execution module (not shown).Wherein, the 5th Determining module may be configured to task parameters and parameter value according to Matching conversation target, whether determine Matching conversation target It completes；First execution module may be configured to execute and Matching conversation target phase in response to determining that Matching conversation target is completed Corresponding predetermined registration operation；Second execution module may be configured in response to determining that Matching conversation target does not complete, according to matching The task parameters and parameter value of conversation object execute default return operation corresponding with Matching conversation target.
In some optional implementations of the present embodiment, above-mentioned attribute information includes triggering mode information and preferential Grade；And for talk with interactive device 500 can also include: modification unit (not shown), may be configured to for Conversation object in conversation object set, in response to determining that the triggering mode information of the conversation object meets preset condition, by this The priority modification of conversation object is higher level.
The device provided by the above embodiment of the application, by intent determination unit 501 in response to receiving user's input Sentence, according to the sentence determine be intended to word.Later, the first adding unit 502 is in response to the preset conversation object set of determination In there is no the conversation objects that match with above-mentioned intention word, obtain the conversation object to match with above-mentioned intention word.Next, Acquired conversation object is determined as Matching conversation target by the first adding unit 502, and acquired conversation object is added Add to above-mentioned conversation object set.Finally, first replys unit 503 according to Matching conversation target, preset return operation is executed. It is engaged in the dialogue interaction to realize according to the conversation object in the conversation object set that can dynamically update.
Below with reference to Fig. 6, it illustrates the computer systems 600 for the electronic equipment for being suitable for being used to realize the embodiment of the present application Structural schematic diagram.Electronic equipment shown in Fig. 6 is only an example, function to the embodiment of the present application and should not use model Shroud carrys out any restrictions.
As shown in fig. 6, computer system 600 includes central processing unit (CPU) 601, it can be read-only according to being stored in Program in memory (ROM) 602 or be loaded into the program in random access storage device (RAM) 603 from storage section 608 and Execute various movements appropriate and processing.In RAM 603, also it is stored with system 600 and operates required various programs and data. CPU 601, ROM 602 and RAM 603 are connected with each other by bus 604.Input/output (I/O) interface 605 is also connected to always Line 604.
I/O interface 605 is connected to lower component: the importation 606 including keyboard, mouse, microphone etc.；Including such as The output par, c 607 of cathode-ray tube (CRT), liquid crystal display (LCD) etc. and loudspeaker etc.；Storage unit including hard disk etc. Divide 608；And the communications portion 609 of the network interface card including LAN card, modem etc..Communications portion 609 via The network of such as internet executes communication process.Driver 610 is also connected to I/O interface 605 as needed.Detachable media 611, such as disk, CD, magneto-optic disk, semiconductor memory etc., are mounted on as needed on driver 610, in order to from The computer program read thereon is mounted into storage section 608 as needed.
Particularly, in accordance with an embodiment of the present disclosure, it may be implemented as computer above with reference to the process of flow chart description Software program.For example, embodiment of the disclosure 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 609, and/or from detachable media 611 are mounted.When the computer program is executed by central processing unit (CPU) 601, limited in execution the present processes Above-mentioned function.
It should be noted that the computer-readable medium of the application can be computer-readable signal media or computer Readable storage medium storing program for executing either the two any combination.Computer readable storage medium for example can be --- but it is unlimited In system, device or the device of --- electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor, or any above combination.It calculates The more specific example of machine readable storage medium storing program for executing can include but is not limited to: have the electrical connection, portable of one or more conducting wires Formula computer disk, hard disk, random access storage device (RAM), read-only memory (ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), light storage device, magnetic memory device or The above-mentioned any appropriate combination of person.In this application, computer readable storage medium can be it is any include or storage program Tangible medium, which can be commanded execution system, device or device use or in connection.And in this Shen Please in, computer-readable signal media may include in a base band or as carrier wave a part propagate data-signal, In carry computer-readable program code.The data-signal of this propagation can take various forms, including but not limited to Electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be computer-readable Any computer-readable medium other than storage medium, the computer-readable medium can send, propagate or transmit for by Instruction execution system, device or device use or program in connection.The journey for including on computer-readable medium Sequence 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.
The calculating of the operation for executing the application can be write with one or more programming languages or combinations thereof Machine program code, described program design language include object-oriented programming language-such as Java, Smalltalk, C+ +, it further include conventional procedural programming language-such as " C " language or similar programming language.Program code can Fully to execute, partly execute on the user computer on the user computer, be executed as an independent software package, Part executes on the remote computer or executes on a remote computer or server completely on the user computer for part. In situations involving remote computers, remote computer can pass through the network of any kind --- including local area network (LAN) Or wide area network (WAN)-is connected to subscriber computer, or, it may be connected to outer computer (such as utilize Internet service Provider is connected by internet).
Flow chart and block diagram in attached drawing are illustrated according to the system of the various embodiments of the application, 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 the module, program segment or code include one or more use The executable instruction of the logic function as defined in realizing.It should also be noted that in some implementations as replacements, being marked in box The function of note can also occur in a different order than that indicated in the drawings.For example, two boxes succeedingly indicated are actually It 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 to infuse Meaning, the combination of each box in block diagram and or flow chart and the box in block diagram and or flow chart can be with holding The dedicated hardware based system of functions or operations as defined in row is realized, or can use specialized hardware and computer instruction Combination realize.
Being described in unit involved in the embodiment of the present application can be realized by way of software, can also be by hard The mode of part is realized.Described unit also can be set in the processor, for example, can be described as: a kind of processor, packet Include intent determination unit, the first adding unit and first replys unit.Wherein, the title of these units is not under certain conditions The restriction to the unit itself is constituted, for example, intent determination unit is also described as " in response to receiving user's input Sentence determines the unit for being intended to word ".
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 said one or multiple programs are held by the electronic equipment When row, so that the electronic equipment: the sentence in response to receiving user's input determines and is intended to word；It is preset right in response to determination There is no the conversation objects to match with the intention word in words target collection, obtain the conversation object to match with the intention word； Acquired conversation object is determined as Matching conversation target；And acquired conversation object is added to the conversation object collection It closes；According to Matching conversation target, preset return operation is executed.
Above description is only the preferred embodiment of the application and the explanation to institute's application technology principle.Those skilled in the art Member is it should be appreciated that invention scope involved in the application, however it is not limited to technology made of the specific combination of above-mentioned technical characteristic Scheme, while should also cover in the case where not departing from foregoing invention design, it is carried out by above-mentioned technical characteristic or its equivalent feature Any combination and the other technical solutions formed.Such as features described above has similar function with (but being not limited to) disclosed herein Can technical characteristic replaced mutually and the technical solution that is formed.

Claims (26)

1. a kind of for talking with interactive method, comprising:
In response to receiving the sentence of user's input, determines and be intended to word；
In response to there is no the conversation object to match with the intention word, acquisition and institutes in the preset conversation object set of determination It states and is intended to the conversation object that word matches；Acquired conversation object is determined as Matching conversation target；And it will be acquired Conversation object is added to the conversation object set；
According to Matching conversation target, preset return operation is executed.
2. according to the method described in claim 1, wherein, before the preset return operation of execution, the method is also wrapped It includes:
It, will be with the intention in response to there is the conversation object to match with the intention word in the determination conversation object set The conversation object that word matches is determined as Matching conversation target.
3. according to the method described in claim 1, wherein, the conversation object in the conversation object set includes attribute information, The attribute information includes at least one of the following: priority, is stored in the time of set；And
The method also includes:
Conversation object associated with the intention word is obtained as guiding target；
According to the attribute information of the conversation object in the guiding target and the conversation object set, the conversation object is determined With the presence or absence of candidate conversation object in set；
In response to there is candidate conversation object in the determination conversation object set, according to the candidate conversation object, execute pre- If return operation.
4. according to the method described in claim 3, wherein, it is candidate right to whether there is in the determination conversation object set Before talking about target, the method also includes:
In response to, there is no the conversation object to match with the guiding target, drawing described in the determination conversation object set It leads target and is added to the conversation object set.
5. according to the method described in claim 4, wherein, the attribute information includes the time of priority and deposit set；With And
With the presence or absence of candidate conversation object in the determination conversation object set, comprising:
The conversation object of highest priority in the conversation object set is determined as the first candidate conversation object；
In response to determining that the number of the first candidate conversation object is 1, the described first candidate conversation object is determined as the second candidate Conversation object；
In response to determining that the number of the first candidate conversation object is greater than 1, according to the deposit set of the first candidate conversation object when Between, the first candidate conversation object for being stored in set at first is determined as the second candidate conversation object；
It is not Matching conversation target in response to the candidate conversation object of determination described second, determines and exist in the conversation object set Candidate conversation object；And the described second candidate conversation object is determined as the candidate conversation object.
6. according to the method described in claim 5, wherein, the method also includes:
In response to determining that the guiding target being added in the conversation object set is not the candidate conversation object, draw described Target is led to delete from the conversation object set.
7. method described in one of -6 according to claim 1, wherein the determining intention word, comprising:
It by the input by sentence to intention word disaggregated model trained in advance, obtains being intended to word, wherein the intention word classification mould Type is used for the corresponding relationship for characterizing sentence and being intended between word.
8. method according to claim 1 or 2, wherein the conversation object in the conversation object set includes attribute letter Breath, the attribute information include the time of deposit set；And
The method also includes:
It determines in the conversation object set with the presence or absence of other conversation objects other than Matching conversation target；
In response to there are other conversation objects other than Matching conversation target in the determination conversation object set, for matching pair Other conversation objects other than words target determine whether the conversation object is completed according to the attribute information of the conversation object；Response In determining conversation object completion, which is deleted from the conversation object set.
9. according to the method described in claim 8, wherein, the attribute information further includes task parameters information, task parameters letter Breath includes task parameters and parameter value；And
Whether the determination conversation object is completed, comprising:
According to the task parameters and parameter value of the conversation object, determine whether the conversation object is completed.
10. according to the method described in claim 9, wherein, the method also includes:
According to Matching conversation target, the task parameters of Matching conversation target are determined；
According to the task parameters of the sentence and identified Matching conversation target, the parameter value of Matching conversation target is generated.
11. it is described that preset return operation is executed according to Matching conversation target according to the method described in claim 10, wherein, Include:
According to the task parameters and parameter value of Matching conversation target, determine whether Matching conversation target is completed；
In response to determining that Matching conversation target is completed, predetermined registration operation corresponding with Matching conversation target is executed；
In response to determining that Matching conversation target does not complete, according to the task parameters and parameter value of Matching conversation target, execute with With the corresponding default return operation of conversation object.
12. the method according to one of claim 3-6, wherein the attribute information includes triggering mode information and preferential Grade；And
The method also includes:
For the conversation object in the conversation object set, in response to determining that it is pre- that the triggering mode information of the conversation object meets It is higher level by the priority modification of the conversation object if condition.
13. a kind of for talking with interactive device, comprising:
Intent determination unit is configured in response to receive the sentence of user's input, determines and be intended to word；
First adding unit, be configured in response to determine in preset conversation object set be not present and the intention word phase The conversation object matched obtains the conversation object to match with the intention word；Acquired conversation object is determined as matching pair Talk about target；And acquired conversation object is added to the conversation object set；
First replys unit, is configured to execute preset return operation according to Matching conversation target.
14. device according to claim 13, wherein described device further include:
First determination unit is configured in response to determine in the conversation object set and is intended to what word matched in the presence of with described The conversation object to match with the intention word is determined as Matching conversation target by conversation object.
15. device according to claim 13, wherein the conversation object in the conversation object set includes attribute letter Breath, the attribute information include at least one of the following: priority, are stored in the time of set；And
Described device further include:
Acquiring unit is configured to obtain conversation object associated with the intention word as guiding target；
Second determination unit is configured to the attribute according to the conversation object in the guiding target and the conversation object set Information determines in the conversation object set with the presence or absence of candidate conversation object；
Second replys unit, is configured in response to determine in the conversation object set there is candidate conversation object, according to institute Candidate conversation object is stated, preset return operation is executed.
16. device according to claim 15, wherein described device further include:
Second adding unit, be configured in response to determine in the conversation object set be not present and the guiding target phase The guiding target is added to the conversation object set by the conversation object matched.
17. device according to claim 16, wherein the attribute information includes the time of priority and deposit set； And
Second determination unit includes:
First determining module is configured to the conversation object of the highest priority in the conversation object set being determined as first Candidate conversation object；
Second determining module is configured in response to determine that the number of the first candidate conversation object is 1, candidate right by described first Words target is determined as the second candidate conversation object；
Third determining module is configured in response to determine that the number of the first candidate conversation object is greater than 1, candidate right according to first The the first candidate conversation object for being stored in set at first is determined as the second candidate dialogue mesh by the time for talking about the deposit set of target Mark；
4th determining module is configured in response to determine that the described second candidate conversation object is not Matching conversation target, determine There is candidate conversation object in the conversation object set；And that the described second candidate conversation object is determined as the candidate is right Talk about target.
18. device according to claim 17, wherein described device further include:
First deletes unit, is configured in response to determine that the guiding target being added in the conversation object set is not described Candidate conversation object deletes the guiding target from the conversation object set.
19. device described in one of 3-18 according to claim 1, wherein the intent determination unit is further configured to:
It by the input by sentence to intention word disaggregated model trained in advance, obtains being intended to word, wherein the intention word classification mould Type is used for the corresponding relationship for characterizing sentence and being intended between word.
20. device described in 3 or 14 according to claim 1, wherein the conversation object in the conversation object set includes attribute Information, the attribute information include the time of deposit set；And
Described device further include:
Third determination unit is configured to determine in the conversation object set with the presence or absence of other other than Matching conversation target Conversation object；
Second deletes unit, is configured in response to determine in the conversation object set that there are its other than Matching conversation target His conversation object, for other conversation objects other than Matching conversation target, according to the attribute information of the conversation object, determining should Whether conversation object is completed；In response to determining that the conversation object is completed, which is deleted from the conversation object set It removes.
21. device according to claim 20, wherein the attribute information further includes task parameters information, task parameters Information includes task parameters and parameter value；And
Described second deletes unit, is further configured to:
According to the task parameters and parameter value of the conversation object, determine whether the conversation object is completed.
22. device according to claim 21, wherein described device further include:
4th determination unit is configured to determine the task parameters of Matching conversation target according to Matching conversation target；
Generation unit is configured to the task parameters according to the sentence and identified Matching conversation target, generates matching pair Talk about the parameter value of target.
23. device according to claim 22, wherein described first, which replys unit, includes:
5th determining module is configured to task parameters and parameter value according to Matching conversation target, determines Matching conversation target Whether complete；
First execution module is configured in response to determine that Matching conversation target is completed, execute corresponding with Matching conversation target Predetermined registration operation；
Second execution module is configured in response to determine that Matching conversation target does not complete, according to the task of Matching conversation target Parameter and parameter value execute default return operation corresponding with Matching conversation target.
24. device described in one of 5-18 according to claim 1, wherein the attribute information includes triggering mode information and excellent First grade；And
Described device further include:
Unit is modified, is configured to for the conversation object in the conversation object set, in response to determining the conversation object Triggering mode information meets preset condition, is higher level by the priority modification of the conversation object.
25. a kind of electronic equipment, comprising:
One or more processors；
Storage device is stored thereon with one or more programs；
When one or more of programs are executed by one or more of processors, so that one or more of processors are real The now method as described in any in claim 1-12.
26. a kind of computer-readable medium, is stored thereon with computer program, wherein the realization when program is executed by processor Method as described in any in claim 1-12.
CN201811146511.9A 2018-09-29 2018-09-29 Method and device for dialogue interaction Active CN109145104B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811146511.9A CN109145104B (en) 2018-09-29 2018-09-29 Method and device for dialogue interaction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811146511.9A CN109145104B (en) 2018-09-29 2018-09-29 Method and device for dialogue interaction

Publications (2)

Publication Number Publication Date
CN109145104A true CN109145104A (en) 2019-01-04
CN109145104B CN109145104B (en) 2023-04-18

Family

ID=64813351

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811146511.9A Active CN109145104B (en) 2018-09-29 2018-09-29 Method and device for dialogue interaction

Country Status (1)

Country Link
CN (1) CN109145104B (en)

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109514586A (en) * 2019-01-30 2019-03-26 第四范式（北京）技术有限公司 Realize the method and system of intelligent customer service robot
CN110232108A (en) * 2019-05-13 2019-09-13 华为技术有限公司 Interactive method and conversational system
CN110543556A (en) * 2019-09-05 2019-12-06 出门问问信息科技有限公司 Dialogue configuration method, storage medium and electronic equipment
CN111143524A (en) * 2019-12-06 2020-05-12 联想(北京)有限公司 User intention determining method and electronic equipment
CN111143523A (en) * 2019-12-02 2020-05-12 北京声智科技有限公司 Intention confirming method and device
CN111831795A (en) * 2019-04-11 2020-10-27 北京猎户星空科技有限公司 Multi-turn conversation processing method and device, electronic equipment and storage medium
CN111966805A (en) * 2020-08-13 2020-11-20 贝壳技术有限公司 Method, device, medium and electronic equipment for assisting in realizing session
CN112035630A (en) * 2020-03-27 2020-12-04 北京来也网络科技有限公司 Dialogue interaction method, device, equipment and storage medium combining RPA and AI
CN112256846A (en) * 2020-09-28 2021-01-22 南方电网深圳数字电网研究院有限公司 Man-machine conversation interaction method and system
CN113127608A (en) * 2019-12-31 2021-07-16 微软技术许可有限责任公司 Plan-guided response provision
CN113132214A (en) * 2019-12-31 2021-07-16 深圳市优必选科技股份有限公司 Conversation method, device, server and storage medium
CN111143523B (en) * 2019-12-02 2024-05-03 北京声智科技有限公司 Intention confirming method and device

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017041372A1 (en) * 2015-09-07 2017-03-16 百度在线网络技术（北京）有限公司 Man-machine interaction method and system based on artificial intelligence
CN106951491A (en) * 2017-03-14 2017-07-14 广东工业大学 A kind of Intelligent dialogue control method and device applied to robot
CN107870994A (en) * 2017-10-31 2018-04-03 北京光年无限科技有限公司 Man-machine interaction method and system for intelligent robot
CN108427722A (en) * 2018-02-09 2018-08-21 卫盈联信息技术（深圳）有限公司 intelligent interactive method, electronic device and storage medium
WO2018149326A1 (en) * 2017-02-16 2018-08-23 阿里巴巴集团控股有限公司 Natural language question answering method and apparatus, and server

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017041372A1 (en) * 2015-09-07 2017-03-16 百度在线网络技术（北京）有限公司 Man-machine interaction method and system based on artificial intelligence
WO2018149326A1 (en) * 2017-02-16 2018-08-23 阿里巴巴集团控股有限公司 Natural language question answering method and apparatus, and server
CN108446286A (en) * 2017-02-16 2018-08-24 阿里巴巴集团控股有限公司 A kind of generation method, device and the server of the answer of natural language question sentence
CN106951491A (en) * 2017-03-14 2017-07-14 广东工业大学 A kind of Intelligent dialogue control method and device applied to robot
CN107870994A (en) * 2017-10-31 2018-04-03 北京光年无限科技有限公司 Man-machine interaction method and system for intelligent robot
CN108427722A (en) * 2018-02-09 2018-08-21 卫盈联信息技术（深圳）有限公司 intelligent interactive method, electronic device and storage medium

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109514586A (en) * 2019-01-30 2019-03-26 第四范式（北京）技术有限公司 Realize the method and system of intelligent customer service robot
CN111831795A (en) * 2019-04-11 2020-10-27 北京猎户星空科技有限公司 Multi-turn conversation processing method and device, electronic equipment and storage medium
CN111831795B (en) * 2019-04-11 2023-10-27 北京猎户星空科技有限公司 Multi-round dialogue processing method and device, electronic equipment and storage medium
CN110232108A (en) * 2019-05-13 2019-09-13 华为技术有限公司 Interactive method and conversational system
CN110232108B (en) * 2019-05-13 2023-02-03 华为技术有限公司 Man-machine conversation method and conversation system
CN110543556A (en) * 2019-09-05 2019-12-06 出门问问信息科技有限公司 Dialogue configuration method, storage medium and electronic equipment
CN111143523A (en) * 2019-12-02 2020-05-12 北京声智科技有限公司 Intention confirming method and device
CN111143523B (en) * 2019-12-02 2024-05-03 北京声智科技有限公司 Intention confirming method and device
CN111143524A (en) * 2019-12-06 2020-05-12 联想(北京)有限公司 User intention determining method and electronic equipment
CN113127608A (en) * 2019-12-31 2021-07-16 微软技术许可有限责任公司 Plan-guided response provision
CN113132214A (en) * 2019-12-31 2021-07-16 深圳市优必选科技股份有限公司 Conversation method, device, server and storage medium
CN113132214B (en) * 2019-12-31 2023-07-18 深圳市优必选科技股份有限公司 Dialogue method, dialogue device, dialogue server and dialogue storage medium
CN112035630A (en) * 2020-03-27 2020-12-04 北京来也网络科技有限公司 Dialogue interaction method, device, equipment and storage medium combining RPA and AI
CN111966805A (en) * 2020-08-13 2020-11-20 贝壳技术有限公司 Method, device, medium and electronic equipment for assisting in realizing session
CN112256846A (en) * 2020-09-28 2021-01-22 南方电网深圳数字电网研究院有限公司 Man-machine conversation interaction method and system

Also Published As

Publication number Publication date
CN109145104B (en) 2023-04-18

Similar Documents

Publication Publication Date Title
CN109145104A (en) For talking with interactive method and apparatus
US11295221B2 (en) Learning user preferences in a conversational system
US10853582B2 (en) Conversational agent
US11394667B2 (en) Chatbot skills systems and methods
CN104584010B (en) Voice-based media research
CN105654950B (en) Adaptive voice feedback method and device
EP2987164B1 (en) Virtual assistant focused user interfaces
US20170277993A1 (en) Virtual assistant escalation
CN107113222B (en) Active chat information system based on environment
US20140012574A1 (en) Interactive timeline for presenting and organizing tasks
JP2019508820A (en) Automatic suggestions for message exchange threads
US20160027440A1 (en) Selective speech recognition for chat and digital personal assistant systems
CN107004410A (en) Voice and connecting platform
CN105446994A (en) Service recommendation method and device with intelligent assistant
CN107562723A (en) Meeting processing method, medium, device and computing device
CN103970791B (en) A kind of method, apparatus for recommending video from video library
WO2013059726A1 (en) Systems, devices and methods for list display and management
CN110022258A (en) A kind of conversation controlling method and device, electronic equipment of instant messaging
CN110990598B (en) Resource retrieval method and device, electronic equipment and computer-readable storage medium
CN109325121A (en) Method and apparatus for determining the keyword of text
CN109739965A (en) Moving method and device, equipment, the readable storage medium storing program for executing of cross-cutting dialog strategy
CN109873756A (en) Method and apparatus for sending information
CN109313649A (en) Voice-based knowledge sharing application for chat robots
KR20190076870A (en) Device and method for recommeding contact information
CN110275691A (en) Automatic reply method, device, terminal and the storage medium that intelligent sound wakes up

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