CN109190114A - Method and apparatus for generating return information - Google Patents

Method and apparatus for generating return information Download PDF

Info

Publication number
CN109190114A
CN109190114A CN201810917669.5A CN201810917669A CN109190114A CN 109190114 A CN109190114 A CN 109190114A CN 201810917669 A CN201810917669 A CN 201810917669A CN 109190114 A CN109190114 A CN 109190114A
Authority
CN
China
Prior art keywords
information
mode
node
target
mentioned
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
CN201810917669.5A
Other languages
Chinese (zh)
Other versions
CN109190114B (en
Inventor
吴金霖
孟思琪
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 CN201810917669.5A priority Critical patent/CN109190114B/en
Publication of CN109190114A publication Critical patent/CN109190114A/en
Application granted granted Critical
Publication of CN109190114B publication Critical patent/CN109190114B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00Handling natural language data
    • G06F40/30Semantic analysis

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Machine Translation (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The embodiment of the present application discloses the method and apparatus for generating return information.One specific embodiment of this method includes: the input information for receiving user and being sent by terminal;Semantic analysis is carried out to input information, obtains the semantic analysis result of input information, wherein semantic analysis result includes target intention classification and confidence level corresponding with target intention classification;Will input information and the question and answer pre-established to the question and answer in information aggregate to information the problem of information match, obtain matching result, wherein question and answer include problem information and answer information to information;Based on semantic analysis result and matching result, target retro information generating mode is determined from preset at least one return information generating mode;Return information is generated according to target retro information generating mode, and return information is sent to terminal.The embodiment makes the return information generated more have specific aim.

Description

Method and apparatus for generating return information
Technical field
The invention relates to field of computer technology, and in particular to the method and apparatus for generating return information.
Background technique
With the development of human-computer interaction technology, the human-computer interaction technology based on language is also widely applied, for example, can To provide service by the artificial user of the intelligent interactions machines such as intelligent online customer service robot, mobile phone assistant.In general, user is chatting Skylight opening inputs after information, and intelligent interaction robot can chatted according to certain logic automatically based on the user's input information Skylight opening provides relevant sentence.
It in practice, for different business scene, needs that different reply logics is arranged for intelligent interaction robot, to meet Demand for services of the user under different business scene.
Summary of the invention
The embodiment of the present application proposes the method and apparatus for generating return information.
In a first aspect, the embodiment of the present application provides a kind of method for generating return information, this method comprises: receiving The input information that user is sent by terminal;Semantic analysis is carried out to above-mentioned input information, obtains the semanteme of above-mentioned input information Analyze result, wherein above-mentioned semantic analysis result includes target intention classification and confidence corresponding with above-mentioned target intention classification Degree;By above-mentioned input information and the question and answer pre-established to the question and answer in information aggregate to information the problem of information match, Obtain matching result, wherein question and answer include problem information and answer information to information;Based on above-mentioned semantic analysis result and above-mentioned Matching result determines target retro information generating mode from preset at least one return information generating mode;It presses Return information is generated according to above-mentioned target retro information generating mode, and above-mentioned return information is sent to above-mentioned terminal.
In some embodiments, above-mentioned at least one return information generating mode includes: based on above-mentioned target intention classification Corresponding root node and digraph generate the mode one of return information, is intended to the corresponding root node of classification based on history and above-mentioned has The mode two of return information is generated to figure, the mode three of return information, base are generated based on historical action node and above-mentioned digraph In mode four of the above-mentioned question and answer to information aggregate and the mode five based on presupposed information set, wherein above-mentioned digraph includes root Node judges node, movement node and expanding node, wherein root node and intention classification correspond, and judge in node in advance Definition has judgment rule, and acting to pre-define in node has return information create-rule, and number of request is pre-defined in expanding node According to rule.
In some embodiments, when above-mentioned target retro information generating mode be mode for the moment, it is above-mentioned according to above-mentioned target Return information generating mode generates return information, and above-mentioned return information is sent to above-mentioned terminal, comprising: above-mentioned target is anticipated The corresponding root node of figure classification is determined as first object root node;It is generated according to above-mentioned first object root node and above-mentioned digraph Return information, and above-mentioned return information is sent to above-mentioned terminal.
In some embodiments, above-mentioned according to above-mentioned target when above-mentioned target retro information generating mode is mode two Return information generating mode generates return information, and above-mentioned return information is sent to above-mentioned terminal, comprising: from dialog information collection The intention classification that a upper sentence for input information that above-mentioned terminal is sent, above-mentioned is obtained in conjunction is intended to classification as history, wherein Above-mentioned dialog information set is for recording and information caused by above-mentioned terminal interaction;Above-mentioned history is intended to classification corresponding Node is determined as the second target root node;Return information is generated according to above-mentioned second target root node and above-mentioned digraph, and Above-mentioned return information is sent to above-mentioned terminal.
In some embodiments, above-mentioned according to above-mentioned target when above-mentioned target retro information generating mode is mode three Return information generating mode generates return information, and above-mentioned return information is sent to above-mentioned terminal, comprising: according to dialog information The information recorded in set determines the corresponding movement node conduct of a upper sentence of input information that above-mentioned terminal is sent, above-mentioned Historical action node has next node in response to the above-mentioned historical action node of determination, by the next of above-mentioned historical action node Node is as destination node, wherein above-mentioned dialog information set is for recording and information caused by above-mentioned terminal interaction;According to Above-mentioned destination node and above-mentioned digraph generate return information, and above-mentioned return information is sent to above-mentioned terminal.
In some embodiments, above-mentioned according to above-mentioned target when above-mentioned target retro information generating mode is mode four Return information generating mode generates return information, and above-mentioned return information is sent to above-mentioned terminal, comprising: according to above-mentioned matching As a result from above-mentioned question and answer to choosing target question and answer in information aggregate to information, and by target question and answer to the answer information in information Above-mentioned terminal is sent to as return information.
In some embodiments, it is above-mentioned be based on above-mentioned semantic analysis result and above-mentioned matching result, from it is preset to Target retro information generating mode is determined in a kind of few return information generating mode, comprising: according to above-mentioned target intention classification Confidence level, above-mentioned matching result and preset scheduling rule determine aforesaid way one, mode two, mode three, mode One of four and mode five are used as target retro information generating mode.
Second aspect, the embodiment of the present application provide a kind of for generating the device of return information, and above-mentioned apparatus includes: to connect Unit is received, is configured to receive the input information that user is sent by terminal;Analytical unit is configured to above-mentioned input information Semantic analysis is carried out, obtains the semantic analysis result of above-mentioned input information, wherein above-mentioned semantic analysis result includes target intention Classification and confidence level corresponding with above-mentioned target intention classification;Matching unit, be configured to by above-mentioned input information with build in advance Vertical question and answer to the question and answer in information aggregate to information the problem of information match, obtain matching result, wherein question and answer are to letter Breath includes problem information and answer information;Determination unit is configured to based on above-mentioned semantic analysis result and above-mentioned matching result, Target retro information generating mode is determined from preset at least one return information generating mode;Generation unit, quilt It is configured to generate return information according to above-mentioned target retro information generating mode, and above-mentioned return information is sent to above-mentioned end End.
In some embodiments, above-mentioned at least one return information generating mode includes: based on above-mentioned target intention classification Corresponding root node and digraph generate the mode one of return information, is intended to the corresponding root node of classification based on history and above-mentioned has The mode two of return information is generated to figure, the mode three of return information, base are generated based on historical action node and above-mentioned digraph In mode four of the above-mentioned question and answer to information aggregate and the mode five based on presupposed information set, wherein above-mentioned digraph includes root Node judges node, movement node and expanding node, wherein root node and intention classification correspond, and judge in node in advance Definition has judgment rule, and acting to pre-define in node has return information create-rule, and number of request is pre-defined in expanding node According to rule.
In some embodiments, when above-mentioned target retro information generating mode be mode for the moment, above-mentioned generation unit is into one Step is configured to: the corresponding root node of above-mentioned target intention classification is determined as first object root node;According to above-mentioned first mesh It marks root node and above-mentioned digraph generates return information, and above-mentioned return information is sent to above-mentioned terminal.
In some embodiments, when above-mentioned target retro information generating mode is mode two, above-mentioned generation unit is into one Step is configured to: the intention class of a upper sentence for input information that above-mentioned terminal is sent, above-mentioned is obtained from dialog information set Not Zuo Wei history be intended to classification, wherein above-mentioned dialog information set for record with above-mentioned terminal interaction caused by information;It will Above-mentioned history is intended to the corresponding root node of classification and is determined as the second target root node;According to above-mentioned second target root node and above-mentioned Digraph generates return information, and above-mentioned return information is sent to above-mentioned terminal.
In some embodiments, when above-mentioned target retro information generating mode is mode three, above-mentioned generation unit is into one Step is configured to: according to the information recorded in dialog information set, determining upper the one of input information that above-mentioned terminal is sent, above-mentioned The corresponding movement node of sentence has next node as historical action node, in response to the above-mentioned historical action node of determination, will be upper The next node of historical action node is stated as destination node, wherein above-mentioned dialog information set is for recording and above-mentioned end Information caused by the interaction of end;Return information is generated according to above-mentioned destination node and above-mentioned digraph, and above-mentioned reply is believed Breath is sent to above-mentioned terminal.
In some embodiments, when above-mentioned target retro information generating mode is mode four, above-mentioned generation unit is into one Step is configured to: according to above-mentioned matching result from above-mentioned question and answer to choosing target question and answer in information aggregate to information, and by mesh Mark question and answer are sent to above-mentioned terminal as return information to the answer information in information.
In some embodiments, above-mentioned determination unit is further configured to: according to the confidence of above-mentioned target intention classification Degree, above-mentioned matching result and preset scheduling rule determine aforesaid way one, mode two, mode three, mode four and side One of formula five is used as target retro information generating mode.
The third aspect, the embodiment of the present application provide a kind of server, which includes: one or more processors; Storage device is stored thereon with one or more programs, when said one or multiple programs are by said one or multiple processors When execution, so that said one or multiple 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, In, the method as described in implementation any in first aspect is realized when which is executed by processor.
Method and apparatus provided by the embodiments of the present application for generating return information first believe the input that terminal is sent Breath carries out semantic analysis and obtains the semantic analysis result of input information, will then input information and the question and answer pre-established to information The problem of question and answer in set are to information information is matched to obtain matching result, is then based on semantic analysis result and matching knot Fruit determines target retro information generating mode from least one return information generating mode, finally according to target retro information Generating mode generates return information and return information is sent to terminal, realizes complete based on semantic analysis result and matching result At the selection of target retro information generating mode, to make the return information generated that more there is specific aim, it is more able to satisfy user Demand.
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 of the method for generating return information of the application;
Fig. 3 is the exemplary diagram of the node and line according to the digraph of the application;
Fig. 4 is the exemplary diagram according to a kind of scheduling rule of the application;
Fig. 5 is the schematic diagram according to an application scenarios of the method for generating return information of the application;
Fig. 6 is the structural schematic diagram according to one embodiment of the device for generating return information of the application;
Fig. 7 is adapted for the structural schematic diagram for the computer system for realizing the server 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, which is shown, can reply letter using the method for being used to generate return information of the embodiment of the present application or for generating The exemplary system architecture 100 of the device of breath.
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..
User can be used terminal device 101,102,103 and be interacted by network 104 with server 105, to receive or send out Send message etc..Various telecommunication customer end applications can be installed, such as web browser is answered on terminal device 101,102,103 With, shopping class application, searching class application, instant messaging tools, mailbox client, social platform software 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 with display screen and supported web page browsing, including but not limited to smart phone, plate Computer, E-book reader, pocket computer on knee and desktop computer etc..When terminal device 101,102,103 is soft When part, it may be mounted in above-mentioned cited electronic equipment.Its may be implemented into multiple softwares or software module (such as Distributed Services are provided), 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, such as to terminal device 101,102,103 received times Complex information provides the background server supported.The input information that background server can send terminal device 101,102,103 It carries out semantic analysis the problems in information information is carried out with question and answer the processing such as to match, and is generated according to processing result and reply letter Breath is sent to terminal device.
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 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.
It should be noted that for generating the method for return information generally by server provided by the embodiment of the present application 105 execute, and correspondingly, the device for generating return information is generally positioned in server 105.
With continued reference to Fig. 2, the stream of one embodiment of the method for generating return information according to the application is shown Journey 200.The method for being used to generate return information, comprising the following steps:
Step 201, the input information that user is sent by terminal is received.
In the present embodiment, for generating the executing subject (such as server 105 shown in FIG. 1) of the method for return information It can carry out terminal (such as Fig. 1 institute of information input using it from user by wired connection mode or radio connection The terminal device 101 that shows, 102,103) receive input information.In practice, input information can include but is not limited to text information, Voice messaging, pictorial information etc..Input information include pictorial information in the case where, above-mentioned executing subject can to picture into Row semantics recognition generates text information relevant to picture;In the case where inputting information includes voice messaging, above-mentioned execution master Body can carry out speech recognition to voice messaging, generate text information relevant to voice messaging.
Step 202, semantic analysis is carried out to input information, obtains the semantic analysis result of input information.
In the present embodiment, above-mentioned executing subject can using various semantic analysis means (for example, participle, part-of-speech tagging, Name Entity recognition etc.) input information obtained in step 201 is analyzed, to obtain the semantic analysis knot of input information Fruit.Herein, above-mentioned semantic analysis result may include target intention classification and confidence level corresponding with target intention classification.Its In, target intention classification can be used to indicate that above-mentioned user sends the classification of the intention of above-mentioned input information.In practice, Ke Yigen Different intention classifications is preset according to different business scenarios, i.e., a kind of corresponding intention classification of business scenario, for example, working as Above-mentioned executing subject is for can be set intention classification for business scenario " telephone expenses inquiry " when providing customer service for operator For " telephone expenses inquiry ", it can be set for business scenario " telephone recharge " and be intended to classification as " charge filling " etc..
As an example, above-mentioned executing subject can first segment above-mentioned input information, obtain target participle Set.Then, target intention classification and confidence corresponding with target intention classification are obtained by the directly matched mode of vocabulary Degree, wherein above-mentioned vocabulary can be technical staff and be pre-established based on the statistics to a large amount of participle set and intention classification , be stored with it is multiple participle set be intended to classification corresponding relationships mapping tables.In this way, above-mentioned executing subject can incite somebody to action Target participle set is successively matched with multiple participle set in the mapping table, obtains the correspondence according to matching result Gather with target participle the corresponding intention classification of most like participle set in relation table as target intention classification, and according to this Participle set determines the corresponding confidence level of target intention classification with the similarity that target participle is gathered.
As another example, above-mentioned input information can be imported the intent classifier mould pre-established by above-mentioned executing subject Type obtains the input corresponding target intention classification of information and confidence level corresponding with target intention classification.Wherein, above-mentioned intention point Class model can be used for characterizing input information and be intended to classification and be intended to the corresponding relationship of the confidence level of classification.Above-mentioned intention point Class model can be to be obtained based on machine learning method, specifically, above-mentioned intent classifier model can be based on naive Bayesian Model (Naive Bayesian Model, NBM) or support vector machines (Support Vector Machine, SVM) etc. are used for What the model training of classification obtained.
Step 203, will input information with the question and answer that pre-establish to the question and answer in information aggregate to information the problem of information It is matched, obtains matching result.
In the present embodiment, question and answer can be previously stored in above-mentioned executing subject to information aggregate, wherein above-mentioned question and answer To each question and answer in information aggregate to information may include associated storage the problem of information and answer information.In this way, above-mentioned The problem of question and answer that executing subject can concentrate the input information received in step 201 and question and answer to information are to information information It is matched, obtains matching result (for example, similarity).
Step 204, it is based on semantic analysis result and matching result, from preset at least one return information generation side Target retro information generating mode is determined in formula.
In the present embodiment, it can be preset in above-mentioned executing subject at least one for generating returning for return information Complex information generating mode.Above-mentioned executing subject can be based on above-mentioned semantic analysis result and above-mentioned matching result from above-mentioned at least one Target retro information generating mode is determined in kind return information generating mode.As an example, can be pre- in above-mentioned executing subject It is first stored with scheduling rule, which can be used for characterizing semantic analysis result, matching result and return information generation side Corresponding relationship between formula.In this way, above-mentioned executing subject after obtaining semantic analysis result and matching result, is somebody's turn to do by inquiry Scheduling rule can determine target retro information generating mode.
In some optional implementations of the present embodiment, above-mentioned at least one return information generating mode may include It is several below:
1) mode one of return information is generated based on the corresponding root node of above-mentioned target intention classification and digraph.
Herein, digraph, which can be, is prestored to above-mentioned executing subject, and digraph may include root node, judgement Node, movement node and expanding node.
Root node can be corresponded with classification is intended to, i.e. the entrance section that a root node can be used as a business scenario Point.
Judge that judgment rule can have been pre-defined in node, in practice, judge that node can set multiple judgement branches, Each judges that branch can be for extracted entity in intention classification, user's input information, interactive process and based on expansion The data such as the exhibition requested result of node are judged.As an example, judging that judgement operator used in node can wrap Include but be not limited to==,!=,>,<,>=,≤, is null, is not null, contains (including), regex (canonical Expression formula class) etc..In addition, judging also to support combination judgement, such as AND or OR in node.In practice, the more of node are judged It may include that an else (other) judges branch in a judgement branch, else judgement branches into all Rule of judgment and is all unsatisfactory for When a branch.
Return information create-rule can have been pre-defined in movement node, be arranged in movement node and be one of the following: etc. To user's input, turn artificial and session termination.For example, the return information that some movement node generates is used to put question to user, Then " waiting user's input " is set by the movement node.In another example the return information that some movement node generates is for prompting User turns manual service, then sets " turning artificial " for the movement node.For another example the return information that some movement node generates For prompting user conversation to terminate, then the movement node is set as " session termination ".
The rule of request data can be pre-defined in expanding node.Expanding node can take in interactive process to outside Be engaged in device initiate http request, and can by the information generated in interactive process (such as, it is intended that classification, entity etc.) be used as http Parameter, the value that http request returns can be used for generating return information for movement node.
In digraph, root node judges that there may be connection relationships between node, movement node and expanding node.Its In, root node may include one outlet point.Judge that node may include an entrance and at least one exit point, wherein The quantity of exit point is identical as the judgement numbers of branches that the judgement node is arranged.Acting node may include an entrance and one A exit point.Expanding node may include an entrance and one outlet point.Herein, exit point is the starting point of line, i.e., The starting point of the oriented line of next node is connected to from a node.Entrance is the terminal of line, i.e., from a node It is connected to the direction point of the oriented line of next node.Wherein, exit point cannot be connected into line, and entrance cannot connect Line, exit point can only connect single line, and entrance can connect more lines.As an example, a node for digraph and showing for line Example diagram can judge node 1,2,3 with as shown in figure 3, including root node 1,2 in Fig. 3, act node 1,2,3,4,5 and extension section Point 1, wherein "○" indicates entrance, and "●" indicates exit point.It should be pointed out that the quantity of various nodes, judgement in Fig. 3 Exit point quantity in node etc. is only schematical, rather than to the quantity of various nodes, judge outlet points in node The restriction of amount.In actual use, it can be configured according to the needs of practical business scene.
2) the corresponding root node of classification is intended to based on history and above-mentioned digraph generates the mode two of return information.At this In, if user is transmitted across sentence before sending above-mentioned input information, above-mentioned executing subject can send user above-mentioned defeated Enter the corresponding intention classification of a upper sentence sent before information and is intended to classification as history.
3) mode three of return information is generated based on historical action node and above-mentioned digraph.Herein, if user sends out It serves and is transmitted across sentence before stating input information, then user can be sent and be sent between above-mentioned input information by above-mentioned executing subject The corresponding movement node of a upper sentence as historical action node, the corresponding movement node of a upper sentence can be pointer The movement node of return information is generated to a upper sentence.
4) based on above-mentioned question and answer to the mode four of information aggregate.
5) based on the mode five of presupposed information set.Herein, it can be preset in above-mentioned presupposed information set useful In indicate do not understand user be intended to if art, for example, " so sorry, I does not understand what you say? ".
In some optional implementations, above-mentioned steps 204 specific as follows can be carried out: according to above-mentioned target intention class Other confidence level, above-mentioned matching result and preset scheduling rule determine aforesaid way one, mode two, mode three, side One of formula four and mode five are used as target retro information generating mode.Herein, above-mentioned scheduling rule can be by technical staff's root It is set according to actual business demand, herein without limitation.
Fig. 4 is referred to, as an example, a kind of scheduling rule can be as follows:
Step S1, judge the confidence level of target intention classification whether be greater than preset threshold and target intention classification have it is corresponding Root node.
Step S2, in response to determining that the confidence level of target intention classification is greater than preset threshold value and target intention classification There is corresponding root node, sets mode one for return information generating mode.
Step S3 determines the corresponding historical action node of a upper sentence that user inputs before inputting above-mentioned input information Whether user is required to input.
In response to the above-mentioned historical action node requirements user input of determination, step S4 is executed.
Step S4, determines whether above-mentioned historical action node has subsequent node.
Step S5 has subsequent node in response to the above-mentioned historical action node of determination, return information generating mode is arranged For mode three.
Step S6 does not have subsequent node in response to the above-mentioned historical action node of determination, and determines that return information generates Mode is vacancy, sets mode two for return information generating mode.
It does not require user to input in response to the above-mentioned historical action node of determination, and determines that return information generating mode is sky It lacks, executes step S7.
Step S7, determines whether above-mentioned question and answer have the question and answer to match to information information aggregate.
Step S8 has the question and answer information to match in response to determination, sets mode four for return information generating mode.
Step S9 sets mode five for return information generating mode in response to determining the question and answer information not matched.
It is not more than preset threshold value and/or target intention class in response to the confidence level of the above-mentioned target intention classification of determination Not without corresponding root node, step S10 is executed.
Step S10 determines that user inputs whether information is initial dialog.If so, thening follow the steps S7;If it is not, Then follow the steps S3.
Step 205, return information is generated according to target retro information generating mode, and return information is sent to end End.
In the present embodiment, above-mentioned executing subject can be according to the target retro information generating mode determined in step 204 Return information is generated, and return information generated is sent to above-mentioned terminal, so that above-mentioned terminal is presented to the user.
In some optional implementations, when above-mentioned target retro information generating mode is mode a period of time, above-mentioned steps 205 can specifically include the following contents: firstly, above-mentioned executing subject can be true by the corresponding root node of above-mentioned target intention classification It is set to first object root node.Later, above-mentioned executing subject can be raw according to above-mentioned first object root node and above-mentioned digraph It is sent to above-mentioned terminal at return information, and by above-mentioned return information, so that above-mentioned terminal is presented.On as an example, Stating executing subject can be executed downwards using first object root node as starting by the line relationship between digraph interior joint, Until generating return information.
In some optional implementations, when above-mentioned target retro information generating mode is mode two, above-mentioned steps 205 can specifically include the following contents: firstly, obtaining input information that above-mentioned terminal is sent, above-mentioned from dialog information set A upper sentence intention classification as history be intended to classification.Herein, above-mentioned dialog information set can be used for record with it is upper State information caused by terminal interaction.Take turns dialogue with executing subject as an example, can store terminal in dialog information set more The various information generated in the process, for example, extracted entity, user input a upper sentence intention classification, be directed to upper one The movement node etc. of sentence generation return information.In practice, when terminal sends an information to executing subject, executing subject It judges, whether includes Session ID (sessionID) in the information, if not including, executing subject can be directed to the letter Breath generates a Session ID, and the various information that the conversation procedure is generated and the Session ID associated storage to dialogue Information aggregate;If including, and included Session ID is not expired, then it can be directly using corresponding to the Session ID Dialog information set, for example, storage information, recalls information etc..Secondly, above-mentioned history can be intended to by above-mentioned executing subject The corresponding root node of classification is determined as the second target root node.Finally, above-mentioned executing subject can be according to above-mentioned second target root Node and above-mentioned digraph generate return information, and above-mentioned return information is sent to above-mentioned terminal, so that terminal is presented to User.
In some optional implementations, when above-mentioned target retro information generating mode is mode three, above-mentioned steps 205 can specifically include the following contents: firstly, above-mentioned executing subject can be according to the letter recorded in above-mentioned dialog information set Breath determines that the corresponding movement node of a upper sentence of input information that above-mentioned terminal is sent, above-mentioned as historical action node, is rung There should be next node in determining above-mentioned historical action node, using the next node of above-mentioned historical action node as target section Point, herein, the corresponding node that acts of a upper sentence can be the movement node for referring to and generating return information for a upper sentence.On Stating dialog information set can be used for recording and information caused by above-mentioned terminal interaction.Finally, above-mentioned executing subject can root Return information is generated according to above-mentioned destination node and above-mentioned digraph, and above-mentioned return information is sent to above-mentioned terminal, for Terminal is presented.As an example, above-mentioned executing subject can be using destination node as starting, by between digraph interior joint Line relationship executes downwards, until generating return information.
In some optional implementations, when above-mentioned target retro information generating mode is mode four, above-mentioned steps 205 can specifically include the following contents: above-mentioned executing subject can according to above-mentioned matching result from above-mentioned question and answer to information aggregate Middle selection target question and answer are to information, as an example, can choose the highest question and answer of matching degree according to matching result to information conduct Target question and answer are to information.And target question and answer are sent to above-mentioned terminal as return information to the answer information in information, for Above-mentioned terminal is presented to the user.
In addition, above-mentioned executing subject can be preset from above-mentioned when above-mentioned target retro information generating mode is mode five (for example, randomly selecting) information is chosen in information aggregate as return information.
It is one of the application scenarios of the method according to the present embodiment for generating return information with continued reference to Fig. 5, Fig. 5 Schematic diagram.In the application scenarios of Fig. 5, user sends input information " I will inquire telephone expenses " by terminal device 501.Later, Server 502 can carry out semantic analysis to input information " I will inquire telephone expenses ", obtain semantic analysis result.Server 502 can also will input information " I will inquire telephone expenses " and question and answer to the question and answer in information aggregate to information the problem of information carry out Matching, obtains matching result.Then, server 502 can be based on semantic analysis result and matching result, from preset 5 Method of determination one is target retro in kind return information generating mode " mode one, mode two, mode three, mode four and mode five " Information generating mode.Finally, generating return information according to mode one, and return information is sent to terminal 501, for terminal 501 are presented to the user.
The method provided by the above embodiment of the application, which is realized, completes target based on semantic analysis result and matching result The selection of return information generating mode is more able to satisfy user demand so that the return information generated be made more to have specific aim.
With further reference to Fig. 6, as the realization to method shown in above-mentioned each figure, this application provides one kind for generating back One embodiment of the device of complex information, the Installation practice is corresponding with embodiment of the method shown in Fig. 2, which specifically may be used To be applied in various electronic equipments.
As shown in fig. 6, the device 600 for generating return information of the present embodiment includes: receiving unit 601, analysis list Member 602, matching unit 603, determination unit 604 and generation unit 605.Wherein, it is logical to be configured to receive user for receiving unit 601 Cross the input information of terminal transmission;Analytical unit 602 is configured to carry out semantic analysis to above-mentioned input information, obtains above-mentioned defeated Enter the semantic analysis result of information, wherein above-mentioned semantic analysis result include target intention classification and with above-mentioned target intention class Not corresponding confidence level;Matching unit 603 is configured to above-mentioned input information and the question and answer pre-established in information aggregate Question and answer to information the problem of information match, obtain matching result, wherein question and answer include problem information and answer to information Information;Determination unit 604 is configured to based on above-mentioned semantic analysis result and above-mentioned matching result, from preset at least one Target retro information generating mode is determined in kind return information generating mode;Generation unit 605 is configured to according to above-mentioned mesh It marks return information generating mode and generates return information, and above-mentioned return information is sent to above-mentioned terminal.
In the present embodiment, for generating the receiving unit 601, analytical unit 602, matching of the device 600 of return information Unit 603, the specific processing of determination unit 604 and generation unit 605 and its brought technical effect can refer to Fig. 2 pairs respectively The related description of step 201 in embodiment, step 202, step 203, step 204 and step 205 is answered, details are not described herein.
In some optional implementations of the present embodiment, above-mentioned at least one return information generating mode includes: base The mode one of return information is generated in the corresponding root node of above-mentioned target intention classification and digraph, classification pair is intended to based on history The root node answered and above-mentioned digraph are generated the mode two of return information, are generated back based on historical action node and above-mentioned digraph The mode three of complex information, based on mode four of the above-mentioned question and answer to information aggregate and the mode five based on presupposed information set, wherein Above-mentioned digraph includes root node, judges node, movement node and expanding node, wherein root node and intention classification one are a pair of It answers, judging to pre-define in node has judgment rule, and acting to pre-define in node has return information create-rule, expanding node In pre-define request data rule.
In some optional implementations of the present embodiment, when above-mentioned target retro information generating mode is mode one When, above-mentioned generation unit 605 is further configured to: the corresponding root node of above-mentioned target intention classification is determined as first object Root node;Return information is generated according to above-mentioned first object root node and above-mentioned digraph, and above-mentioned return information is sent To above-mentioned terminal.
In some optional implementations of the present embodiment, when above-mentioned target retro information generating mode is mode two When, above-mentioned generation unit 605 is further configured to: input that above-mentioned terminal is sent, above-mentioned is obtained from dialog information set The intention classification of a upper sentence for information as history be intended to classification, wherein above-mentioned dialog information set for record with it is above-mentioned Information caused by terminal interaction;Above-mentioned history is intended to the corresponding root node of classification and is determined as the second target root node;According to Above-mentioned second target root node and above-mentioned digraph generate return information, and above-mentioned return information is sent to above-mentioned terminal.
In some optional implementations of the present embodiment, when above-mentioned target retro information generating mode is mode three When, above-mentioned generation unit 605 is further configured to: according to the information recorded in dialog information set, determining above-mentioned terminal hair The corresponding movement node of a upper sentence of input information send, above-mentioned is as historical action node, in response to the above-mentioned history of determination Movement node has next node, using the next node of above-mentioned historical action node as destination node, wherein above-mentioned dialogue letter Breath set is for recording and information caused by above-mentioned terminal interaction;It is generated and is replied according to above-mentioned destination node and above-mentioned digraph Information, and above-mentioned return information is sent to above-mentioned terminal.
In some optional implementations of the present embodiment, when above-mentioned target retro information generating mode is mode four When, above-mentioned generation unit 605 is further configured to: according to above-mentioned matching result from above-mentioned question and answer to choosing mesh in information aggregate It marks question and answer and above-mentioned terminal is sent to as return information to the answer information in information to information, and using target question and answer.
In some optional implementations of the present embodiment, above-mentioned determination unit 604 is further configured to: according to upper Confidence level, above-mentioned matching result and the preset scheduling rule for stating target intention classification determine aforesaid way one, mode Two, one of mode three, mode four and mode five are used as target retro information generating mode.
Below with reference to Fig. 7, it illustrates the computer systems 700 for the server for being suitable for being used to realize the embodiment of the present application Structural schematic diagram.Server shown in Fig. 7 is only an example, should not function and use scope band to the embodiment of the present application Carry out any restrictions.
As shown in fig. 7, computer system 700 includes central processing unit (CPU, Central Processing Unit) 701, it can be according to the program being stored in read-only memory (ROM, Read Only Memory) 702 or from storage section 706 programs being loaded into random access storage device (RAM, Random Access Memory) 703 and execute various appropriate Movement and processing.In RAM 703, also it is stored with system 700 and operates required various programs and data.CPU 701,ROM 702 and RAM 703 is connected with each other by bus 704.Input/output (I/O, Input/Output) interface 705 is also connected to Bus 704.
I/O interface 705 is connected to lower component: the storage section 706 including hard disk etc.;And including such as LAN (local Net, Local Area Network) card, modem etc. network interface card communications portion 707.Communications portion 707 passes through Communication process is executed by the network of such as internet.Driver 708 is also connected to I/O interface 705 as needed.Detachable media 709, such as disk, CD, magneto-optic disk, semiconductor memory etc., are mounted on as needed on driver 708, in order to from The computer program read thereon is mounted into storage section 706 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 707, and/or from detachable media 709 are mounted.When the computer program is executed by central processing unit (CPU) 701, limited in execution the present processes Above-mentioned function.
It should be noted that computer-readable medium described herein can be computer-readable signal media or meter Calculation machine readable storage medium storing program for executing either the two any combination.Computer readable storage medium for example can be --- but not Be limited to --- electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor system, device or device, or any above combination.Meter The more specific example of calculation machine readable storage medium storing program for executing can include but is not limited to: have the electrical connection, just of one or more conducting wires Taking formula computer disk, hard disk, random access storage device (RAM), read-only memory (ROM), erasable type may be programmed read-only storage Device (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), light storage device, magnetic memory device, Or above-mentioned any appropriate combination.In this application, computer readable storage medium can be it is any include or storage journey The tangible medium of sequence, the program can be commanded execution system, device or device use or in connection.And at this In application, computer-readable signal media may include in a base band or as carrier wave a part propagate data-signal, Wherein carry computer-readable program code.The data-signal of this propagation can take various forms, including but unlimited In electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be that computer can Any computer-readable medium other than storage medium is read, which can send, propagates or transmit and be used for By the use of instruction execution system, device or device or program in connection.Include on computer-readable medium Program code can transmit with any suitable medium, including but not limited to: wireless, electric wire, optical cable, RF etc. are above-mentioned Any appropriate combination.
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 program 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 receiving unit, analytical unit, matching unit, determination unit and generation unit.Wherein, the title of these units is in certain situation Under do not constitute restriction to the unit itself, for example, receiving unit is also described as " receiving user and passing through terminal transmission Input information unit ".
As on the other hand, present invention also provides a kind of computer-readable medium, which be can be Included in device described in above-described embodiment;It is also possible to individualism, and without in the supplying device.Above-mentioned calculating Machine readable medium carries one or more program, when said one or multiple programs are executed by the device, so that should Device: the input information that user is sent by terminal is received;Semantic analysis is carried out to above-mentioned input information, obtains above-mentioned input letter The semantic analysis result of breath, wherein above-mentioned semantic analysis result include target intention classification and with above-mentioned target intention classification pair The confidence level answered;By above-mentioned input information with the question and answer that pre-establish to the question and answer in information aggregate to information the problem of information into Row matching, obtains matching result, wherein question and answer include problem information and answer information to information;Based on above-mentioned semantic analysis knot Fruit and above-mentioned matching result determine that target retro information generates from preset at least one return information generating mode Mode;Return information is generated according to above-mentioned target retro information generating mode, and above-mentioned return information is sent to above-mentioned end End.
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 (16)

1. a kind of method for generating return information, comprising:
Receive the input information that user is sent by terminal;
Semantic analysis is carried out to the input information, obtains the semantic analysis result of the input information, wherein described semantic point Analysing result includes target intention classification and confidence level corresponding with the target intention classification;
By the input information and the question and answer pre-established to the question and answer in information aggregate to information the problem of information match, Obtain matching result, wherein question and answer include problem information and answer information to information;
Based on the semantic analysis result and the matching result, from preset at least one return information generating mode Determine target retro information generating mode;
Return information is generated according to the target retro information generating mode, and the return information is sent to the end End.
2. according to the method described in claim 1, wherein, at least one return information generating mode includes:
The mode one of return information is generated based on the corresponding root node of the target intention classification and digraph, is intended to based on history The corresponding root node of classification and the digraph generate the mode two of return information, are based on historical action node and the digraph Generate the mode three of return information, based on the question and answer to the mode four of information aggregate and the mode based on presupposed information set Five, wherein the digraph includes root node, judges node, movement node and expanding node, wherein root node and intention class It does not correspond, judging to pre-define in node has judgment rule, and acting to pre-define in node has return information create-rule, The rule of request data is pre-defined in expanding node.
3. according to the method described in claim 2, wherein, when the target retro information generating mode be mode for the moment, it is described Return information is generated according to the target retro information generating mode, and the return information is sent to the terminal, comprising:
The corresponding root node of the target intention classification is determined as first object root node;
Return information is generated according to the first object root node and the digraph, and the return information is sent to institute State terminal.
4. described when the target retro information generating mode is mode two according to the method described in claim 2, wherein Return information is generated according to the target retro information generating mode, and the return information is sent to the terminal, comprising:
The intention classification that a upper sentence for the input information that the terminal is sent, described is obtained from dialog information set, which is used as, to be gone through History is intended to classification, wherein the dialog information set is for recording and information caused by the terminal interaction;
The history is intended to the corresponding root node of classification and is determined as the second target root node;
Return information is generated according to the second target root node and the digraph, and the return information is sent to institute State terminal.
5. described when the target retro information generating mode is mode three according to the method described in claim 2, wherein Return information is generated according to the target retro information generating mode, and the return information is sent to the terminal, comprising:
According to the information recorded in dialog information set, a upper sentence pair for the input information that the terminal is sent, described is determined The movement node answered has next node as historical action node, in response to the determination historical action node, by the history The next node of node is acted as destination node, wherein the dialog information set is for recording and the terminal interaction Generated information;
Return information is generated according to the destination node and the digraph, and the return information is sent to the end End.
6. described when the target retro information generating mode is mode four according to the method described in claim 2, wherein Return information is generated according to the target retro information generating mode, and the return information is sent to the terminal, comprising:
According to the matching result from the question and answer to choosing target question and answer in information aggregate to information, and by target question and answer pair Answer information in information is sent to the terminal as return information.
7. being set according to the method described in claim 2, described be based on the semantic analysis result and the matching result from advance Target retro information generating mode is determined in fixed at least one return information generating mode, comprising:
The side is determined according to the confidence level of the target intention classification, the matching result and preset scheduling rule One of formula one, mode two, mode three, mode four and mode five are used as target retro information generating mode.
8. a kind of for generating the device of return information, comprising:
Receiving unit is configured to receive the input information that user is sent by terminal;
Analytical unit is configured to carry out semantic analysis to the input information, obtains the semantic analysis knot of the input information Fruit, wherein the semantic analysis result includes target intention classification and confidence level corresponding with the target intention classification;
Matching unit is configured to the input information and the question and answer pre-established to the question and answer in information aggregate to information Problem information is matched, and matching result is obtained, wherein question and answer include problem information and answer information to information;
Determination unit is configured to based on the semantic analysis result and the matching result, from preset at least one Target retro information generating mode is determined in return information generating mode;
Generation unit is configured to generate return information according to the target retro information generating mode, and by the reply Information is sent to the terminal.
9. device according to claim 8, wherein it is described at least one return information generating mode include:
The mode one of return information is generated based on the corresponding root node of the target intention classification and digraph, is intended to based on history The corresponding root node of classification and the digraph generate the mode two of return information, are based on historical action node and the digraph Generate the mode three of return information, based on the question and answer to the mode four of information aggregate and the mode based on presupposed information set Five, wherein the digraph includes root node, judges node, movement node and expanding node, wherein root node and intention class It does not correspond, judging to pre-define in node has judgment rule, and acting to pre-define in node has return information create-rule, The rule of request data is pre-defined in expanding node.
10. device according to claim 9, wherein when the target retro information generating mode be mode for the moment, it is described Generation unit is further configured to:
The corresponding root node of the target intention classification is determined as first object root node;
Return information is generated according to the first object root node and the digraph, and the return information is sent to institute State terminal.
11. device according to claim 9, wherein described when the target retro information generating mode is mode two Generation unit is further configured to:
The intention classification that a upper sentence for the input information that the terminal is sent, described is obtained from dialog information set, which is used as, to be gone through History is intended to classification, wherein the dialog information set is for recording and information caused by the terminal interaction;
The history is intended to the corresponding root node of classification and is determined as the second target root node;
Return information is generated according to the second target root node and the digraph, and the return information is sent to institute State terminal.
12. device according to claim 9, wherein described when the target retro information generating mode is mode three Generation unit is further configured to:
According to the information recorded in dialog information set, a upper sentence pair for the input information that the terminal is sent, described is determined The movement node answered has next node as historical action node, in response to the determination historical action node, by the history The next node of node is acted as destination node, wherein the dialog information set is for recording and the terminal interaction Generated information;
Return information is generated according to the destination node and the digraph, and the return information is sent to the end End.
13. device according to claim 9, wherein described when the target retro information generating mode is mode four Generation unit is further configured to:
According to the matching result from the question and answer to choosing target question and answer in information aggregate to information, and by target question and answer pair Answer information in information is sent to the terminal as return information.
14. device according to claim 9, wherein the determination unit is further configured to:
The side is determined according to the confidence level of the target intention classification, the matching result and preset scheduling rule One of formula one, mode two, mode three, mode four and mode five are used as target retro information generating mode.
15. a kind of server, 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 1-7.
16. a kind of computer-readable medium, is stored thereon with computer program, wherein real when described program is executed by processor The now method as described in any in claim 1-7.
CN201810917669.5A 2018-08-13 2018-08-13 Method and device for generating reply information Active CN109190114B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810917669.5A CN109190114B (en) 2018-08-13 2018-08-13 Method and device for generating reply information

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810917669.5A CN109190114B (en) 2018-08-13 2018-08-13 Method and device for generating reply information

Publications (2)

Publication Number Publication Date
CN109190114A true CN109190114A (en) 2019-01-11
CN109190114B CN109190114B (en) 2022-06-07

Family

ID=64921658

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810917669.5A Active CN109190114B (en) 2018-08-13 2018-08-13 Method and device for generating reply information

Country Status (1)

Country Link
CN (1) CN109190114B (en)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109753561A (en) * 2019-01-16 2019-05-14 长安汽车金融有限公司 A kind of generation method automatically replied and device
CN109948155A (en) * 2019-03-12 2019-06-28 广东小天才科技有限公司 A kind of selection method and device, terminal device of more intentions
CN110008314A (en) * 2019-04-12 2019-07-12 广东小天才科技有限公司 A kind of intents method and device
CN110008050A (en) * 2019-04-11 2019-07-12 北京百度网讯科技有限公司 Method and apparatus for handling information
CN110445707A (en) * 2019-07-23 2019-11-12 精硕科技(北京)股份有限公司 Message processing method and device
CN110532440A (en) * 2019-08-21 2019-12-03 深圳追一科技有限公司 Information processing method, device, equipment and storage medium
CN110929027A (en) * 2019-09-30 2020-03-27 珠海格力电器股份有限公司 Prompting system, prompting method, computer and waste accommodating device
CN111881266A (en) * 2019-07-19 2020-11-03 马上消费金融股份有限公司 Response method and device
CN113010654A (en) * 2021-03-17 2021-06-22 北京十一贝科技有限公司 Question reply method and device applied to insurance industry, electronic equipment and medium
CN113761138A (en) * 2020-06-02 2021-12-07 阿里巴巴集团控股有限公司 Dialog system, dialog method, dialog device, electronic device, and computer-readable storage medium
CN117196035A (en) * 2023-08-31 2023-12-08 摩尔线程智能科技(北京)有限责任公司 Reply content processing method and device, electronic equipment and storage medium

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140222433A1 (en) * 2011-09-19 2014-08-07 Personetics Technologies Ltd. System and Method for Evaluating Intent of a Human Partner to a Dialogue Between Human User and Computerized System
CN105050103A (en) * 2015-07-20 2015-11-11 大唐移动通信设备有限公司 Signalling process identification method and signalling process identification device
CN105068661A (en) * 2015-09-07 2015-11-18 百度在线网络技术(北京)有限公司 Man-machine interaction method and system based on artificial intelligence
CN105094315A (en) * 2015-06-25 2015-11-25 百度在线网络技术(北京)有限公司 Method and apparatus for smart man-machine chat based on artificial intelligence
CN105589848A (en) * 2015-12-28 2016-05-18 百度在线网络技术(北京)有限公司 Dialog management method and device
CN105845137A (en) * 2016-03-18 2016-08-10 中国科学院声学研究所 Voice communication management system
CN106486120A (en) * 2016-10-21 2017-03-08 上海智臻智能网络科技股份有限公司 Interactive voice response method and answering system
WO2017041370A1 (en) * 2015-09-07 2017-03-16 百度在线网络技术(北京)有限公司 Human-computer chatting method and device based on artificial intelligence
CN108090177A (en) * 2017-12-15 2018-05-29 上海智臻智能网络科技股份有限公司 The generation methods of more wheel question answering systems, equipment, medium and take turns question answering system more
US20180174055A1 (en) * 2016-12-19 2018-06-21 Giridhar S. Tirumale Intelligent conversation system

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140222433A1 (en) * 2011-09-19 2014-08-07 Personetics Technologies Ltd. System and Method for Evaluating Intent of a Human Partner to a Dialogue Between Human User and Computerized System
CN105094315A (en) * 2015-06-25 2015-11-25 百度在线网络技术(北京)有限公司 Method and apparatus for smart man-machine chat based on artificial intelligence
CN105050103A (en) * 2015-07-20 2015-11-11 大唐移动通信设备有限公司 Signalling process identification method and signalling process identification device
CN105068661A (en) * 2015-09-07 2015-11-18 百度在线网络技术(北京)有限公司 Man-machine interaction method and system based on artificial intelligence
WO2017041370A1 (en) * 2015-09-07 2017-03-16 百度在线网络技术(北京)有限公司 Human-computer chatting method and device based on artificial intelligence
CN105589848A (en) * 2015-12-28 2016-05-18 百度在线网络技术(北京)有限公司 Dialog management method and device
CN105845137A (en) * 2016-03-18 2016-08-10 中国科学院声学研究所 Voice communication management system
CN106486120A (en) * 2016-10-21 2017-03-08 上海智臻智能网络科技股份有限公司 Interactive voice response method and answering system
US20180174055A1 (en) * 2016-12-19 2018-06-21 Giridhar S. Tirumale Intelligent conversation system
CN108090177A (en) * 2017-12-15 2018-05-29 上海智臻智能网络科技股份有限公司 The generation methods of more wheel question answering systems, equipment, medium and take turns question answering system more

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109753561B (en) * 2019-01-16 2021-04-27 长安汽车金融有限公司 Automatic reply generation method and device
CN109753561A (en) * 2019-01-16 2019-05-14 长安汽车金融有限公司 A kind of generation method automatically replied and device
CN109948155A (en) * 2019-03-12 2019-06-28 广东小天才科技有限公司 A kind of selection method and device, terminal device of more intentions
CN110008050B (en) * 2019-04-11 2023-06-30 北京百度网讯科技有限公司 Method and device for processing information
CN110008050A (en) * 2019-04-11 2019-07-12 北京百度网讯科技有限公司 Method and apparatus for handling information
CN110008314B (en) * 2019-04-12 2022-07-26 广东小天才科技有限公司 Intention analysis method and device
CN110008314A (en) * 2019-04-12 2019-07-12 广东小天才科技有限公司 A kind of intents method and device
CN111881266A (en) * 2019-07-19 2020-11-03 马上消费金融股份有限公司 Response method and device
CN111881266B (en) * 2019-07-19 2024-06-07 马上消费金融股份有限公司 Response method and device
CN110445707A (en) * 2019-07-23 2019-11-12 精硕科技(北京)股份有限公司 Message processing method and device
CN110532440A (en) * 2019-08-21 2019-12-03 深圳追一科技有限公司 Information processing method, device, equipment and storage medium
CN110929027A (en) * 2019-09-30 2020-03-27 珠海格力电器股份有限公司 Prompting system, prompting method, computer and waste accommodating device
CN113761138A (en) * 2020-06-02 2021-12-07 阿里巴巴集团控股有限公司 Dialog system, dialog method, dialog device, electronic device, and computer-readable storage medium
CN113761138B (en) * 2020-06-02 2024-02-02 阿里巴巴集团控股有限公司 Dialogue system, dialogue method, dialogue device, electronic equipment and computer readable storage medium
CN113010654A (en) * 2021-03-17 2021-06-22 北京十一贝科技有限公司 Question reply method and device applied to insurance industry, electronic equipment and medium
CN117196035A (en) * 2023-08-31 2023-12-08 摩尔线程智能科技(北京)有限责任公司 Reply content processing method and device, electronic equipment and storage medium

Also Published As

Publication number Publication date
CN109190114B (en) 2022-06-07

Similar Documents

Publication Publication Date Title
CN109190114A (en) Method and apparatus for generating return information
CN109460513A (en) Method and apparatus for generating clicking rate prediction model
CN108986805B (en) Method and apparatus for sending information
CN110298906A (en) Method and apparatus for generating information
CN108121800A (en) Information generating method and device based on artificial intelligence
CN108520470A (en) Method and apparatus for generating customer attribute information
CN108932220A (en) article generation method and device
CN110096584A (en) A kind of answer method and device
CN108388674A (en) Method and apparatus for pushed information
CN109299477A (en) Method and apparatus for generating text header
CN109325121A (en) Method and apparatus for determining the keyword of text
CN110046254A (en) Method and apparatus for generating model
CN107656996A (en) Man-machine interaction method and device based on artificial intelligence
CN108629011A (en) Method and apparatus for sending feedback information
CN109815365A (en) Method and apparatus for handling video
CN110399465A (en) Method and apparatus for handling information
CN110263142A (en) Method and apparatus for output information
CN109766418A (en) Method and apparatus for output information
CN108182472A (en) For generating the method and apparatus of information
CN109829164A (en) Method and apparatus for generating text
CN110084317A (en) The method and apparatus of image for identification
CN110164429A (en) Voice interactive method and device
CN107590484A (en) Method and apparatus for information to be presented
CN110059623A (en) Method and apparatus for generating information
CN110532983A (en) Method for processing video frequency, device, medium and equipment

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