WO2019205679A1 - 一种信息处理方法及装置 - Google Patents

一种信息处理方法及装置 Download PDF

Info

Publication number
WO2019205679A1
WO2019205679A1 PCT/CN2018/122196 CN2018122196W WO2019205679A1 WO 2019205679 A1 WO2019205679 A1 WO 2019205679A1 CN 2018122196 W CN2018122196 W CN 2018122196W WO 2019205679 A1 WO2019205679 A1 WO 2019205679A1
Authority
WO
WIPO (PCT)
Prior art keywords
slot
task engine
key information
information
task
Prior art date
Application number
PCT/CN2018/122196
Other languages
English (en)
French (fr)
Inventor
严哲峰
尚利峰
蔡涛
钱莉
Original Assignee
华为技术有限公司
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 华为技术有限公司 filed Critical 华为技术有限公司
Priority to EP18915826.4A priority Critical patent/EP3779725A4/en
Publication of WO2019205679A1 publication Critical patent/WO2019205679A1/zh
Priority to US16/868,970 priority patent/US20200264923A1/en

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/02Reservations, e.g. for tickets, services or events
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/95Retrieval from the web
    • G06F16/953Querying, e.g. by the use of web search engines
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/903Querying
    • G06F16/9035Filtering based on additional data, e.g. user or group profiles
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33Querying
    • G06F16/332Query formulation
    • G06F16/3329Natural language query formulation or dialogue systems
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/903Querying
    • G06F16/9032Query formulation
    • G06F16/90332Natural language query formulation or dialogue systems
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46Multiprogramming arrangements
    • G06F9/48Program initiating; Program switching, e.g. by interrupt
    • G06F9/4806Task transfer initiation or dispatching
    • G06F9/4843Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
    • G06F9/4881Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • G06Q30/0282Rating or review of business operators or products
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10Services
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2209/00Indexing scheme relating to G06F9/00
    • G06F2209/48Indexing scheme relating to G06F9/48
    • G06F2209/486Scheduler internals

Definitions

  • the present application relates to the field of computer technologies, and in particular, to an information processing method and apparatus.
  • the input of the human-computer interaction system may be a request for asking for weather, ordering a ticket, encyclopedia knowledge, and a chat, and the human-computer interaction system may feed back information that meets the user's needs according to the user's request.
  • the request input by the user may be "Beijing weather conditions tomorrow"
  • the information that the human-computer interaction system gives back to the user can be "Beijing tomorrow's weather is sunny and cloudy.”
  • the human-computer interaction system must obtain the critical information that is satisfied in the user's request information. If the key information that satisfies the condition is not obtained in the request information of the user, the human-computer interaction system will feedback the failed information to the user, thereby making the user experience poor and the human-computer interaction efficiency low.
  • the application provides an information processing method and device, so that a better user experience is better, and the human-computer interaction efficiency is higher.
  • the application provides an information processing method, including: receiving first request information input by a user; determining a first task engine that is responsible for the first request information, where the first task engine is provided with a first slot Bit; extracting key information in the first request information according to the first slot;
  • the key information is not extracted in the first request information according to the first slot, or if the key information is extracted in the first request information according to the first slot, and the extracted key information is obtained. If the condition is not met, the target key information is obtained in the shared parameter list of the user, where the shared parameter list includes at least a correspondence between the second slot and the target key information, where the second slot and the location.
  • the first slot has the same slot type, the second slot is a slot in the second task engine, and the first task engine is different from the second task engine.
  • the target key information is obtained in the shared parameter list of the user, compared to the prior art.
  • the key information is not extracted or the extracted key information does not satisfy the condition, that is, the information that the feedback processing fails, so that the user inputs the requested solution again, and a better user experience can be obtained, and the human-computer interaction efficiency is high.
  • the shared parameter list includes a correspondence between a first task engine and one or more task engines, and a correspondence between slots included in any two task engines of the corresponding relationship, and a Corresponding relationship between the two slots and the target key information, wherein the one or more task engines include the second task engine;
  • the target key information may be obtained in the shared parameter list of the user by: determining, according to the correspondence between the first task engine and one or more task engines, the second task engine; according to the correspondence between the slots And determining, in the slot included in the second task engine included in the shared parameter list, a second slot that has a corresponding relationship with the first slot of the first task engine; Corresponding relationship between the slot and the target key information determines the target key information.
  • the second task engine is configured to be responsible for the second request information of the user, where the second request information and the first request information occur within a preset time, and the second request information is The first request information occurs within a preset number of dialog rounds, and the round of dialogue may specifically refer to a process in which the user inputs a request information and the human-machine interaction system feeds back a message.
  • the shared parameter list is combined with the time when the request information is generated, and the second task engine is determined, and the second task engine can be accurately positioned to improve the accuracy of the target key information filled in the second slot.
  • the mapping between the second slot and the target key information is stored in the shared parameter list, and the target key information is obtained in the shared parameter list of the user in the following manner: Determining a context parameter of the user, determining a second task engine adjacent to the time of the first task engine; determining, in a slot of the second task engine, a second slot that is the same as the first slot type And determining the target key information according to the correspondence between the second slot and the target key information.
  • the second task engine is configured to be responsible for the second request information of the user, where the second task engine is provided with the same slot type as the first slot type, and the time neighbor refers to the The occurrence time of the second request information is adjacent to the occurrence time of the first request information.
  • the preset may be preset. The priority information is selected.
  • the shared parameter list only one correspondence relationship is stored in the shared parameter list, that is, the correspondence between different slots of the task engine and the target key information, and the implementation manner of the shared parameter list (in the above shared parameter list)
  • the three corresponding relationships need to be stored.
  • the amount of data stored in the shared parameter list can be reduced, and the rate of accessing the shared parameter list is improved.
  • the shared parameter list is a shared parameter list exclusive to the second task engine
  • the shared parameter list exclusive to the second task engine may include a second slot and key information.
  • the target key information may be obtained in the shared parameter list of the user by: determining, according to the correlation between the first task engine and the second task engine, the second task engine, the first task engine Correlating with the second task engine is greater than a preset threshold; determining, according to the second task engine, a shared parameter list exclusive to the second task engine; a shared parameter list exclusive to the second task engine
  • the slot is the same as the slot type of the first slot, and the target slot is determined according to the relationship between the second slot and the target key information pair.
  • each task engine of the human-computer interaction system maintains a shared parameter list, and the shared parameter list stores the correspondence between the slot of the task engine and the target key information, corresponding to the above.
  • the two shared parameter lists one for maintaining the three correspondences, and the other for storing the correspondence between the slots of the multiple task engines and the target key information.
  • the shared parameter list of the present application can further reduce the shared parameter list. The amount of data stored in the information further increases the rate at which the shared parameter list is accessed.
  • the key information that does not satisfy the condition includes the reference information, and in the corresponding relationship between the second slot and the target key information included in the shared parameter list, the second slot
  • the target key information corresponding to the bit is unique, and the key information corresponding to the second slot is updatable.
  • the target information corresponding to the second slot may be updated in the following manner, specifically: receiving third request information input by the user, where the third request information is used by the second task engine responsible for: in the third request information, extracting key information corresponding to the second slot of the second task engine; and according to the third request information, extracting key information corresponding to the second slot And updating the target key information corresponding to the second slot in the correspondence between the second slot and the target key information in the shared parameter list.
  • a plurality of slots may be disposed in the first task engine, and the first slot may be determined in the multiple slots in the following manner. Specifically, determining that the first task engine is configured a plurality of slots, the plurality of slots having priority information; determining the first slot according to priority information of the multiple slots.
  • the slot type may be specifically a time, a place, a behavior, a character, or the like.
  • the shared parameter list may be stored at the local end, such as a memory in the information processing device, or may be acquired from an external storage device by a communication network or an access interface.
  • the present application provides an information processing apparatus for a terminal device or a server, comprising: a unit or means for performing the steps of the above first aspect.
  • the present application provides an information processing apparatus for a terminal device or a server, comprising at least one processing element and at least one storage element, wherein the at least one storage element is configured to store a program and data, the at least one processing The elements are for performing the methods provided by the first aspect of the application.
  • the present application provides a computer readable storage medium having stored therein instructions that, when run on a computer, cause the computer to perform the method of the first aspect described above.
  • the application provides a computer program comprising computer instructions that, when executed by a computer, cause the computer to perform the method of the first aspect.
  • the application provides a human-computer interaction system, including a central control module and a task engine module, wherein the central control module is implemented by an information processing apparatus provided by the application, and the engine module is used to implement a first task engine and a The function of the second task engine.
  • the human-computer interaction system further includes a multi-task parameter sharing management module, and the shared parameter list may be stored in the sharing parameter management module.
  • Figure 1 is a human-computer interaction system provided by the present application.
  • FIG. 3 is an information processing flow provided by the present application.
  • FIG. 4 is another information processing flow provided by the present application.
  • FIG. 5 is a schematic diagram of a shared parameter list provided by the present application.
  • FIG. 6 is another schematic diagram of a shared parameter list provided by the present application.
  • FIG. 7 is still another schematic diagram of a shared parameter list provided by the present application.
  • FIG. 8 is still another information processing flow provided by the present application.
  • Figure 9 is a structural diagram of a communication device provided by the present application.
  • FIG. 10 is still another structural diagram of a communication device provided by the present application.
  • the present application provides a human-computer interaction system, which mainly includes a central control module and a task engine module.
  • the human-computer interaction system may further include a task state library module.
  • the central control module is mainly responsible for distributing the dialog request to one or more related task engine modules, and then summarizing the results returned by the multiple task engine modules to obtain the final result.
  • the task engine module includes a plurality of task engines.
  • Each task engine is mainly used to be responsible for a certain domain of conversation services, for example, parsing request information input by a user to obtain key information that satisfies the conditions.
  • key information that satisfies the ticket booking task such as the origin information, the destination information, and the time information, may be extracted from the request information input by the user.
  • the task state library module is mainly used to store the intermediate state of the user dialogue task.
  • the task engine module may execute the corresponding task according to the key information, or the central control module may perform the corresponding task according to the key information, or
  • the smart terminal may perform a corresponding task according to the key information, or may develop a new module in the human-computer interaction system for performing a corresponding task according to the key information.
  • the subject who performs the corresponding task based on the key information is not specifically limited.
  • the function of the human-machine interaction system may be implemented by a server, may be implemented by a terminal device, or may be implemented by a server and a terminal device.
  • the human-computer interaction system described above mainly includes a central control module and a task engine module. It can be understood that when the function of the human-computer interaction system is implemented by a server, the server includes a central control module and a task engine.
  • the module when the function of the human-machine interaction system is implemented by the terminal device, includes a central control module and a task engine module.
  • the "the task engine module sends the key information to the central control module after the key information satisfying the condition is acquired, and the central control module executes the corresponding task" as an example to describe the entire human-computer interaction in detail.
  • the user inputs the request information to the smart terminal, wherein the request information may be voice information, or text information, or image information.
  • the smart terminal forwards the user's request information to the central control module of the human-computer interaction system.
  • the central control module forwards the request information of the user to one or more task engines in the task engine module.
  • the task engine may parse the request information, obtain key information that satisfies the condition, and send the key information that satisfies the condition to the central control module.
  • the central control module can perform corresponding tasks according to key information fed back by the task engine module.
  • the task engine can store the corresponding parsing result into the corresponding task state in the task state library module.
  • the smart terminal can send a request for "user A wants to book a ticket from Beijing to Shanghai tomorrow" to the central control module.
  • the central control module may first determine that the request is processed by the "booking ticket task engine”, and then send the request of "user A to book a ticket from Beijing to Shanghai tomorrow" to the ticket booking task engine in the task engine module.
  • the feedback information is ****", and the feedback information is sent to the smart terminal, and the smart terminal can display the feedback information to the user by means of text, voice or image.
  • the human-computer interaction system will feedback “processing failure, please re-enter information” and other prompts.
  • the user needs to re-enter the request information, which makes the human-computer interaction inefficient. For example, if the user first enters the request information of "I want to book a ticket from Beijing to Shanghai tomorrow", and then enters the request information of "How is the weather there?", because of the request information on "How is the weather there?"
  • key information such as weather and location that meets the weather query task cannot be extracted.
  • the human-computer interaction system feedback “I would like to ask the weather conditions of that city”, which makes the user experience worse, human-computer interaction. Inefficient.
  • the present application provides a human-machine interaction system, which includes at least: a central control module, a task engine module, and a multi-task parameter sharing management module.
  • the method further includes: a task state library module.
  • the shared parameter list may be stored in the multi-task parameter sharing management module, and the number of the shared parameter lists may be one, and the n task engines in the human-computer interaction system may share the shared parameter list. Alternatively, the number of shared parameter lists may also be multiple, and each task engine in the human-computer interaction system monopolizes a shared parameter list. For example, the entire human-computer interaction system includes n task engines. Then, the multi-task parameter sharing management module may also include n shared parameter lists, and each of the n task engines exclusively shares one of the n shared parameter lists. parameter list.
  • the multi-task parameter management module may further include a multi-task parameter sharing management unit, configured to manage the shared parameter list.
  • the shared parameter list may be generated according to a historical session parameter of the user.
  • the task engine module when the task engine module does not extract the key information in the request information input by the user, or extracts the key information but the extracted key information does not satisfy the condition, the task module may obtain the shared parameter list.
  • the key information that meets the requirements of the target does not require repeated input by the user, the user experience is better, and the efficiency of human-computer interaction is effectively improved.
  • the present application provides an information processing method, and the execution body of the method may be the smart terminal shown in FIG. 1 or FIG. 2, or may be operated by using FIG. 1 or FIG. 2 above.
  • a server showing a human-computer interaction system which may be, but is not limited to, a cloud server.
  • the server running the human-computer interaction system is taken as an example of an execution subject, and as shown in FIG. 3, the flow of the information processing method is described in detail.
  • Step S301 The server receives the first request information input by the user.
  • the first request information may be voice information, text information, image information, or the like.
  • the user may input the first request information to the smart terminal, and the smart terminal may forward the first request information to the server.
  • Step S302 The server performs intent identification on the received first request information, and determines a first task engine that is responsible for the first request information.
  • the central control module in the server may perform intent identification on the first request information, and determine a first task engine that is responsible for the first request information according to the intention identification result. For example, if the central control module determines that the intent recognition result of the first request information is a booking ticket, then determining that the first task engine responsible for the first request information is the booking ticket task engine.
  • each task engine may be provided with a different slot, and the slot may be specifically a variable, and the slot view may be specifically corresponding to the slot.
  • Key Information In the embodiment of the present application, the slot may also be referred to as an information slot, and the key information corresponding to the slot may also be referred to as slot information.
  • the process of the present application is described in detail by taking the first slot in the first task engine as an example. It should be noted that, in the embodiment of the present application, the words “first, second” in the first task engine, the second task engine, the first slot, and the second slot are only used to distinguish the purpose of description. It is not to be understood as indicating or implying relative importance, nor as an indication or implied order.
  • Step S303 The server extracts key information in the first request information according to the first slot.
  • the first task engine is the “scheduled ticket task engine”, and the first slot is “starting slot”, and the server may extract the “starting slot” corresponding to the first request information.
  • Key Information For example, if the first slot is the “destination slot”, the server may extract the key information corresponding to the “destination slot” in the first request information.
  • Step S304 If key information is not extracted in the first request information according to the first slot, or if key information is extracted in the first request information according to the first slot, and extracted The key information does not satisfy the condition, and the server acquires the target key information in the shared parameter list of the user.
  • the shared parameter list includes at least a corresponding relationship between the second slot and the target key information, where the second slot has the same slot type as the first slot, and the slot type is time, Location, behavior, or character, etc., the second slot is a slot in the second task engine, and the first task engine is different from the second task engine.
  • the key information that does not satisfy the condition may be specifically referred to as information, and the reference information may specifically be that the key information extracted in the first request information is a pronoun, and the pronoun may be “this” or “that”. , "here", "where” and so on.
  • Step S305 The server executes the corresponding task according to the target key information.
  • the shared parameter can be used in the user.
  • the target key information is obtained in the list, and the corresponding task is executed according to the target key information.
  • a plurality of slots may be provided in the first task engine, and the levels of each slot are the same.
  • the key information corresponding to each slot may be filled in any order.
  • the first task engine has "starting slot”, "destination slot” and "time slot”, which can be in accordance with "starting slot”, "destination slot” and "time slot”.
  • the order of the bits is to fill in the key information corresponding to each slot in turn, or to fill each slot in the order of "time slot", "destination slot” and "start slot”. Corresponding key information, etc.
  • the first task engine may be provided with multiple slots, each slot having a different priority, and the key information corresponding to each slot may be filled according to the priority of the slot.
  • the booking ticket engine there are three slots, which are “starting slot”, “destination slot” and “time slot”, and setting “time slot” has the highest priority.
  • the priority of the "starting slot” is the highest, and the priority of the "destination slot” is the lowest.
  • the key information corresponding to the "time slot” can be filled first, and then the "starting slot” is filled.
  • the corresponding key information is finally filled with the key information corresponding to the “destination slot”.
  • the process of filling the key information corresponding to each slot is performed by taking the first slot as an example, and the process is described in detail: first, in the request information input by the user, the corresponding slot is extracted. The key information, if the key information can be extracted, and the extracted key information meets the condition, and the filling of the first slot is completed. If the key information is not extracted in the request information, or the key information can be extracted, but the extracted key information does not satisfy the condition, the target key information is acquired in the shared parameter list of the user, and the target key information is used as the Key information corresponding to the first slot.
  • the method may further include: the server updating the shared parameter list of the user according to a semantic recognition result of each task engine, where the semantic recognition result includes intent information and a key that satisfies the condition information.
  • the central control module of the machine interaction system performs intent identification on the first request information input by the user, and finds that the first request information corresponds to “booking”, and the entire human-computer interaction system includes “booking ticket task engine” and “booking train ticket”.
  • the task engine and the "custom ticket ticket task engine”, the central control module can send the first request information to the "book ticket task engine", "book ticket task engine” and "book ticket task engine” respectively.
  • the first request information may be referred to as the above-mentioned "scheduled train ticket task engine", "booking ticket ticket task engine”, and “booking ticket task engine”, corresponding to three task engines.
  • the N+1 task engines include the first task engine and the other N task engines
  • the N+1 is
  • the task engine may send the semantic recognition result for the first request information to the central control module, and the central control module may select a semantic recognition result satisfying the rule, such as an optimal semantic, from the N+1 semantic recognition results. Identify the result and update the shared parameter list of the user.
  • the server shown in FIG. 3 includes the central control module, the task engine module, and the multi-task parameter sharing management module, and the information processing shown in FIG. 3 is described in detail.
  • Step S401 The smart terminal receives the first request information input by the user, and sends the first request information to the central control module of the human-machine interaction system.
  • Step S402 The central control module performs intent identification on the first request information, acquires one or more task engines related to the first request information, and sends the first request information to one or more of the task engine modules. Task engines.
  • Step S403 Determine a slot corresponding to the task engine for a task engine.
  • the slots of the different task engines may be corresponding to different slots.
  • the slots corresponding to the ticket booking engine may be “starting slot”, “destination slot” or “location slot”.
  • the slot corresponding to the weather query task engine can be a “slot slot” or a “time slot”.
  • Step S404 The task engine extracts key information in the first request information according to the slot corresponding to the task engine.
  • the corresponding slot may specifically include a “location slot” and a “time slot”, and the weather query task engine may sequentially extract the “first request information input by the user”.
  • the first request information input by the user may be “query weather conditions of Beijing tomorrow”, and then the weather query task engine may sequentially extract the key information “Beijing” corresponding to the “location slot” in the first request information. , extract the key information "Tomorrow” corresponding to the "time slot”.
  • Step S405 For a slot, if the key information can be extracted in the first request information, and the extracted key information satisfies the condition, step S409 is performed, otherwise step S406 is performed.
  • Step S406 The task engine sends a read command to the multi-task parameter sharing management module.
  • Step S407 The multitasking parameter sharing management module acquires target key information from the shared parameter list of the user.
  • Step S408 The multi-task parameter sharing management module sends the target key information to the task engine.
  • Step S409 The task engine acquires a semantic recognition result for the first request information.
  • the semantic recognition result includes intent information for the first request information and key information that satisfies the condition.
  • Step S410 The task engine sends the semantic recognition result for the first request information to the central control module.
  • Step S411 The central control module updates the shared parameter list of the user in the multitasking parameter sharing management module according to the semantic recognition result sent by the task engine.
  • the task engine module reads the information in the shared parameter list, and the central control module updates the shared parameter list as an example.
  • the body that reads the information in the shared parameter list and the body that updates the shared parameter list are not limited.
  • the central control module can be responsible for reading information from the shared parameter list and updating the shared parameter list.
  • the task engine module is responsible for reading information from the shared parameter list and updating the shared parameter list.
  • the central control module is responsible for reading information from the shared parameter list, and the task engine module is responsible for updating the shared parameter list.
  • the task engine module is responsible for reading information from the shared parameter list, and the central control module is responsible for updating the shared parameter list.
  • the embodiment of the present application provides the following two shared parameter lists, which are a first shared parameter list and a second shared parameter list.
  • the target key information can be obtained by using the first shared parameter list or the second shared parameter list.
  • the first type of shared parameter list the number of shared parameter lists is one, and all the task engines in the human-computer interaction system share the shared parameter list, that is, only one shared parameter list is stored in the multi-task parameter sharing management module.
  • the first type of shared parameter list may specifically include: a correspondence between the first task engine and one or more task engines, and any two of the corresponding relationships exist. Correspondence between the slots included in the task engine and the correspondence between the second slot and the target key information.
  • the second task engine is included in the one or more task engines.
  • the second task engine is configured to be responsible for second request information of the user, where the second request information and the first request information may occur within a preset time, and the second request information And the first request information occurs within a preset number of dialog rounds.
  • the first shared parameter list may specifically store three corresponding relationships, namely: a correspondence between different task engines, and if there is a correspondence between the two task engines, the two task engines include Correspondence between slots and the corresponding relationship between the slots included in each task engine and the target key information.
  • the process of acquiring target key information may be: first acquiring a preset time (such as 10 minutes) and a preset number of dialogue rounds (such as 5 rounds of dialogue,
  • the process of one round of dialogue specifically including the user inputting a request information, the human-computer interaction system feeding back a message, and the task engine interacting with the user, for example, the task engine having the interaction may include an ⁇ air ticket reservation task engine, a subscription take-out task engine ⁇ .
  • the task engine that has a corresponding relationship with the weather query task engine in the shared parameter list is obtained, for example, may be a ⁇ booking ticket task engine, a train ticket task engine, a hotel reservation task engine ⁇ .
  • the two take the intersection that is, the ⁇ ticket reservation task engine, order takeaway task engine ⁇ and ⁇ booking ticket task engine, booking train ticket task engine, hotel reservation task engine ⁇ take the intersection, it can be seen that the intersection of the two is ⁇ order Ticket mission engine ⁇ .
  • the slot of the ticket reservation task engine the slot corresponding to the location slot of the weather query task engine is queried.
  • Bit which is the destination slot of the ticketing task engine; finally, the target key information corresponding to the destination slot, namely Shanghai, is used as the key information of the location slot of the weather query task engine.
  • the first type of shared parameter list may specifically include a correspondence between the second slot and the target key information.
  • the following key methods can be used to determine the target key information:
  • the second task engine is configured to be responsible for the second request information of the user, where the second task engine is provided with the same slot type as the first slot type, and the time neighbor refers to the The occurrence time of the second request information is adjacent to the occurrence time of the first request information;
  • the first task engine is adjacent to the second task engine, and may specifically include the following two situations.
  • the first task engine is responsible for the first
  • the task request is adjacent to the second task request that is responsible for the second task engine, that is, the first task request input by the user is adjacent to the second task request in time, for example, the user is first in the human machine
  • the first task request "book tickets from Beijing to Shanghai” is input
  • the second task request "how weather is there" is input in the human-computer interaction system.
  • the first task request may be considered to be temporally adjacent to the second task request.
  • the above-mentioned "first task request” is set by the "air ticket reservation” task engine, and the above-mentioned “second task request” is responsible for the "weather inquiry” task engine, then the above-mentioned "air ticket reservation” task engine and "weather inquiry” can be identified.
  • the task engine is adjacent in time.
  • the process of obtaining the target key information may be: first, in the context parameter of the user, acquiring a task engine adjacent to the weather query task engine, that is, the user is Before interacting with the weather query task engine, the task engine interacting with it, for example, as shown in FIG. 6, the task engine that is temporally adjacent to the weather query task engine is a ticket reservation task engine. Then, in the slot of the ticket reservation task engine, query the slot with the same slot type as the slot type, for example, the slot of the same type is the destination slot. Finally, use the ticket booking engine in the shared parameter list.
  • the key information corresponding to the destination slot namely Shanghai, is the key information corresponding to the location slot in the weather query field.
  • the second shared parameter list the number of shared parameter lists is multiple, and each task engine monopolizes a shared parameter list.
  • the entire human-computer interaction system includes N task engines, and the multi-task parameter sharing management module can store N shared parameter lists, and each of the N task engines exclusively shares one of the N shared parameter lists. parameter list.
  • the corresponding relationship between the second slot and the target key information may be specifically stored in the shared parameter list.
  • Obtaining the target key information in the following manner, specifically: obtaining the first task engine corresponding to the first request information; obtaining the second task engine, wherein the second task engine corresponds to the second vertical domain, The first task engine corresponds to the first vertical domain, the correlation between the first vertical domain and the second vertical domain is greater than a preset threshold, and the correlation between the first vertical domain and the second vertical domain is passed Obtained by the historical data clustering; in the shared parameter list exclusive to the second task engine, searching for the same slot type as the slot type of the first slot as the second slot, according to the second The relationship between the slot and the target key information pair determines the target key information.
  • the shared parameter list may be generated according to historical conversation parameters of the user and the human-computer interaction system. For example, if the user inputs the request information of “a ticket for booking Beijing to Shanghai today” to the human-computer interaction system, the ticket reservation task engine may extract the key information corresponding to the “starting slot” in the request information. "Key information” in Beijing, “destination slot” corresponds to the key information "Shanghai”, "time slot” corresponding to the key information "today”. Finally, according to the key information corresponding to the different slots, the shared parameter list exclusive to the booking ticket task engine is updated.
  • the above example is still used.
  • the user first inputs the request information of “a ticket for booking Beijing to Shanghai today” in the human-computer interaction system, and generates the shared parameter list shown in FIG. 7 above.
  • the user inputs the request information of “how is the weather there”, and can find that the request information is processed by the weather query task engine, and then the weather query task engine can be in the request information input by the user.
  • the “missing information” and the “referring information” may not be set to meet the conditions. In this case, the following key information may be obtained by using the following methods:
  • the correlation between the first vertical domain and the second vertical domain can be obtained by historical data clustering.
  • the target key information is obtained in the shared parameter list exclusive to the second task engine.
  • the second task engine can be set as the ticket booking task engine, and then the key information satisfying the condition can be obtained in the shared parameter list shown in FIG. 7 above.
  • the key information corresponding to the “location slot” in the weather task engine may be queried in the shared parameter list shown in FIG. 7 , and it can be seen from Table 7 above.
  • “Shanghai” is the key information that satisfies the conditions corresponding to the “location slot” in the weather query task engine.
  • the key information corresponding to the “time slot” of the weather task engine can be queried. It can be seen from the above table 7 that “today” is the weather query task engine. The time slot corresponds to the key information that satisfies the condition.
  • the correspondence between the second slot and the target key information is stored.
  • the key information corresponding to the second slot is unique and updateable.
  • the key information corresponding to the second slot may be updated in the following manner: the server receives third request information input by the user, and the third request information is responsible by the second task engine; Determining, in the third request information, key information corresponding to the second slot of the second task engine; the server according to the key information corresponding to the second slot extracted in the third request information, And updating the target key information corresponding to the second slot in the correspondence between the second slot and the target key information in the shared parameter list.
  • the present application provides another process of information processing, including:
  • the first round of dialogue the user inputs the request information of "How is the weather in Beijing today" to the human-computer interaction system, and the human-computer interaction system can perform intent identification on the request information input by the user, and determine that the request information belongs to the weather query task engine for processing.
  • the weather query task engine's location slot can extract the key information "Beijing" in the above request information.
  • the time slot of the weather query task engine can extract the key information "Today” in the above request information, and the last generated historical parameter list can be seen in Table 1 below.
  • the second round of dialogue the user continues to input the request information of “booking a ticket for Beijing to wholesome Aires tomorrow” to the human-computer interaction system.
  • the human-computer interaction system can identify the user's request information. And determining that the request information belongs to the ticket reservation task engine. Then, the ticket reservation task engine can extract the key information "Beijing" and the "destination slot” corresponding to the key information "wholesome Aires” in the request information.
  • the key information corresponding to the time slot is “Tomorrow”.
  • the historical parameter list shown in Table 1 above and the historical parameter list shown in Table 2 above may be stored in the first shared parameter list, or may be stored in the second shared parameter list. And if stored in the second shared parameter list, the historical parameter list shown in Table 1 above may be a shared parameter list exclusive to the weather query task engine, and the historical parameter list shown in Table 2 above may be a ticket reservation task engine. A list of shared parameters that are exclusive.
  • the third round of dialogue the user enters the request information of "How is the weather there" to the human-computer interaction system.
  • Step S801 Receive request information of the third round of dialogue input by the user, "How is the weather there?"
  • Step S802 extracting key information in the request information of the third round of dialogue.
  • step S803 the method may further include: step S804, step S805, step S806, and step S807.
  • Step S804 determining whether the filling of the information slot has a predefined rule; if not, executing step S805; if yes, executing step S808.
  • the predefined rule refers to whether the information slot can be filled by using a shared parameter list
  • Step S805 determining whether the missing information slot can supplement the key information by using the sensor. If yes, go to step S806; otherwise, go to step S807;
  • Step S806 Filling in "location information" with a sensor.
  • Step S807 prompting the user to input "place information”.
  • the purpose of setting step S804 to step S807 is to set different manners for the filling of the information slot, for example, a method of sharing a parameter list, a method of using a sensor, or a manual by a user may be adopted.
  • the way to enter for example, a method of sharing a parameter list, a method of using a sensor, or a manual by a user may be adopted. The way to enter.
  • step S808 may be directly performed after step S803, that is, the key information of the information slot is filled by directly setting the shared parameter list.
  • Step S810 Query the historical "location" parameter of the user in the shared parameter list according to the user ID.
  • Step S811 It is judged whether the history "place" parameter has a value; if there is no value, step S812 is performed; otherwise, step S813 is performed.
  • Step S812 prompt the user to input "place" information.
  • Step S813 Fill the value of the historical "location” parameter into the "location” slot.
  • the "destination: wholesome Aires" shown in Table 2 above can be filled into the "location" slot to generate the key information shown in Table 4.
  • Step S814 Query the historical “time” parameter of the user in the shared parameter list according to the user ID.
  • Step S815 Determine whether the historical "time" parameter has a value. If there is a value, step S816 is performed; otherwise, step S817 is performed.
  • Step S816 Fill the value of the historical "time” parameter into the "time” slot.
  • the “time: tomorrow” shown in Table 2 above may be filled into the time slot to generate key information shown in Table 5.
  • Step S817 prompt the user to input "time” information.
  • the historical conversation parameter of the user can be utilized to obtain the key information that satisfies the condition, so that the user does not need to Enter the request information again, the user experience is better, and the human-computer interaction efficiency is higher.
  • an information processing apparatus 900 which may be specifically a terminal device or a server, and the like, and includes:
  • the transceiver unit 901 is configured to receive first request information input by the user.
  • the transceiver unit 901 may be a hardware device such as a transceiver, or may be a logical unit such as a data interface, which is not limited in this application.
  • the processing unit 902 is configured to perform intent identification on the first request information, and determine a first task engine that is responsible for the first request information, where the first task engine is provided with a first slot, according to the first
  • the slot extracts the key information in the first request information, if the key information is not extracted in the first request information according to the first slot, or if the first slot is in the first If the key information is extracted from the request information and the extracted key information does not satisfy the condition, the target key information is obtained in the shared parameter list of the user;
  • the shared parameter list includes at least a corresponding relationship between the second slot and the target key information, where the second slot has the same slot type as the first slot, and the second slot is A slot in the second task engine, the first task engine being different from the second task engine.
  • the embodiment of the present application further provides a structure of the information processing apparatus 100.
  • the information processing apparatus 100 may be specifically a terminal device or a server.
  • the information processing apparatus 100 may be specific.
  • the central control module in the server, or the function thereof is implemented by the central control module in the server; further optionally, the information processing device 100 may be specifically a central control module in the terminal device, or The function is implemented by a central control module in the terminal device.
  • the information processing apparatus 100 may include a communication interface 101 and a processor 102.
  • the memory 103 may also be included in the information processing apparatus 100.
  • the memory 103 may be provided inside the information processing apparatus 100, and may be provided outside the information processing apparatus.
  • the processing unit 902 shown in FIG. 9 above may be implemented by the processor 102.
  • the transceiver unit 901 can be implemented by the communication interface 101.
  • the processor 102 receives the traffic data through the communication interface 101.
  • the communication interface 101 in the embodiment of the present application may be a circuit, a bus, a transceiver, or any other device that can be used for information interaction.
  • the processor 102 in the embodiment of the present application may be a general-purpose processor, a digital signal processor, an application specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or a transistor logic device, and a discrete hardware component, which may be implemented or executed.
  • a general purpose processor can be a microprocessor or any conventional processor or the like.
  • the steps of the method disclosed in the embodiments of the present application may be directly implemented as a hardware processor, or may be performed by a combination of hardware and software units in the processor.
  • Program code for execution by processor 102 to implement the above methods may be stored in memory 103.
  • Memory 103 is coupled to processor 102.
  • Processor 102 may operate in conjunction with memory 103.
  • the memory 103 may be a non-volatile memory, such as a hard disk drive (HDD) or a solid-state drive (SSD), or a volatile memory such as a random access memory (random). -access memory, RAM).
  • Memory 103 is any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited thereto.
  • connection medium between the communication interface 101, the processor 102, and the memory 103 is not limited in the embodiment of the present application.
  • the memory 103, the processor 102, and the communication interface 101 are connected by a bus in FIG. 10, and the bus is indicated by a thick line in FIG. 10, and the connection manner between other components is only schematically illustrated. Not limited to limits.
  • the bus can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only one thick line is shown in FIG. 10, but it does not mean that there is only one bus or one type of bus.
  • the embodiment of the present application further provides a computer storage medium, where the software program stores a software program, and the software program can implement any one or more of the foregoing when being read and executed by one or more processors.
  • the computer storage medium may include various media that can store program codes, such as a USB flash drive, a removable hard disk, a read only memory, a random access memory, a magnetic disk, or an optical disk.
  • the present application provides a computer program comprising computer instructions that, when executed by a computer, cause the computer to perform the methods provided by any one or more of the above embodiments.
  • embodiments of the present application can be provided as a method, system, or computer program product.
  • the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment in combination of software and hardware.
  • the application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.
  • the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
  • the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
  • the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Databases & Information Systems (AREA)
  • General Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • Mathematical Physics (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Software Systems (AREA)
  • Human Resources & Organizations (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Finance (AREA)
  • Artificial Intelligence (AREA)
  • Accounting & Taxation (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Human Computer Interaction (AREA)
  • Primary Health Care (AREA)
  • Game Theory and Decision Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Information Transfer Between Computers (AREA)

Abstract

一种信息处理方法及装置,该方法包括:接收用户输入的第一请求信息;确定负责所述第一请求信息的第一任务引擎,所述第一任务引擎中设有第一槽位;根据所述第一槽位在所述第一请求信息中提取关键信息;若根据所述第一槽位在所述第一请求信息中提取不到关键信息,或者,若根据所述第一槽位在所述第一请求信息中提取到关键信息且提取到的关键信息不满足条件,则在所述用户的共享参数列表中,获取目标关键信息。可获得较好的用户体验较好,人机交互效率较高。

Description

一种信息处理方法及装置
本申请要求于2018年04月26日提交中国国家知识产权局、申请号为201810388248.8、发明名称为“一种信息处理方法及装置”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及计算机技术领域,尤其涉及一种信息处理方法及装置。
背景技术
随着互联网的飞速发展,人机交互系统的应用越来越广泛。其中,人机交互系统的输入可以是询问天气、订购机票、百科知识和闲聊等请求,而人机交互系统可根据用户的请求,反馈给满足用户需求的信息,比如,用户输入的请求可以是“北京明天的天气情况”,而人机交互系统反馈给用户的信息可以为“北京明天的天气为晴转多云”。
在现有技术中,人机交互系统必须在用户的请求信息中,获取满足的关键信息。如果在用户的请求信息中,获取不到满足条件的关键信息,人机交互系统将会反馈处理失败的信息至用户,从而使得用户体验较差,人机交互效率低下。
发明内容
本申请提供一种信息处理方法及装置,以获得较好的用户体验较好,人机交互效率较高。
第一方面,本申请提供一种信息处理方法,包括:接收用户输入的第一请求信息;确定负责所述第一请求信息的第一任务引擎,所述第一任务引擎中设有第一槽位;根据所述第一槽位在所述第一请求信息中提取关键信息;
若根据所述第一槽位在所述第一请求信息中提取不到关键信息,或者,若根据所述第一槽位在所述第一请求信息中提取到关键信息且提取到的关键信息不满足条件,则在所述用户的共享参数列表中,获取目标关键信息;其中,所述共享参数列表中至少包括第二槽位与目标关键信息的对应关系,所述第二槽位与所述第一槽位具有相同的槽位类型,所述第二槽位为所述第二任务引擎中的槽位,所述第一任务引擎与所述第二任务引擎不同。
在本申请实施例中,如果针对第一槽位提取不到关键信息,或提取到的关键信息不满足条件,则在用户的共享参数列表中,获取目标关键信息,相对于现有技术中的方案,提取不到关键信息或提取到的关键信息不满足条件,即反馈处理失败的信息,使用户再次输入请求的方案,可获得较好的用户体验,人机交互效率较高。
在一种可能的实现中,所述共享参数列表中包括第一任务引擎与一个或多个任务引擎的对应关系,存在对应关系的任两个任务引擎包括的槽位间的对应关系,以及第二槽位与目标关键信息的对应关系,所述一个或多个任务引擎中包括所述第二任务引擎;
可采用以下方式,在用户的共享参数列表中,获取目标关键信息:根据所述第一任务引擎与一个或多个任务引擎的对应关系,确定第二任务引擎;根据所述槽位间的对应关系,在所述共享参数列表中包括的所述第二任务引擎包括的槽位中,确定与所述第一任务引擎的第一槽位存在对应关系的第二槽位;根据所述第二槽位与目标关键信息的对应关系,确定所述目标关键信息。
可选的,所述第二任务引擎用于负责所述用户的第二请求信息,所述第二请求信息与所述第一请求信息发生在预设时间内,且所述第二请求信息与所述第一请求信息发生在预设的对话轮数之内,一轮对话可具体指用户输入一请求信息,人机交互系统反馈一信息的过程。
在本申请实施例中,将共享参数列表与请求信息的发生时间相结合,确定第二任务引擎,可较精准的定位第二任务引擎,提升对第二槽位填充的目标关键信息的准确性。
在一种可能的实现中,所述共享参数列表中存储有第二槽位与目标关键信息的对应关系,可采用以下方式,在所述用户的共享参数列表中,获取目标关键信息:通过所述用户的上下文参数,确定与所述第一任务引擎时间相邻的第二任务引擎;在所述第二任务引擎的槽位中,确定与所述第一槽位类型相同的第二槽位;根据所述第二槽位与目标关键信息的对应关系,确定所述目标关键信息。其中,所述第二任务引擎用于负责所述用户的第二请求信息,所述第二任务引擎中设有与所述第一槽位类型相同的槽位,所述时间相邻指所述第二请求信息的发生时间与所述第一请求信息的发生时间相邻。在一种可能的实现中,确定与所述第一槽位类型相同的第二槽位时,当与所述第一槽位类型相同的槽位有两个或两个以上时可以按预设的优先级信息选定。
在本申请实施例中,所述共享参数列表中仅存储一种对应关系,即任务引擎的不同槽位与目标关键信息的对应关系,相对于上述共享参数列表的实现方式(上述共享参数列表中需要存储三种对应关系),采用本申请的方式,可减少共享参数列表所存储信息的数据量,提高访问共享参数列表的速率。
在一种可能的实现中,所述共享参数列表为所述第二任务引擎所独占的共享参数列表,所述第二任务引擎所独占的共享参数列表中可包括第二槽位与关键信息的对应关系。可采用以下方式,在所述用户的共享参数列表中,获取目标关键信息:根据所述第一任务引擎与所述第二任务引擎的相关性,确定第二任务引擎,所述第一任务引擎与所述第二任务引擎的相关性大于预设阈值;根据所述第二任务引擎,确定所述第二任务引擎所独占的共享参数列表;在所述第二任务引擎所独占的共享参数列表中,查找与所述第一槽位的槽位类型相同的槽位作为所述第二槽位;根据所述第二槽位与目标关键信息对的应关系,确定所述目标关键信息。
在本申请实施例中,人机交互系统的每个任务引擎各自维护一张共享参数表列,该共享参数列表中存储有该任务引擎的槽位与目标关键信息的对应关系,相对应于上述两种共享参数列表,一种为维护三种对应关系,另一种为存储多个任务引擎的槽位与目标关键信息的对应关系,采用本申请的共享参数列表,可进一步减少共享参数列表所存储信息的数据量,进一步提高访问共享参数列表的速率。
在本申请实施例中,所述不满足条件的关键信息包括指代信息,且在所述共享参数列表中包括的所述第二槽位与目标关键信息的对应关系中,所述第二槽位所对应的目标关键 信息是唯一的,且第二槽位所对应的关键信息是可更新的。
在一种可能的实现方式中,可采用以下方式更新第二槽位所对应的目标关键信息,具体为:接收用户输入的第三请求信息,所述第三请求信息由所述第二任务引擎负责;在所述第三请求信息中,提取所述第二任务引擎的第二槽位所对应的关键信息;根据在所述第三请求信息中,提取的第二槽位所对应的关键信息,更新所述共享参数列表中的第二槽位与目关键信息的对应关系中,所述第二槽位所对应的目标关键信息。
在本申请实施例中,第一任务引擎中可设有多个槽位,可采用以下方式,在多个槽位中确定第一槽位,具体为:确定所述第一任务引擎所设有的多个槽位,所述多个槽位具有优先级信息;根据所述多个槽位的优先级信息,确定所述第一槽位。
在本申请的一示例中,所述槽位类型可具体为时间、地点、行为或人物等。
在可能的实现中,所述共享参数列表可以存储在本端,比如所述信息处理装置中的存储器,也可以由通信网络或存取接口从外部存储设备中获取。
第二方面,本申请提供一种信息处理装置,用于终端设备或服务器,包括:包括用于执行以上第一方面各个步骤的单元或手段(means)。
第三方面,本申请提供一种信息处理装置,用于终端设备或服务器,包括至少一个处理元件和至少一个存储元件,其中所述至少一个存储元件用于存储程序和数据,所述至少一个处理元件用于执行本申请第一方面种提供的方法。
第四方面,本申请提供一种计算机可读存储介质,所述计算机可读存储介质中存储有指令,当其在计算机上运行时,使得计算机执行上述第一方面所述的方法。
第五方面,本申请提供一种计算机程序,所述计算机程序包括计算机指令,当所述计算机指令被计算机执行时,使得所述计算机执行第一方面所述的方法。
第六方面,本申请提供一种人机交互系统,包括中控模块和任务引擎模块,其中中控模块由本申请所提供的信息处理装置实现,所述引擎模块用于实现第一任务引擎和第二任务引擎的功能。在一种可能的实现中,所述人机交互系统还包括多任务参数分享管理模块,所述共享参数列表可存储在所述分享参数管理模块中。
附图说明
图1为本申请提供的一种人机交互系统;
图2为本申请提供的另一种人机交互系统;
图3为本申请提供的一种信息处理流程;
图4为本申请提供的另一种信息处理流程;
图5为本申请提供的共享参数列表的一示意图;
图6为本申请提供的共享参数列表的另一示意图;
图7为本申请提供的共享参数列表的又一示意图;
图8为本申请提供的又一种信息处理流程;
图9为本申请提供的通信装置的一结构图;
图10为本申请提供的通信装置的又一结构图。
具体实施方式
如图1所示,本申请提供一种人机交互系统,该人机交互系统主要包括中控模块、任 务引擎模块;所述人机交互系统还可以包括任务状态库模块。
其中,中控模块,主要负责将对话请求分发至一个或多个相关的任务引擎模块,然后对多个任务引擎模块返回的结果进行汇总决策,以得到最终的结果。
任务引擎模块,包括多个任务引擎,每个任务引擎主要用于负责某个特定领域的对话业务,比如,对用户输入的请求信息进行解析,以获取满足条件的关键信息。比如,对于订机票任务引擎,可从用户输入的请求信息中,提取满足订机票任务的关键信息,比如,起始地信息、目的地信息以及时间信息等。
任务状态库模块,主要用于用于存储用户对话任务的中间状态。
在本申请中,在任务引擎模块获取满足条件的关键信息后,可由任务引擎模块根据所述关键信息,执行相应的任务,或者,可由中控模块根据所述关键信息,执行相应的任务,或者,可由智能终端根据所述关键信息,执行相应的任务,或者,可在人机交互系统中,开发一新模块,用于根据所述关键信息,执行相应的任务。在本申请中,并不具体限定根据关键信息,执行相应任务的主体。
需要说明的是,所述人机交互系统的功能可以由服务器实现,也可以由终端设备实现,还可以由服务器与终端设备共同实现。
因此,上文所述的人机交互系统主要包括中控模块、任务引擎模块,可以理解为,当所述人机交互系统的功能由服务器实现时,所述服务器中包括中控模块、任务引擎模块,当所述人机交互系统的功能由终端设备实现时,所述终端设备中包括中控模块、任务引擎模块。
在本申请的一示例中,以“任务引擎模块在获取满足条件的关键信息后,将所述关键信息发送至中控模块,由中控模块执行相应任务”为例,详细说明整个人机交互系统的工作过程:
用户输入请求信息至智能终端,其中,所述请求信息可为语音信息,或为文本信息,或为图像信息。所述智能终端将用户的请求信息转发至人机交互系统的中控模块。所述中控模块将所述用户的请求信息转发至任务引擎模块中的一个或多个任务引擎。所述任务引擎可解析所述请求信息,获取满足条件的关键信息,并发送所述满足条件的关键信息至中控模块。所述中控模块可根据任务引擎模块反馈的关键信息,执行相应的任务。同时,所述任务引擎可将相应的解析结果存储至任务状态库模块中的相应任务状态中。
比如,用户A在智能终端中输入“我要预订明天从北京到上海的机票”的请求,那么智能终端可将“用户A要预订明天从北京到上海的机票”的请求发送至中控模块,而中控模块可首先确定所述请求归“订机票任务引擎”所处理,然后将所述“用户A要预订明天从北京到上海的机票”的请求发送至任务引擎模块中的订机票任务引擎,订机票任务引擎可对“用户A要预订明天从北京至上海的机票”的请求信息进行解析,获得以下满足条件的关键信息{意图=“订机票”,时间=“明天”,起始地=“北京”,目的地=“上海”},然后“订机票任务引擎”可将所述关键信息反馈至中控模块,而中控模块可根据“订机票任务引擎”反馈的关键信息为用户A预订“明天从北京至上海的机票”,最后生成处理结果“已经为您预订明天从北京至上海的机票,且订票信息为****”的反馈信息,且发送所述反馈信息至智能终端,而智能终端可通过文字、语音或图像等方式展示所述反馈信息至用户。
通过上述分析可以看出,在图1所示的人机交互系统中,在对用户的任务请求进行处 理时,必须要获取满足执行该任务条件的关键信息,才能对用户的任务请求进行处理,比如,对于订机票的任务请求,必须要获取满足订机票任务的“时间、起始地和目的地”等关键信息,才能根据用户的任务请求,为用户预订机票;对于天气查询任务请求,必须要获取满足天气查询任务的“地点与时间”等关键信息,才能为用户查询天气。如果在某一任务请求中,提取不到关键信息,或能提取到关键信息,但提取到的关键信息不满足条件,人机交互系统将会反馈“处理失败,请重新输入信息”等提示,而用户需要重新输入请求信息,从而使得人机交互的效率低下。比如,如果用户首先输入“我要预订明天从北京至上海的机票”的请求信息,然后,再输入“那边天气怎么样”的请求信息,由于在对“那边天气怎么样”的请求信息进行处理时,提取不到满足天气查询任务的天气与地点等关键信息,结果会导致人机交互系统反馈“请问您想问那个城市的天气情况”,从而使得用户的体验较差,人机交互的效率低下。
基于以上场景,如图2所示,本申请提供一种人机交互系统,至少包括:中控模块、任务引擎模块以及多任务参数分享管理模块。可选的,还可包括:任务状态库模块。
其中,中控模块、任务引擎模块以及任务状态库模块的相应功能可参见上述图1中的记载所示,在此不再赘述。
所述多任务参数分享管理模块中可存储有共享参数列表,所述共享参数列表的数量可为一个,所述人机交互系统中的n个任务引擎可共享该共享参数列表。或者,所述共享参数列表的数量也可为多个,人机交互系统中的每个任务引擎独占一个共享参数列表。比如,整个人机交互系统包括n个任务引擎,那么,多任务参数分享管理模块中也可包括n个共享参数列表,n个任务引擎中每个任务引擎独占n个共享参数列表中的一个共享参数列表。
可选的,所述多任务参数管理模块中还可包括多任务参数分享管理单元,用于对共享参数列表进行管理。
在本申请实施例中,所述共享参数列表可根据用户的历史对话参数所生成。在本申请实施例中,当任务引擎模块在用户输入的请求信息中,提取不到关键信息,或可提取到关键信息但提取到的关键信息不满足条件时,可在共享参数列表中,获取满足条件的目标关键信息,无需用户重复输入,用户体验较好,人机交互的效率有效提高。
基于图2所示的人机交互系统,本申请提供一种信息处理方法,该方法的执行主体可为上述图1或图2所示的智能终端,也可为运行上述图1或图2所示人机交互系统的服务器,所述服务器可但不限于为云端服务器。在本申请实施例中,以运行所述人机交互系统的服务器为执行主体举例,如图3所示,详细说明信息处理方法的流程。
步骤S301:服务器接收用户输入的第一请求信息。
在本申请实施例中,所述第一请求信息可为语音信息、文本信息或图像信息等。所述用户可输入第一请求信息至智能终端,所述智能终端可转发所述第一请求信息至服务器。
步骤S302:服务器对接收的第一请求信息进行意图识别,确定负责所述第一请求信息的第一任务引擎。
在本申请实施例中,服务器中的中控模块可对第一请求信息进行意图识别,并根据意图识别结果,确定负责第一请求信息的第一任务引擎。比如,中控模块确定第一请求信息的意图识别结果为订机票,那么可确定负责所述第一请求信息的第一任务引擎为订机票任务引擎。
在本申请实施例中,每个任务引擎中可设有不同的槽位(slot),所述槽位可具体为一个变量,槽位的值(slot view)可具体为该槽位所对应的关键信息。在本申请实施例中,所述槽位也可称为信息槽,所述槽位所对应的关键信息也可称为槽信息。在本申请实施例中,以第一任务引擎中设有第一槽位为例,详细说明本申请的过程。需要说明的是,在本申请实施例中,第一任务引擎、第二任务引擎、第一槽位、第二槽位中的“第一、第二”等词汇,仅用于区分描述的目的,而不能理解为指示或暗示相对重要性,也不能理解为指示或暗示顺序。
步骤S303:服务器根据所述第一槽位在所述第一请求信息中提取关键信息。
比如,第一任务引擎为“订机票任务引擎”,所述第一槽位为“起始地槽位”,那么服务器可在第一请求信息中,提取“起始地槽位”所对应的关键信息。再如,第一槽位为“目的地槽位”,那么服务器可在第一请求信息中,提取“目的地槽位”所对应的关键信息。
步骤S304:若根据所述第一槽位在所述第一请求信息中提取不到关键信息,或者,若根据所述第一槽位在所述第一请求信息中提取到关键信息且提取到的关键信息不满足条件,所述服务器则在所述用户的共享参数列表中,获取目标关键信息。
其中,所述共享参数列表中至少包括第二槽位与目标关键信息的对应关系,所述第二槽位与所述第一槽位具有相同的槽位类型,所述槽位类型是时间、地点、行为或人物等,所述第二槽位为所述第二任务引擎中的槽位,所述第一任务引擎与所述第二任务引擎不同。所述不满足条件的关键信息可具体为指代信息,所述指代信息可具体为在第一请求信息中提取的关键信息为指代词,所述指代词可为“这”、“那”、“这里”、“那里”等指代词。
步骤S305:服务器根据所述目标关键信息,执行相应的任务。
由上可见,在本申请实施例中,当在用户输入的第一请求信息中提取不到关键信息,或可提取到关键信息,但提取的关键信息不满足条件时,可在用户的共享参数列表中获取目标关键信息,并根据所述目标关键信息,执行相应任务。相对于上述图1所示的方案,当提取不到关键信息,或提取的关键信息不满足条件时,即请求用户再次输入方案,采用本申请的方案,用户体验较好,人机交互效率较高。
在本申请的一示例中,第一任务引擎中可设有多个槽位,每个槽位的级别是相同的,可按照任意的顺序,填充每个槽位所对应的关键信息。比如,第一任务引擎中设有“起始地槽位”,“目的地槽位”以及“时间槽位”,可按照“起始地槽位”,“目的地槽位”以及“时间槽位”的顺序,依次填充每个槽位所对应的关键信息,也可按照“时间槽位”,“目的地槽位”以及“起始地槽位”的顺序,依次填充每个槽位所对应的关键信息等。
在本申请的另一示例中,第一任务引擎中可设有多个槽位,每个槽位具有不同的优先级,可根据槽位的优先级,填充每个槽位所对应的关键信息,比如,对于订机票任务引擎中包括三个槽位,分别为“起始地槽位”,“目的地槽位”以及“时间槽位”,且设定“时间槽位”的优先级最高,“起始地槽位”的优先级次高,“目的地槽位”的优先级最低,那么,可首先填充“时间槽位”所对应的关键信息,再填充“起始地槽位”所对应的关键信息,最后再填充“目的地槽位”所对应的关键信息。
在本申请实施例中,填充每个槽位所对应的关键信息的过程,以第一槽位为例,详细说明该过程:首先在用户输入的请求信息中,提取第一槽位所对应的关键信息,如果能提取到关键信息,且提取的关键信息满足条件,且第一槽位的填充完成。如果在请求信息中, 提取不到关键信息,或能提取到关键信息,但提取的关键信息不满足条件,则在用户的共享参数列表中,获取目标关键信息,所述目标关键信息作为所述第一槽位所对应的关键信息。
可选的,在图3所示的流程中,还可包括:服务器根据每个任务引擎的语义识别结果,更新所述用户的共享参数列表,所述语义识别结果包括意图信息以及满足条件的关键信息。
比如,用户A输入的第一请求为“预订明天从北京到上海的机票”,那么对于订机票任务引擎所获取的语义识别结果,可为{意图=“订机票”;时间=“明天”,起始地=“北京”,目的地=“上海”}。其中,意图信息为{意图=“订机票”},满足条件的关键信息为{时间=“明天”,起始地=“北京”,目的地=“上海”}。
在本申请实施例中,当第一请求信息除对应第一任务引擎外,还对于N个其它任务引擎时,比如,用户输入第一请求信息“预订明天从北京至上海的票”,那么人机交互系统的中控模块对用户输入的第一请求信息进行意图识别,发现第一请求信息对应于“订票”,而整个人机交互系统中包括“订机票任务引擎”,“订火车票任务引擎”以及“订汽车票任务引擎”,那么中控模块可将上述第一请求信息,分别发送至“订火车票任务引擎”、“订汽车票任务引擎”以及“订机票任务引擎”。可称为第一请求信息与上述”订火车票任务引擎”、“订汽车票任务引擎”以及“订机票任务引擎”,三个任务引擎相对应。
在本申请实施例中,当第一任务请求与N+1个任务引擎相对应时,所述N+1个任务引擎包括第一任务引擎以及上述其它N个任务引擎,那么,上述N+1个任务引擎可将各自针对第一请求信息的语义识别结果发送至中控模块,所述中控模块可从上述N+1个语义识别结果中选择满足规则的语义识别结果,比如最优的语义识别结果,更新所述用户的共享参数列表。
如图4所示,在本申请的一示例中,以上述图3所示的服务器包括中控模块、任务引擎模块和多任务参数分享管理模块为例,详细说明上述图3所示的信息处理流程:
步骤S401:智能终端接收用户输入的第一请求信息,发送所述第一请求信息至人机交互系统的中控模块。
步骤S402:中控模块对所述第一请求信息进行意图识别,获取与该第一请求信息相关的一个或多个任务引擎,并发送所述第一请求信息至任务引擎模块中的一个或多个任务引擎。
步骤S403:针对一个任务引擎,确定该任务引擎所对应的槽位。
其中,不同的任务引擎可对应不同的槽位,比如,对于订机票任务引擎所对应的槽位可为“起始地槽位”、“目的地槽位”或“地点槽位”。对于天气查询任务引擎所对应的槽位可为“地点槽位”或“时间槽位”。
步骤S404:任务引擎根据所述该任务引擎所对应的槽位,在第一请求信息中提取关键信息。
比如,对于天气查询任务引擎,其所对应的槽位可具体包括“地点槽位”和“时间槽位”,所述天气查询任务引擎,可在用户输入的第一请求信息中,依次提取“地点槽位”和“时间槽位”所对应的关键信息。再如,用户输入的第一请求信息可为“查询北京明天的天气状况”,那么天气查询任务引擎可依次在上述第一请求信息中,提取“地点槽位”所对应的关键信息“北京”,提取“时间槽位”所对应的关键信息“明天”。
步骤S405:针对一槽位,如果可在第一请求信息中提取到关键信息,且提取的关键信息满足条件,则执行步骤S409,否则执行步骤S406。
步骤S406:任务引擎发送读取指令,至多任务参数分享管理模块。
步骤S407:多任务参数分享管理模块,从所述用户的共享参数列表中,获取目标关键信息。
步骤S408:多任务参数分享管理模块,发送目标关键信息至任务引擎。
步骤S409:任务引擎获取针对第一请求信息的语义识别结果。
其中,所述语义识别结果包括针对第一请求信息的意图信息以及满足条件的关键信息。
步骤S410:任务引擎将针对第一请求信息的语义识别结果发送至中控模块。
步骤S411:中控模块根据任务引擎发送的语义识别结果更新所述多任务参数分享管理模块中用户的共享参数列表。
需要说明的是,在上述图4所示的流程中,是以任务引擎模块在共享参数列表中,读取信息,中控模块更新共享参数列表为例进行说明的。在本申请实施例中,并不限定在共享参数列表中读取信息的主体,以及更新共享参数列表的主体。比如,在本申请中,中控模块可负责从共享参数列表中读取信息,以及更新共享参数列表。或者,任务引擎模块负责从共享参数列表中读取信息,以及更新共享参数列表。或者,中控模块负责从共享参数列表中读取信息,任务引擎模块负责更新共享参数列表。或者,任务引擎模块负责从共享参数列表中读取信息,中控模块负责更新共享参数列表。
在本申请实施例中,将详细说明,在上述图3或图4所示流程中,“如何根据用户的共享参数列表,获取目标关键信息”。
其中,本申请实施例提供以下两种共享参数列表,分别为第一种共享参数列表和第二种共享参数列表。利用所述第一种共享参数列表或第二种共享参数列表,均可获取目标关键信息。
第一种共享参数列表:共享参数列表的数量为一个,人机交互系统中的所有任务引擎共享该共享参数列表,也即上述多任务参数分享管理模块中仅存储一个共享参数列表。
在本申请的一示例中,针对第一种共享参数列表,所述第一种共享参数列表中可具体包括:第一任务引擎与一个或多个任务引擎的对应关系,存在对应关系的任两个任务引擎包括的槽位间的对应关系,以及第二槽位与目标关键信息的对应关系。所述一个或多个任务引擎中包括所述第二任务引擎。针对第一种共享参数列表,可具体采用以下方式,获取目标关键信息:
根据所述第一任务引擎与一个或多个任务引擎的对应关系,确定第二任务引擎;
可选的,所述第二任务引擎可用于负责所述用户的第二请求信息,所述第二请求信息与所述第一请求信息可发生在预设时间内,且所述第二请求信息与所述第一请求信息发生在预设对话轮数内。
根据所述槽位间的对应关系,在所述共享参数列表中包括的所述第二任务引擎包括的槽位中,确定与所述第一任务引擎的第一槽位存在对应关系的第二槽位;
根据所述第二槽位与目标关键信息的对应关系,确定所述目标关键信息。
在本示例中,所述第一种共享参数列表中可具体存储有三个对应关系,分别为:不同任务引擎间的对应关系,以及如果两个任务引擎间存在对应关系,两个任务引擎所包括槽 位的对应关系,以及针对每个任务引擎所包括的槽位与目标关键信息的对应关系。
比如,如图1所示,以天气查询任务引擎为例,获取目标关键信息的过程可为:首先获取在预设时间(比如10分钟)和预设对话轮数(比如5轮对话,所述1轮对话的过程,具体包括用户输入一请求信息,人机交互系统反馈一信息),与用户存在交互的任务引擎,比如,存在交互的任务引擎可包括{机票预定任务引擎、订外卖任务引擎}。然后,获取在共享参数列表中,与天气查询任务引擎存在对应关系的任务引擎,比如,可为{订机票任务引擎、订火车票任务引擎、酒店预定任务引擎}。那么两者取交集,即将{机票预定任务引擎、订外卖任务引擎}与{订机票任务引擎、订火车票任务引擎、酒店预定任务引擎}取交集,可以看出,两者的交集为{订机票任务引擎}。然后,在机票预定任务引擎的槽位中,查询与天气查询任务引擎的地点槽位存在对应关系的槽位,通过图5可以看出,与天气查询任务引擎的地点槽位存在对应关系的槽位,为订机票任务引擎的目的地槽位;最后将目的地槽位所对应的目标关键信息即上海,作为天气查询任务引擎的地点槽位的关键信息。
在本申请的另一示例中,针对第一种共享参数列表,所述第一种共享参数列表中可具体包括第二槽位与目标关键信息的对应关系。针对第一种共享参数列表,可具体采用以下方式,确定目标关键信息:
通过所述用户的上下文参数,确定与所述第一任务引擎时间相邻的第二任务引擎;
其中,所述第二任务引擎用于负责所述用户的第二请求信息,所述第二任务引擎中设有与所述第一槽位类型相同的槽位,所述时间相邻指所述第二请求信息的发生时间与所述第一请求信息的发生时间相邻;
在所述第二任务引擎的槽位中,确定与所述第一槽位类型相同的第二槽位;
根据所述第二槽位与目标关键信息的对应关系,确定所述目标关键信息。
需要说明的是,在本申请示例中,所述第一任务引擎与第二任务引擎时间相邻,可具体包括以下两种情况,第一种情况:所述第一任务引擎所负责的第一任务请求,与所述第二任务引擎所负责的第二任务请求,在时间上相邻,即用户输入的第一任务请求与第二任务请求在时间上相邻,比如,用户首先在人机交互系统中输入第一任务请求“预定北京至上海的机票”,然后在人机交互系统中输入第二任务请求“那边天气怎么样”。可认为上述第一任务请求与上述第二任务请求在时间上相邻。同时,设定上述“第一任务请求”由“机票预定”任务引擎负责,上述“第二任务请求”由“天气查询”任务引擎负责,那么可认定上述“机票预定”任务引擎与“天气查询”任务引擎在时间上相邻。第二种情况:第一任务引擎所负责的第一任务请求,与所述第二任务引擎所负责的第二任务请求,在时间上相隔预设时间,且两者在预定对话轮数之内。
在本示例中,整个共享参数列表中,仅存储有一个对应关系,即每个任务引擎的槽位与关键信息的对应关系。
仍以填充天气查询任务引擎的地点槽位为例,进行说明,获取目标关键信息的过程可为:首先在用户的上下文参数中,获取与天气查询任务引擎相邻近的任务引擎,即用户在与天气查询任务引擎交互前,与之交互的任务引擎,比如,如图6所示,与天气查询任务引擎时间上相邻近的任务引擎为机票预定任务引擎。然后,在机票预定任务引擎的槽位中,查询与地点槽位类型相同的槽位,比如查询出的类型相同的槽位为目的地槽位,最后,用共享参数列表中的订机票任务引擎的目的地槽位所对应的关键信息,即上海,作为天气查 询领域的地点槽位所对应的关键信息。
第二种共享参数列表:共享参数列表的数量为多个,每个任务引擎独占一个共享参数列表。比如,整个人机交互系统,包括N个任务引擎,那么多任务参数分享管理模块中可存储N个共享参数列表,N个任务引擎中的每个任务引擎独占N个共享参数列表中的一个共享参数列表。
针对第二种共享参数列表,所述共享参数列表中可具体存储第二槽位与目标关键信息的对应关系。可采用以下方式,获取目标关键信息,具体为:获得所述第一请求信息所对应的第一任务引擎;获得第二任务引擎,其中,所述第二任务引擎对应于第二垂域,所述第一任务引擎对应于第一垂域,所述第一垂域与所述第二垂域的相关性大于预设阈值,所述第一垂域与所述第二垂域的相关性通过历史数据聚类所获得的;在所述第二任务引擎所独占的共享参数列表中,查找与第一槽位的槽位类型相同的槽位作为所述第二槽位,根据所述第二槽位与目标关键信息对的应关系确定目标关键信息。
在本申请的一示例中,比如对于订机票任务引擎所维护的共享参数列表,可如图7所示。
在本申请实施例中,可根据用户与人机交互系统的历史对话参数,生成所述共享参数列表。比如,用户输入“预定今天北京至上海的机票”的请求信息至人机交互系统,那么机票预定任务引擎,可在上述请求信息中,分别提取“起始地槽位”所对应的关键信息“北京”,“目的地槽位”所对应的关键信息“上海”,“时间槽位”所对应的关键信息“今天”。最后,根据不同槽位所对应的关键信息,更新所述订机票任务引擎所独占的共享参数列表。
在本申请实施例中,仍沿用上述举例,用户首先在人机交互系统中,输入“预定今天北京至上海的机票”的请求信息,且生成上述图7所示的共享参数列表。然后用户在人机交互系统中,再次输入“那边天气怎么样”的请求信息,可以发现该请求信息归属天气查询任务引擎所处理,然后天气查询任务引擎,可在用户输入的请求信息中,分别提取“地点槽位”和“时间槽位”所对应的关键信息。可见发现,对于“地点槽位”所提取到的关键信息“那边”为指代信息,对于“时间槽位”提取不到的关键信息,即对于“时间槽位”缺失信息。在本申请实施例中,可设定“缺失信息”和“指代信息”均不满足条件,此时,可采用以下方式,获取目标关键信息:
获取天气查询任务引擎所对应的第一垂域;
获取与第一垂域的相关性大于预设阈值的第二垂域。
其中,第一垂域与第二垂域的相关性可通过历史数据聚类所获得的。
获取第二垂域所对应的第二任务引擎。
在第二任务引擎所所独占的共享参数列表中,获取目标关键信息。
在本申请中,可设定第二任务引擎为订机票任务引擎,那么,可在上述图7所示的共享参数列表中,获取满足条件的关键信息。
比如,在本申请的一示例中,可在上述图7所示的共享参数列表中,查询天气任务引擎中“地点槽位”所对应的满足条件的关键信息,通过上述表7可以看出,“上海”为天气查询任务引擎中“地点槽位”所对应的满足条件的关键信息。同理,可在上述图7所示的共享参数列表中,查询天气任务引擎的“时间槽位”所对应的关键信息,通过上述表7可以看出,“今天”为天气查询任务引擎中“时间槽位”所对应的满足条件的关键信息。
通过以上分析可以看出,在上述第一种共享参数列表和第二种共享参数列表中,均存储有第二槽位与目标关键信息的对应关系。在本申请实施例中,所述第二槽位所对应的关键信息是唯一的,且可更新的。
在本申请的一示例中,可采用以下方式,更新第二槽位所对应的关键信息:服务器接收用户输入的第三请求信息,所述第三请求信息由所述第二任务引擎负责;服务器在所述第三请求信息中,提取所述第二任务引擎的第二槽位所对应的关键信息;服务器根据在所述第三请求信息中,提取的第二槽位所对应的关键信息,更新所述共享参数列表中的第二槽位与目关键信息的对应关系中,所述第二槽位所对应的目标关键信息。
针对上述的第一种共享参数列表和第二种共享参数列表,如图8所示,本申请提供了信息处理的另一流程,包括:
在图8所示的流程中,以用户与人机交互系统进行三轮对话为例,详细进行说明:
第一轮对话:用户输入“北京今天的天气怎么样”的请求信息至人机交互系统,人机交互系统可对用户输入的请求信息进行意图识别,确定该请求信息归属天气查询任务引擎进行处理。天气查询任务引擎的地点槽位,可在上述请求信息中提取关键信息“北京”。天气查询任务引擎的时间槽位,可在上述请求信息中提取关键信息“今天”,最后生成的历史参数列表,可参见下述表1所示。
用户ID 意图 时间 地点
*** 查天气 今天 北京
表1
第二轮对话:用户继续输入“订一张明天北京至布宜诺斯艾利斯的机票”的请求信息至人机交互系统,同理,人机交互系统可对用户的请求信息进行意图识别,确定该请求信息归属机票预定任务引擎所处理。然后机票预定任务引擎可在请求信息中,分别提取“起始地槽位”所对应的关键信息“北京”,“目的地槽位”所对应的关键信息“布宜诺斯艾利斯”,时间槽位所对应的关键信息为“明天”。最后,生成的历史参数列表,可参见下述表2所示。
用户ID 意图 出发地 目的地 时间
*** 订机票 北京 布宜诺斯艾利斯 明天
表2
在本申请实施例中,可将上述表1所示的历史参数列表和上述表2所示的历史参数列表存储于第一种共享参数列表中,也可存储于第二种共享参数列表中。且如果存储于第二共享参数列表中,对于上述表1所示的历史参数列表可为天气查询任务引擎所独占的共享参数列表,对于上述表2所示的历史参数列表可为机票预定任务引擎所独占的共享参数列表。
第三轮对话:用户输入“那边的天气怎么样”的请求信息至人机交互系统。
步骤S801:接收用户输入的第三轮对话的请求信息,“那边的天气怎么样”。
步骤S802:在第三轮对话的请求信息中,提取关键信息。
其中,提取的关键信息,可参见下述表3所示。
用户ID 意图 时间 地点
*** 查询天气 那边
表3
步骤S803:获得在天气查询场景下,缺少的必要信息槽集合L0={“时间”,“地点”}。
可选的,在步骤S803之后,还可包括:步骤S804、步骤S805、步骤S806以及步骤S807。
步骤S804:判断信息槽的填充是否有预定义规则;如果否,执行步骤S805;如果是,执行步骤S808。
其中,所述预定义规则是指是否可采用利用共享参数列表,填充信息槽;
步骤S805:判断缺失的信息槽是否可利用传感器补充关键信息。如果是,执行步骤S806;否则,执行步骤S807;
步骤S806:用传感器填充“地点信息”。
步骤S807:提示用户输入“地点信息”。
在本申请实施例中,设置步骤S804至步骤S807的目的是,可为信息槽的填充,设置不同的方式,比如可采用共享参数列表的方式,也可采用传感器的方式,也可采用用户手动输入的方式。
而在本申请实施例中,也可在步骤S803之后,直接执行步骤S808,即直接设置采用共享参数列表的方式,填充信息槽的关键信息。
步骤S808:据预定义规则,逐个判断L0的槽位在场景开发时是否预定义到共享参数列给中的某个值,且如果是的话,加入集合L1={“时间”,“地点”}。
步骤S809:L1非空时,根据L1中的每个槽位的优先级,得到集合L2={“地点”,“时间”}。
步骤S810:根据用户ID在共享参数列表中,查询用户的历史“地点”参数。
步骤S811:判断历史“地点”参数是否有值;如果没有值,执行步骤S812;否则执行步骤S813。
步骤S812:提示用户输入“地点”信息。
步骤S813:将历史“地点”参数的值填充到“地点”槽位。
在本申请实施例中,可将上述表2所示的“目的地:布宜诺斯艾利斯”填充到“地点”槽位,生成表4所示的关键信息。
用户ID 意图 时间 地点
*** 查询天气 布宜诺斯艾利斯
表4
步骤S814:根据用户ID在共享参数列表中,查询用户的历史“时间”参数。
步骤S815:判断历史“时间”参数是否有值。如果有值,执行步骤S816;否则执行步骤S817。
步骤S816:将历史“时间”参数的值填充到“时间”槽位。
在本申请实施例中,可将上述表2所示的“时间:明天”填充到时间槽位,生成表5所示的关键信息。
用户ID 意图 时间 地点
*** 查询天气 明天 布宜诺斯艾利斯
表5
步骤S817:提示用户输入“时间”信息。
通过以上方式,可以看出在本申请实施例中,在用户输入的请求信息中提取不到满足条件的关键信息时,可利用用户的历史对话参数,获取到满足条件的关键信息,从而用户无需再次输入请求信息,用户体验较好,人机交互效率较高。
与上述构思相同,如图9所示,本申请提供一种信息处理装置900,该信息处理装置900可具体为终端设备或服务器等,包括:
其中,收发单元901,用于接收用户输入的第一请求信息;
需要说明的是,所述收发单元901可以是硬件装置如收发器,也可以是逻辑单元如数据接口,在本申请中不对此做限定。
处理单元902,用于对所述第一请求信息进行意图识别,确定负责所述第一请求信息的第一任务引擎,所述第一任务引擎中设有第一槽位,根据所述第一槽位在所述第一请求信息中提取关键信息,若根据所述第一槽位在所述第一请求信息中提取不到关键信息,或者,若根据所述第一槽位在所述第一请求信息中提取到关键信息且提取到的关键信息不满足条件,则在所述用户的共享参数列表中,获取目标关键信息;
其中,所述共享参数列表中至少包括第二槽位与目标关键信息的对应关系,所述第二槽位与所述第一槽位具有相同的槽位类型,所述第二槽位为所述第二任务引擎中的槽位,所述第一任务引擎与所述第二任务引擎不同。
关于收发单元901和处理单元902的具体介绍,可参数上述图3所示的信息处理方法的介绍,在此不再赘述。
与上述构思相同,如图10所示,本申请实施例还提供一种信息处理装置100的结构,信息处理装置100可具体为终端设备或服务器,可选的,所述信息处理装置100可以具体为所述服务器中的中控模块,或者其功能由所述服务器中的中控模块实现;进一步可选的,所述信息处理装置100可以具体为所述终端设备中的中控模块,或者其功能由所述终端设备中的中控模块实现。如图10所示,信息处理装置100中可以包括通信接口101、处理器102。信息处理装置100中还可以包括存储器103。存储器103可以设置于信息处理装置100的内部,还可以设置于信息处理装置的外部。上述图9中所示的处理单元902可以由处理器102实现。收发单元901可以由通信接口101实现。处理器102通过通信接口101接收业务数据。
本申请实施例中通信接口101可以是电路、总线、收发器或者其它任意可以用于进行信息交互的装置。
本申请实施例中处理器102可以是通用处理器、数字信号处理器、专用集成电路、现场可编程门阵列或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件,可以实现或者执行本申请实施例中的公开的图3、图4或图8所示的信息处理方法。通用处理器可以是微处理器或者任何常规的处理器等。结合本申请实施例所公开的方法的步骤可以直接体现为硬件处理器执行完成,或者用处理器中的硬件及软件单元组合执行完成。处理器102用于实现上述方法所执行的程序代码可以存储在存储器103中。存储器103和处理器102耦合。处理器102可能和存储器103协同操作。存储器103可以是非易失性存储 器,比如硬盘(hard disk drive,HDD)或固态硬盘(solid-state drive,SSD)等,还可以是易失性存储器(volatile memory),例如随机存取存储器(random-access memory,RAM)。存储器103是能够用于携带或存储具有指令或数据结构形式的期望的程序代码并能够由计算机存取的任何其他介质,但不限于此。
本申请实施例中不限定上述通信接口101、处理器102以及存储器103之间的具体连接介质。本申请实施例在图10中以存储器103、处理器102以及通信接口101之间通过总线连接,总线在图10中以粗线表示,其它部件之间的连接方式,仅是进行示意性说明,并不引以为限。所述总线可以分为地址总线、数据总线、控制总线等。为便于表示,图10中仅用一条粗线表示,但并不表示仅有一根总线或一种类型的总线。
基于以上实施例,本申请实施例还提供了一种计算机存储介质,该存储介质中存储软件程序,该软件程序在被一个或多个处理器读取并执行时可实现上述任意一个或多个实施例提供的方法。所述计算机存储介质可以包括:U盘、移动硬盘、只读存储器、随机存取存储器、磁碟或者光盘等各种可以存储程序代码的介质。
基于以上实施例,本申请提供一种计算机程序,所述计算机程序包括计算机指令,当所述计算机指令被计算机执行时,使得所述计算机执行上述任意一个或多个实施例提供的方法。
本领域内的技术人员应明白,本申请的实施例可提供为方法、系统、或计算机程序产品。因此,本申请可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且,本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请是参照根据本申请的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器,使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中,使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品,该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上,使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理,从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
显然,本领域的技术人员可以对本申请进行各种改动和变型而不脱离本申请的范围。这样,倘若本申请的这些修改和变型属于本申请权利要求及其等同技术的范围之内,则本申请也意图包含这些改动和变型在内。

Claims (22)

  1. 一种信息处理方法,其特征在于,包括:
    接收用户输入的第一请求信息;
    确定负责所述第一请求信息的第一任务引擎,所述第一任务引擎中设有第一槽位;
    根据所述第一槽位在所述第一请求信息中提取关键信息;
    若根据所述第一槽位在所述第一请求信息中提取不到关键信息,或者,若根据所述第一槽位在所述第一请求信息中提取到关键信息且提取到的关键信息不满足条件,则在所述用户的共享参数列表中,获取目标关键信息;
    其中,所述共享参数列表中至少包括第二槽位与目标关键信息的对应关系,所述第二槽位与所述第一槽位具有相同的槽位类型,所述第二槽位为第二任务引擎中的槽位,所述第一任务引擎与所述第二任务引擎不同。
  2. 根据权利要求1所述的方法,其特征在于,所述共享参数列表中还包括所述第一任务引擎与一个或多个任务引擎的对应关系,以及存在对应关系的任两个任务引擎包括的槽位间的对应关系,所述一个或多个任务引擎中包括所述第二任务引擎;
    所述在所述用户的共享参数列表中,获取目标关键信息,包括:
    根据所述第一任务引擎与一个或多个任务引擎的对应关系,确定第二任务引擎;
    根据所述槽位间的对应关系,在所述共享参数列表中包括的所述第二任务引擎包括的槽位中,确定与所述第一任务引擎的第一槽位存在对应关系的第二槽位;
    根据所述第二槽位与目标关键信息的对应关系,确定所述目标关键信息。
  3. 根据权利要求1或2所述的方法,其特征在于,所述第二任务引擎用于负责所述用户的第二请求信息,所述第二请求信息与所述第一请求信息发生在预设时间内。
  4. 根据权利要求1所述的方法,其特征在于,所述在所述用户的共享参数列表中,获取目标关键信息,包括:
    通过所述用户的上下文参数,确定与所述第一任务引擎时间相邻的第二任务引擎;
    其中,所述第二任务引擎用于负责所述用户的第二请求信息,所述第二任务引擎中设有与所述第一槽位类型相同的槽位,所述时间相邻指所述第二请求信息的发生时间与所述第一请求信息的发生时间相邻;
    在所述第二任务引擎的槽位中,确定与所述第一槽位类型相同的第二槽位;
    根据所述第二槽位与目标关键信息的对应关系,确定所述目标关键信息。
  5. 根据权利要求1所述的方法,其特征在于,所述共享参数列表为所述第二任务引擎所独占的共享参数列表;
    所述在所述用户的共享参数列表中,获取目标关键信息,包括:
    根据所述第一任务引擎与所述第二任务引擎的相关性,确定第二任务引擎,所述第一任务引擎与所述第二任务引擎的相关性大于预设阈值;
    根据所述第二任务引擎,确定所述第二任务引擎所独占的共享参数列表;
    在所述第二任务引擎所独占的共享参数列表中,查找与所述第一槽位的槽位类型相同的槽位作为所述第二槽位;
    根据所述第二槽位与目标关键信息对的应关系,确定所述目标关键信息。
  6. 根据权利要求1至5任一项所述的方法,其特征在于,所述不满足条件的关键信息包括指代信息。
  7. 根据权利要求1至6任一项所述的方法,其特征在于,在所述共享参数列表中包括的所述第二槽位与目标关键信息的对应关系中,所述第二槽位所对应的目标关键信息是唯一的。
  8. 根据权利要求1至7任一项所述的方法,其特征在于,所述方法还包括:
    接收用户输入的第三请求信息,所述第三请求信息由所述第二任务引擎负责;
    在所述第三请求信息中,提取所述第二任务引擎的第二槽位所对应的关键信息;
    根据在所述第三请求信息中,提取的第二槽位所对应的关键信息,更新所述共享参数列表中的第二槽位与目关键信息的对应关系中,所述第二槽位所对应的目标关键信息。
  9. 根据权利要求1至8任一项所述的方法,其特征在于,在所述根据所述第一槽位在所述第一请求信息中提取关键信息之前,所述方法还包括:
    确定所述第一任务引擎所设有的多个槽位,所述多个槽位具有优先级信息;
    根据所述多个槽位的优先级信息,确定所述第一槽位。
  10. 根据权利要求1至9任一项所述的方法,其特征在于,所述槽位类型是时间、地点、行为或人物。
  11. 一种信息处理装置,其特征在于,包括:
    收发单元,用于接收用户输入的第一请求信息;
    处理单元,用于确定负责所述第一请求信息的第一任务引擎,所述第一任务引擎中设有第一槽位,根据所述第一槽位在所述第一请求信息中提取关键信息,若根据所述第一槽位在所述第一请求信息中提取不到关键信息,或者,若根据所述第一槽位在所述第一请求信息中提取到关键信息且提取到的关键信息不满足条件,则在所述用户的共享参数列表中,获取目标关键信息;
    其中,所述共享参数列表中至少包括第二槽位与目标关键信息的对应关系,所述第二槽位与所述第一槽位具有相同的槽位类型,所述第二槽位为第二任务引擎中的槽位,所述第一任务引擎与所述第二任务引擎不同。
  12. 根据权利要求11所述的装置,其特征在于,所述共享参数列表中还包括第一任务引擎与一个或多个任务引擎的对应关系,以及存在对应关系的任两个任务引擎包括的槽位间的对应关系,所述一个或多个任务引擎中包括所述第二任务引擎;
    所述处理单元在所述用户的共享参数列表中,获取目标关键信息时,具体用于:
    根据所述第一任务引擎与一个或多个任务引擎的对应关系,确定第二任务引擎;
    根据所述槽位间的对应关系,在所述共享参数列表中包括的所述第二任务引擎包括的槽位中,确定与所述第一任务引擎的第一槽位存在对应关系的第二槽位;
    根据所述第二槽位与目标关键信息的对应关系,确定所述目标关键信息。
  13. 根据权利要求11或12所述的装置,其特征在于,所述第二任务引擎用于负责所述用户的第二请求信息,所述第二请求信息与所述第一请求信息发生在预设时间内。
  14. 根据权利要求11所述的装置,其特征在于,所述处理单元在所述用户的共享参数列表中,获取目标关键信息时,具体用于:
    通过所述用户的上下文参数,确定与所述第一任务引擎时间相邻的第二任务引擎;
    其中,所述第二任务引擎用于负责所述用户的第二请求信息,所述第二任务引擎中设有与所述第一槽位类型相同的槽位,所述时间相邻指所述第二请求信息的发生时间与所述第一请求信息的发生时间相邻;
    在所述第二任务引擎的槽位中,确定与所述第一槽位类型相同的第二槽位;
    根据所述第二槽位与目标关键信息的对应关系,确定所述目标关键信息。
  15. 根据权利要求11所述的装置,其特征在于,所述共享参数列表为所述第二任务引擎所独占的共享参数列表;
    所述处理单元在所述用户的共享参数列表中,获取目标关键信息,具体用于:
    根据所述第一任务引擎与所述第二任务引擎的相关性,确定第二任务引擎,所述第一任务引擎与所述第二任务引擎的相关性大于预设阈值;
    根据所述第二任务引擎,确定所述第二任务引擎所独占的共享参数列表;
    在所述第二任务引擎所独占的共享参数列表中,查找与所述第一槽位的槽位类型相同的槽位作为所述第二槽位;
    根据所述第二槽位与目标关键信息对的应关系,确定所述目标关键信息。
  16. 根据权利要求11至15任一项所述的装置,其特征在于,所述不满足条件的关键信息包括指代信息。
  17. 根据权利要求11至16任一项所述的装置,其特征在于,在所述共享参数列表中包括的所述第二槽位与目标关键信息的对应关系中,所述第二槽位所对应的目标关键信息是唯一的。
  18. 根据权利要求11至17任一项所述的装置,其特征在于,所述收发单元,还用于:接收用户输入的第三请求信息,所述第三请求信息由所述第二任务引擎负责;
    所述处理单元,还用于:在所述第三请求信息中,提取所述第二任务引擎的第二槽位所对应的关键信息;根据在所述第三请求信息中,提取的第二槽位所对应的关键信息,更新所述共享参数列表中的第二槽位与目关键信息的对应关系中,所述第二槽位所对应的目标关键信息。
  19. 根据权利要求11至18任一项所述的装置,其特征在于,所述处理单元还用于:
    确定所述第一任务引擎所设有的多个槽位,所述多个槽位具有优先级信息;
    根据所述多个槽位的优先级信息,确定所述第一槽位。
  20. 根据权利要求11至19任一项所述的装置,其特征在于,所述槽位类型是时间、地点、行为或人物。
  21. 一种信息处理装置,其特征在于,包括处理器和存储器;
    所述存储器用于存储计算机执行指令;
    所述处理器用于执行所述存储器所存储的计算机执行指令,以使所述信息处理装置执行如权利要求1至10任一项所述的方法。
  22. 一种计算机可读存储介质,其特征在于,所述计算机可读存储介质中存储有指令,当其在计算机上运行时,使得计算机执行如权利要求1至10任一项所述的方法。
PCT/CN2018/122196 2018-04-26 2018-12-19 一种信息处理方法及装置 WO2019205679A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP18915826.4A EP3779725A4 (en) 2018-04-26 2018-12-19 PROCESS AND DEVICE FOR PROCESSING INFORMATION
US16/868,970 US20200264923A1 (en) 2018-04-26 2020-05-07 Information Processing Method and Apparatus

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810388248.8 2018-04-26
CN201810388248.8A CN109948017B (zh) 2018-04-26 2018-04-26 一种信息处理方法及装置

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/868,970 Continuation US20200264923A1 (en) 2018-04-26 2020-05-07 Information Processing Method and Apparatus

Publications (1)

Publication Number Publication Date
WO2019205679A1 true WO2019205679A1 (zh) 2019-10-31

Family

ID=67006229

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/122196 WO2019205679A1 (zh) 2018-04-26 2018-12-19 一种信息处理方法及装置

Country Status (4)

Country Link
US (1) US20200264923A1 (zh)
EP (1) EP3779725A4 (zh)
CN (2) CN113536093A (zh)
WO (1) WO2019205679A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111091825A (zh) * 2019-12-03 2020-05-01 北京蓦然认知科技有限公司 一种构建交互引擎簇的方法及装置
CN111881270A (zh) * 2020-07-01 2020-11-03 北京嘀嘀无限科技发展有限公司 一种智能对话方法和系统

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113536093A (zh) * 2018-04-26 2021-10-22 华为技术有限公司 一种信息处理方法及装置
CN110798506B (zh) * 2019-09-27 2023-03-10 华为技术有限公司 执行命令的方法、装置及设备
CN111124348B (zh) * 2019-12-03 2023-12-05 光禹莱特数字科技(上海)有限公司 一种生成交互引擎簇的方法及装置
CN111128153B (zh) * 2019-12-03 2020-10-02 北京蓦然认知科技有限公司 一种语音交互方法及装置
CN111128154B (zh) * 2019-12-03 2022-06-03 杭州蓦然认知科技有限公司 一种聚合形成交互引擎簇的方法及装置
CN111125452B (zh) * 2019-12-26 2024-02-20 深圳市旦米科技有限公司 一种基于信息关联的语音交互方法、装置
CN111061453B (zh) * 2019-12-26 2023-12-26 北京官匠空间科技有限公司 一种用于app生态系统的语音交互方法、装置
CN111124866B (zh) * 2019-12-26 2023-12-08 光禹莱特数字科技(上海)有限公司 一种语音交互的方法和装置
CN111048088A (zh) * 2019-12-26 2020-04-21 北京蓦然认知科技有限公司 一种多应用程序的语音交互方法和装置
CN111026628B (zh) * 2019-12-26 2023-06-30 杭州蓦然认知科技有限公司 一种语音交互方法、装置
CN115605861A (zh) * 2020-06-02 2023-01-13 深圳市欢太科技有限公司(Cn) 文本信息的分类方法、移动终端及计算机可读存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105068661A (zh) * 2015-09-07 2015-11-18 百度在线网络技术(北京)有限公司 基于人工智能的人机交互方法和系统
CN105159996A (zh) * 2015-09-07 2015-12-16 百度在线网络技术(北京)有限公司 基于人工智能的深度问答服务提供方法和装置
CN105183850A (zh) * 2015-09-07 2015-12-23 百度在线网络技术(北京)有限公司 基于人工智能的信息查询方法及装置
CN106202127A (zh) * 2015-05-08 2016-12-07 深圳市腾讯计算机系统有限公司 一种垂直搜索引擎对检索请求的处理方法及装置
CN107807933A (zh) * 2016-09-09 2018-03-16 阿里巴巴集团控股有限公司 一种针对提问的回复方法和装置

Family Cites Families (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2001356796A (ja) * 2000-06-12 2001-12-26 Atr Onsei Gengo Tsushin Kenkyusho:Kk サービス予約システムおよびサービスを予約する情報端末
US10241752B2 (en) * 2011-09-30 2019-03-26 Apple Inc. Interface for a virtual digital assistant
US9858925B2 (en) * 2009-06-05 2018-01-02 Apple Inc. Using context information to facilitate processing of commands in a virtual assistant
KR101194296B1 (ko) * 2010-09-01 2012-10-24 주식회사 다음커뮤니케이션 관련 검색어 제공 방법 및 시스템
US9842168B2 (en) * 2011-03-31 2017-12-12 Microsoft Technology Licensing, Llc Task driven user intents
EP2575128A3 (en) * 2011-09-30 2013-08-14 Apple Inc. Using context information to facilitate processing of commands in a virtual assistant
KR101593720B1 (ko) * 2011-12-27 2016-02-17 한국전자통신연구원 시맨틱 네트워크 기반 콘텐츠 검색추천 장치 및 방법
US10417037B2 (en) * 2012-05-15 2019-09-17 Apple Inc. Systems and methods for integrating third party services with a digital assistant
CN103870476A (zh) * 2012-12-12 2014-06-18 百度在线网络技术(北京)有限公司 检索方法及设备
CN113268498A (zh) * 2014-07-11 2021-08-17 华为技术有限公司 业务推荐方法和具有智能助手的装置
US9767794B2 (en) * 2014-08-11 2017-09-19 Nuance Communications, Inc. Dialog flow management in hierarchical task dialogs
CN104281698B (zh) * 2014-10-15 2017-07-07 国云科技股份有限公司 一种高效的大数据查询方法
US10482184B2 (en) * 2015-03-08 2019-11-19 Google Llc Context-based natural language processing
US9832504B2 (en) * 2015-09-15 2017-11-28 Google Inc. Event-based content distribution
CN107256267B (zh) * 2017-06-19 2020-07-24 北京百度网讯科技有限公司 查询方法和装置
CN107301227A (zh) * 2017-06-21 2017-10-27 北京百度网讯科技有限公司 基于人工智能的搜索信息解析方法及装置
CN107369443B (zh) * 2017-06-29 2020-09-25 北京百度网讯科技有限公司 基于人工智能的对话管理方法及装置
CN113536093A (zh) * 2018-04-26 2021-10-22 华为技术有限公司 一种信息处理方法及装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106202127A (zh) * 2015-05-08 2016-12-07 深圳市腾讯计算机系统有限公司 一种垂直搜索引擎对检索请求的处理方法及装置
CN105068661A (zh) * 2015-09-07 2015-11-18 百度在线网络技术(北京)有限公司 基于人工智能的人机交互方法和系统
CN105159996A (zh) * 2015-09-07 2015-12-16 百度在线网络技术(北京)有限公司 基于人工智能的深度问答服务提供方法和装置
CN105183850A (zh) * 2015-09-07 2015-12-23 百度在线网络技术(北京)有限公司 基于人工智能的信息查询方法及装置
CN107807933A (zh) * 2016-09-09 2018-03-16 阿里巴巴集团控股有限公司 一种针对提问的回复方法和装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3779725A4

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111091825A (zh) * 2019-12-03 2020-05-01 北京蓦然认知科技有限公司 一种构建交互引擎簇的方法及装置
CN111091825B (zh) * 2019-12-03 2022-09-23 杭州蓦然认知科技有限公司 一种构建交互引擎簇的方法及装置
CN111881270A (zh) * 2020-07-01 2020-11-03 北京嘀嘀无限科技发展有限公司 一种智能对话方法和系统

Also Published As

Publication number Publication date
CN109948017B (zh) 2021-03-30
US20200264923A1 (en) 2020-08-20
EP3779725A1 (en) 2021-02-17
EP3779725A4 (en) 2021-06-30
CN109948017A (zh) 2019-06-28
CN113536093A (zh) 2021-10-22

Similar Documents

Publication Publication Date Title
WO2019205679A1 (zh) 一种信息处理方法及装置
CN109003605B (zh) 智能语音交互处理方法、装置、设备及存储介质
CN108664650B (zh) 一种区块链网络的事务处理方法、装置、设备及存储介质
US9619545B2 (en) Naïve, client-side sharding with online addition of shards
US9558242B2 (en) Social where next suggestion
CN108604177B (zh) 用于序列相关数据消息整合的系统和方法
US20130151467A1 (en) Slave Consistency in a Synchronous Replication Environment
US10182024B1 (en) Reallocating users in content sharing environments
US20190324795A1 (en) Composite task execution
KR20190005224A (ko) 액션에 대한 엔티티를 결정하고 그리고 액션 및 엔티티에 관련된 하나 이상의 컴퓨터 액션들을 수행하기 위해 촉구들에 대한 사용자 응답들의 촉진 및 분석의 자동화된 생성을 위한 방법 및 장치
US20190147029A1 (en) Method and system for generating conversational user interface
JP2020506470A (ja) 取引対象予約システム、方法、及び装置
US10599380B2 (en) Method and system for automatically managing content in an electronic device
CN111813900A (zh) 多轮对话处理方法、装置、电子设备及存储介质
CN114997414B (zh) 数据处理方法、装置、电子设备和存储介质
CN104111957A (zh) 一种分布式事务同步的方法及系统
AU2023201871A1 (en) Automated assistant for generating, in response to a request from a user, application input content using application data from other sources
KR102164409B1 (ko) 사용자 제공 식별자들을 크라우드소싱하고 그것들을 브랜드 아이덴티티들 연관시키기
US10033737B2 (en) System and method for cross-cloud identity matching
WO2016155384A1 (zh) 一种搜索优化方法、装置和系统
US11863635B2 (en) Enhanced processing of user profiles using data structures specialized for graphical processing units (GPUs)
CN115762503A (zh) 车载语音系统、车载语音的自主学习方法、设备及介质
US11475221B2 (en) Techniques for selecting content to include in user communications
CN108287853B (zh) 一种数据关系分析方法及其系统
WO2019165762A1 (zh) 一种抽样查询的方法和装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18915826

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2018915826

Country of ref document: EP

Effective date: 20201125

ENP Entry into the national phase

Ref document number: 2018915826

Country of ref document: EP

Effective date: 20200831