WO2021239002A1

WO2021239002A1 - Group chat robot autonomous response method and system

Info

Publication number: WO2021239002A1
Application number: PCT/CN2021/096153
Authority: WO
Inventors: 杨明晖; 林川杰; 刘威; 崔恒斌; 刘洋
Original assignee: 支付宝(杭州)信息技术有限公司
Priority date: 2020-05-27
Filing date: 2021-05-26
Publication date: 2021-12-02
Also published as: CN111368059A; CN111368059B; CN112052323A

Abstract

A group chat robot autonomous response method and system. After an input statement sent by a target user in a target communication group is received, a candidate list related to the input statement is generated on the basis of the input statement and a preset knowledge base, and the candidate list is sent to a target client or a first group of clients comprising the target client, so that the influence on other users is reduced; after a management client of the first group of clients makes a selection on the candidate list, a target answer matching the selection is sent to a second group of clients. According to the method, questions raised by a user can be actively answered, moreover, other users in the group chat are not disturbed, and the user experience is improved while the working efficiency is improved.

Description

Method and system for autonomous response of group chat robot

Technical field

This specification relates to the field of Internet technology, and in particular to a method and system for autonomous response of group chat robots.

Background technique

With the rapid development of information technology and Internet technology, people's work, life, study, etc. are closely connected to the network. Instant Messaging (IM) technology supports two or more people to communicate through the network, which is widely used. In people's work and life, such as daily work and life communication, etc., it brings all kinds of convenience to people's work and life. Common IM software from the early MSN, ICQ, OICQ, to today's QQ, WeChat, Dingding, Baidu Hi, etc., instant messaging technology is constantly evolving in the direction of more convenient users. Many instant messaging applications support group chat technology. For example, WeChat, Dingding and Baidu Hi all support group chat technology. Group chat technology can allow multiple users in real life to gather together to interact and share information. In some service groups, there are both users of the service group and customer service personnel, such as customer service robots. In the use of the service group, users will conduct dialogues or business consultations in the service group. When a user consults about affairs in a service group, the user often hopes that the customer service robot can actively answer the question, but at the same time does not want the customer service robot to disturb other users. Therefore, how to ensure that the customer service robot can actively respond to the affairs consulted by the user without disturbing the dialogue and communication between the users is a problem that needs to be solved urgently.

Therefore, in order to ensure that the robot actively responds to the affairs of the user's consultation without disturbing the user's dialogue and communication, a more efficient method and system for group chat robots to autonomously respond is needed.

Summary of the invention

In order to ensure that the robot actively responds to the affairs of the user's consultation without disturbing the user's dialogue and communication, this manual provides a more efficient method and system for group chat robots to respond autonomously.

The group chat robot autonomous response method and system provided in this manual, after receiving the input sentence sent by the user in the group chat, generate a candidate list related to the input sentence based on the input sentence and the preset knowledge base, and combine the candidates The list is sent to the user who sent the input sentence. After the user makes a selection on the candidate list, the answer that matches the selection result is sent to other users in the group chat. The method and system can not only actively answer questions raised by users, but also will not cause interference to other users in the group chat. While improving work efficiency, it can also improve user experience.

In the first aspect, this specification provides a method for group chat robots to autonomously respond, including: receiving an input sentence, the input sentence being input by a target client on an instant messaging group interface; based on the input sentence and a preset knowledge base Generate a candidate list, the candidate list including at least one recommended knowledge point related to the input sentence; sending the candidate list to the instant messaging group, the candidate list being marked as being only available to users in the instant messaging group The first group of clients is visible, and the first group of clients are part of the clients including the target client; receiving the selection of the candidate list by the management client in the first group of clients; and Send a target answer matching the selection to the instant messaging group, where the target answer is marked as visible to a second group of clients in the instant messaging group, and the second group of clients includes the At least some clients in the instant messaging group.

In some embodiments, the sending the candidate list to the instant messaging group includes: sending first icon data to the instant messaging group, and the first icon data is marked as only for the first group of customers It can be seen from the end that the first icon data includes a list display request; receiving a signal for triggering the list display request by the first group of clients; sending the candidate list to the instant messaging group, and the candidate list is marked as Visible to the client that triggered the list display request.

In some embodiments, the sending the candidate list to the instant messaging group includes: sending the candidate list and the first icon data to the instant messaging group, wherein the first icon data is marked as only Visible to the first group of clients, the first icon data includes a list display request, and when the first icon data is triggered by the client, the candidate list is displayed on the client that triggered the first icon data .

In some embodiments, the sending the target answer matching the selection to the instant messaging group includes: sending the target answer and second icon data to the instant messaging group, wherein the second icon The data is marked as visible to the second group of clients, the second icon data includes an answer display request, and when the second icon data is triggered by the client, it is displayed on the client that triggered the second icon data The target answer.

In some embodiments, the generating a candidate list based on the input sentence and a preset knowledge base includes: identifying whether the input sentence includes a target question related to a preset transaction, and the preset transaction is directed to the Customized by the instant messaging group; determining that the input sentence includes the target question; and generating the candidate list based on the input sentence and the preset knowledge base.

In some embodiments, the identifying whether the input sentence includes a target question related to a preset transaction includes: matching the input sentence with a plurality of preset question templates to identify whether the input sentence is Question sentence; and when the input sentence is a question sentence, classify the input sentence, and identify whether the input sentence includes a target question related to the preset transaction.

In some embodiments, the method for the group chat robot to autonomously answer further includes: sending an answer evaluation request to the instant messaging group, the answer evaluation request being marked as visible only to the target client; and receiving the answer Describe the answer evaluation of the target client.

In some embodiments, the management client includes the target client.

In some embodiments, the second group of clients includes: clients selected by the management client from the instant messaging group or all clients in the instant messaging group.

In a second aspect, this specification provides a system for autonomous response of group chat robots, including at least one storage medium and at least one processor, the at least one storage medium includes at least one instruction set for autonomous response of group chat robots; A processor is in communication connection with the at least one storage medium, wherein, when the system is running, the at least one processor reads the at least one instruction set, and executes this specification according to the instructions of the at least one instruction set The method for autonomous response of group chat robots described in the first aspect.

In a third aspect, this specification provides a method for group chat robots to autonomously respond, including: receiving an input sentence and sending it to the server, the input sentence being input by the target user on the instant messaging group interface; receiving the message sent by the server The candidate list is displayed on the instant messaging group interface, the candidate list is generated based on the input sentence and the preset knowledge base, and the candidate list includes at least one recommendation related to the input sentence Knowledge points; receiving the target user’s selection of the candidate list and sending it to the server; sending the selection to the instant messaging group; and receiving the target that matches the selection sent by the server Answer.

In some embodiments, the receiving the candidate list sent by the server includes: receiving the candidate list and the first icon data sent by the server, wherein the first icon data includes a list display request, when When the first icon data is triggered by the client, display the candidate list on the client that triggered the first icon data; receive the trigger operation of the target user on the first icon data; and in the instant The communication group displays the candidate list.

In a fourth aspect, this specification provides a system for autonomous response of group chat robots, including at least one storage medium and at least one processor, the at least one storage medium includes at least one instruction set for autonomous response of group chat robots; A processor is in communication connection with the at least one storage medium, wherein, when the system is running, the at least one processor reads the at least one instruction set, and executes this specification according to the instructions of the at least one instruction set The method for autonomous response of group chat robots described in the third aspect.

The group chat robot autonomous response method and other functions of the system provided in this manual will be partially listed in the following description. According to the description, the content introduced by the following figures and examples will be obvious to those of ordinary skill in the art. The creative aspects of the group chat robot autonomous response method, system, and storage medium provided in this specification can be fully explained through practice or use of the methods, devices, and combinations described in the following detailed examples.

Description of the drawings

In order to explain the technical solutions in the embodiments of this specification more clearly, the following will briefly introduce the drawings needed in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the specification. For those of ordinary skill in the art, without creative work, other drawings can be obtained based on these drawings.

Fig. 1 shows a schematic diagram of a system for autonomous response of group chat robots according to an embodiment of this specification;

FIG. 2 shows a schematic diagram of a device structure for autonomous response of a group chat robot according to an embodiment of this specification;

Figure 3 shows a flow chart of a method for group chat robots to autonomously respond according to an embodiment of this specification;

Fig. 4 shows a flow chart for generating a candidate list according to an embodiment of the present specification;

Figure 5 shows a flow chart of a method for group chat robots to autonomously respond according to an embodiment of this specification; and

Fig. 6 shows a schematic diagram of the instant messaging group interface provided according to an embodiment of the present specification.

Detailed ways

The following description provides specific application scenarios and requirements of this specification, with the purpose of enabling those skilled in the art to make and use the content in this specification. For those skilled in the art, various partial modifications to the disclosed embodiments are obvious, and without departing from the spirit and scope of this specification, the general principles defined herein can be applied to other embodiments and application. Therefore, the present specification is not limited to the illustrated embodiment, but is the widest scope consistent with the claims.

The terminology used here is only for the purpose of describing specific example embodiments, and is not restrictive. For example, unless the context clearly indicates otherwise, as used herein, the singular form "a", "an" and "the" may also include plural forms. When used in this specification, the term "comprises", "includes" or "contains" means that the associated integers, steps, operations, elements or components exist, but do not exclude one or more other features, integers, and steps. The existence of, operations, elements, components or groups or other features, integers, steps, operations, elements, components or groups can be added to the system/method.

In consideration of the following description, the operation and function of these and other features of this specification, as well as the related elements of the structure, as well as the combination of components and the economics of manufacturing can be significantly improved. With reference to the drawings, all of these form a part of this specification. However, it should be clearly understood that the drawings are only for illustration and description purposes, and are not intended to limit the scope of this specification. It should also be understood that the drawings are not drawn to scale.

The flowchart used in this specification shows operations implemented by the system according to some embodiments in this specification. It should be clearly understood that the operations of the flowchart can be implemented out of order. Conversely, the operations can be reversed or performed simultaneously. In addition, one or more other operations can be added to the flowchart. One or more operations can be removed from the flowchart.

Group chat technology is an important function of instant messaging application software, which allows multiple users in real life to gather together to interact and share information. In order to improve work efficiency, a service group was launched on the basis of ordinary group chat. The service group refers to a group chat established for the purpose of serving users. There are not only users but also customer service in the service group. The user may be a served person in the service group. The user can conduct group chat communication in the service group, and can also conduct business consultation, business help or business feedback in the service group. The customer service may be a server in the service group, and may answer users' questions, respond to user requests, or send notifications or announcements to users in the service group, and so on. The customer service may include a human customer service, and may also include a customer service robot. The customer service robot can help the manual customer service respond to the user's request, reduce the pressure of the manual customer service, and improve the work efficiency of the service group. The customer service robot can provide a series of auxiliary functions for manual customer service, including service reminders, answer recommendations, and question and answer storage. Customer service robots can also directly answer user questions.

The group chat robot autonomous response method and system provided in this manual, after receiving the input sentence sent by the user in the group chat, generate a candidate list related to the input sentence based on the input sentence and the preset knowledge base, and combine the candidates The list is sent to the user who sent the input sentence. After the user makes a selection on the candidate list, the answer that matches the selection result is sent to other users in the group chat. The method and system can not only actively answer questions raised by users, but also will not cause interference to other users in the group chat, and improve user experience while improving work efficiency.

FIG. 1 shows a schematic diagram of a system 100 for autonomous response of a group chat robot. The system 100 may include a server 200, a client 300, a network 120, and a database 150. The client 300 may include a target client 301, a client 302, a client 303, a client 304, and so on.

The server 200 may store data or instructions for executing the group chat robot autonomous response method described in this specification, and may execute or be used to execute the data or instructions. The server 200 may include the robot server. The server 200 may further include a server that provides the group chat service. The group chat service may also be provided by the robot server.

As shown in FIG. 1, the user 110 is a user of the client 300. The user 110 may include a target user 111, a user 112, a user 113, a user 114, and so on. The target user 111 is the user of the target client 301, the user 112 is the user of the client 302, the user 113 is the user of the client 303, the user 114 is the user of the client 304, and so on. The client 300 is a connection device for establishing communication between the target user 110 and the server 200 in general. The client 300 communicates with the server 200. The server 200 can communicate with multiple clients 300 at the same time. The client 300 may be installed with one or more applications (APP). The APP can provide the user 110 with the ability and interface to interact with the outside world through the network 120. The APP including but not limited to: Chat class APP program, shopping category APP program, video classes APP program, wealth management APP program, etc., such as Alipay ^TM, Taobao ^TM, Jingdong ^TM or such as banking and financial services institutions, financial products such as APP program. The client 300 may include a mobile device, a tablet computer, a notebook computer, a built-in device of a motor vehicle, or the like, or any combination thereof. The mobile device may include a smart home device, a smart mobile device, a virtual reality device, an augmented reality device, or similar devices, or any combination thereof. The smart home device may include a smart TV, a desktop computer, etc., or any combination. The smart mobile device may include a smart phone, a personal digital assistant, a game device, a navigation device, etc., or any combination thereof. The virtual reality device or augmented reality device may include a virtual reality helmet, virtual reality glasses, virtual reality patch, augmented reality helmet, augmented reality glasses, augmented reality patch or similar content, or any combination thereof. For example, the virtual reality device or the augmented reality device may include Google glasses, head-mounted displays, gear VR, and so on. The built-in devices in the motor vehicle may include a vehicle-mounted computer, a vehicle-mounted TV, and the like. The client 300 may be a device with positioning technology for locating the position of the client 300.

The client 300 is loaded with instant messaging application software corresponding to the server 200, for example, DingTalk ^™ . The instant messaging application software is provided with an instant messaging group. The user 110 is a user in the instant messaging group. The user 110 can input chat information in the instant messaging group. The instant messaging group includes at least two users 110.

The network 120 may facilitate the exchange of information or data. As shown in FIG. 1, the client 300, the server 200, and the database 150 can be connected to the network 120, and transmit information or data to each other through the network 120. For example, the server 200 may obtain input information from the client 300 through the network 120. The network 120 can be any type of wired or wireless network, or a combination thereof. For example, the network 120 may include a cable network, a wired network, an optical fiber network, a telecommunication communication network, an intranet, the Internet, a local area network (LAN), a wide area network (WAN), a wireless local area network (WLAN), a metropolitan area network (MAN), and a wide area network. (WAN), Public Switched Telephone Network (PSTN), Bluetooth network, ZigBee network, Near Field Communication (NFC) network or similar network. The network 120 may include one or more network access points. For example, the network 120 may include wired or wireless network access points, such as base stations or Internet exchange points 120-1, 120-2,..., through which one or more of the client 300, the server 200, and the database 150 Each component can be connected to the network 120 to exchange data or information.

The database 150 may store data or instructions. The database 150 may store data obtained from the server 200 or the client 300. The database 150 may store data or instructions that can be executed by the server 200 or used to execute the group chat robot autonomous response method described in this specification. The database 150 may store a knowledge base related to the preset affairs of the instant messaging group. The server 200 and the client 300 may have access to the database 150, and the server 200 and the client 300 may access data or instructions stored in the database 150 through the network 120. The database 150 can be directly connected to the server 200 and the client 300. The database 150 may be a part of the server 200. The database 150 may include mass storage, removable storage, volatile read-write memory, read-only memory (ROM), or the like, or any combination thereof. Exemplary mass storage may include non-transitory storage medium (non-transitory storage medium) such as magnetic disks, optical disks, and solid-state drives. Removable storage may include flash drives, floppy disks, CDs, memory cards, zip disks, tapes, etc. Typical volatile read-write memory may include random access memory (RAM). RAM may include dynamic RAM (DRAM), double date rate synchronous dynamic RAM (DDR SDRAM), static RAM (SRAM), thyristor RAM (T-RAM), and zero capacitance RAM (Z-RAM). ROM may include mask ROM (MROM), programmable ROM (PROM), virtual programmable ROM (PEROM), electronic programmable ROM (EEPROM), compact disk (CD-ROM) and digital versatile disk ROM, etc. The database 150 can be implemented on a cloud platform. Merely as an example, the cloud platform may include private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, inter-cloud cloud, etc., or forms similar to the foregoing forms, or any combination of the foregoing forms.

As shown in Figure 1, the target user 111 inputs chat information through the instant messaging group of the instant messaging application software on the target client 301, and the chat information is transmitted to the server 200 through the network 120; / Or the group chat robot in the database 150 instructions for the autonomous answering method, answering the questions contained in the input chat information to generate a candidate list, and sending the candidate list to the target user 111 or some users in the instant messaging group ; After the target user 111 makes a selection on the candidate list, the server 200 sends a target answer that matches the selection to all users or a part of the users in the instant messaging group.

Figure 2 shows a schematic diagram of the structure of an autonomous response device for a group chat robot. The device may be the server 200 or the client 300. In the following description, the server 200 is taken as an example to introduce the device.

The server 200 can perform the autonomous response method of the group chat robot described in this specification. The method is described in other parts of this specification. For example, in the description of FIG. 3 to FIG. 4, a method P200 for the server 200 to execute the autonomous response of the group chat robot is introduced.

As shown in FIG. 2, the server 200 includes at least one storage medium 230 and at least one processor 220. The server 200 may also include a communication port 250 and an internal communication bus 210. Meanwhile, the server 200 may further include an I/O component 260.

The internal communication bus 210 can connect different system components, including the storage medium 230 and the processor 220.

The I/O component 260 supports input/output between the server 200 and other components.

The storage medium 230 may include a data storage device. The data storage device may be a non-transitory storage medium or a temporary storage medium. For example, the data storage device may include one or more of a magnetic disk 232, a read-only storage medium (ROM) 234, or a random access storage medium (RAM) 236. The storage medium 230 also includes at least one instruction set stored in the data storage device. The instructions are computer program codes, and the computer program codes may include programs, routines, objects, components, data structures, procedures, modules, etc., that execute the group chat robot autonomous response method provided in this specification.

The communication port 250 is used for data communication between the server 200 and the outside world. For example, the server 200 may be connected by a communication port 250 network 120, the target receives the user input sentence 111 by the APP universal instant messaging (e.g. nails ^(TM)) input, 250 111 further recommended to the target user answers through the communication port.

At least one processor 220 and at least one storage medium 230 are communicatively connected through an internal communication bus 210. At least one processor 220 is used to execute the above-mentioned at least one instruction set. When the system 100 is running, at least one processor 220 reads the at least one instruction set, and executes the group chat robot autonomous response method P200 provided in this specification according to the instructions of the at least one instruction set. The processor 220 can execute all the steps included in the method P200 for group chat robot autonomous response. The processor 220 may be in the form of one or more processors, and the processor 220 may include one or more hardware processors, such as a microcontroller, a microprocessor, a reduced instruction set computer (RISC), an application specific integrated circuit (ASIC), Application-specific instruction set processor (ASIP), central processing unit (CPU), graphics processing unit (GPU), physical processing unit (PPU), microcontroller unit, digital signal processor (DSP), field programmable gate Array (FPGA), advanced RISC machine (ARM), programmable logic device (PLD), any circuit or processor capable of performing one or more functions, etc., or any combination thereof. Just to illustrate the problem, only one processor 220 is described in the server 200 in this specification. However, it should be noted that the server 200 in this specification may also include multiple processors. Therefore, the operations or method steps disclosed in this specification may be executed by one processor as described in this specification, or may be jointly executed by multiple processors. . For example, if the processor 220 of the server 200 performs step A and step B in this specification, it should be understood that step A and step B can also be performed jointly or separately by two different processors 220 (for example, the first processor performs step A, the second processor executes step B, or the first and second processors jointly execute steps A and B).

Although the above structure describes the server 200, this structure is also applicable to the client 300. The target client 301 can execute the group chat robot autonomous response method described in this specification. The method is described in other parts of this specification. For example, in the description of FIG. 5 and FIG. 6, a method P300 for the target client 301 to perform the autonomous response of the group chat robot is introduced.

Fig. 3 shows a flowchart of a method P200 for autonomous response of a group chat robot. As mentioned above, the server 200 can execute the group chat robot autonomous response method P200 provided in this specification. Specifically, the processor 220 in the server 200 can read the instruction set stored in its local storage medium and/or the database 150, and then execute the group chat robot autonomous response method provided in this specification according to the provisions of the instruction set. P200. The method P200 may include execution by at least one processor 220 on the server 200 side: S220: receiving an input sentence, the input sentence being input by the target client 301 on the instant messaging group interface.

As mentioned above, the method P200 for autonomous response of group chat robots is mainly used for group chat services. The target client 301 may be installed with general instant messaging application software, for example, Dingding ^TM APP. DingTalk ^TM APP may have a service group, that is, the instant messaging group. The service of the instant messaging group can be provided by the server 200, can be provided by the group chat service server in the server 200, or can be provided by the robot server. The instant messaging group provides shared instant public messaging services on the interface for at least two users 110. For example, when a user (for example, target user 111) enters a message in the instant messaging group, the message can be instantly seen by all other users (for example, user 112, user 113, user 114, etc.). Any user 110 can answer the input message. All users 110 and the robots in the instant messaging group can be online at the same time. The target user 111 is a user in the instant messaging group.

The instant messaging group may be a transaction service group. Further, the instant messaging group may be a business service group customized for a specific group of people. The customer service robot in the instant messaging group can provide transaction consulting services for users in the instant messaging group. The transaction is a transaction associated with the instant messaging group. Different users have different definitions of the transaction. For example, the affairs defined by the automobile company may be related to the automobile, such as the engine of the automobile, the tire of the automobile, and so on. The affairs defined by the financial company are those that the finance wants to do, for example, stocks, funds, insurance, and so on. Different departments in the same automobile company have different definitions of affairs. For example, the engine department's definition of business may be the model of the engine, the power of the engine, the displacement of the engine, and so on. For example, the definition of business in the automobile tire department can be the structure of the tire, the pressure of the tire, the pattern of the tire, and so on. Since the transactions defined by different users are different, the transactions associated with different communication groups are also different. The instant messaging group often presets transactions according to the needs of users, so the preset transactions are customized for the instant messaging group. The preset affairs can be supplemented, modified and changed. Affairs consulting services provided by robots in the instant messaging group according to the preset affairs.

The target user 111 can input the input sentence through the interface of the instant messaging group on the target client 301. The input sentence may be real-time chat information input by the target user 111 in the instant messaging group. The input sentence may be voice data or text data. When the input sentence is voice data, the server 200 may convert the voice data into text data. The input sentence may be any sentence input by the target user 111 in the instant messaging group, but may be a communication sentence with the user 110, or some doubts or difficulties encountered, and so on. The input sentence may be data related to the preset transaction, or data not related to the preset transaction. The input sentence may be the data for the users in the instant messaging group to communicate with each other, or it may be a question asked by the target user 111 to the robot or other users 110. The input sentence may not have any mark to ask the robot a question. For example, in a context of dialogue and communication between users, the input sentence is a natural dialogue between users, so there is no mark to ask the robot a question. There may also be some questions in the input sentence, such as "What is Huabei".

The working mode of the instant messaging group may be: the target user 111 inputs the input sentence on the instant messaging group interface of the target client 301, and the input sentence is transmitted to the server 200 through the network 120, and the instant messaging group Clients of other users, such as client 302, client 303, client 304, etc., will actively monitor the server 200. Once the input sentence is monitored, it will actively obtain the input sentence from the server 200 and display it on On the interface of the instant messaging group.

S230: Generate a candidate list based on the input sentence and the preset knowledge base.

The server 200 may respond to the input sentence according to the content of the input sentence. Specifically, the server 200 may generate a candidate list related to the input sentence according to the content of the input sentence and the knowledge base preset by the instant messaging group. The predetermined knowledge base may be a collection of knowledge points related to the predetermined transaction. The knowledge base can be used to describe the functions or rules of the preset affairs. For example, the instant messaging group is an Alipay transaction group, the preset transaction may be a transaction related to Alipay, and the preset knowledge base may include multiple knowledge points related to functions supported by Alipay. The knowledge base may include a plurality of preset knowledge points, and each of the plurality of knowledge points includes a knowledge point title and a knowledge point answer. Generally, the title of the knowledge point may be a question asked about the predetermined transaction, and the answer of the knowledge point may be the answer content to the question. For example, in the Alipay transaction group, the knowledge points of the knowledge base are "Title: What is Huabei? Answer: Huabei is a loan product.", "Title: What is Yu'e Bao? Answer: Yu'e Bao is one A currency fund.", "Title: What should I do if I forget my password? Answer: Use my mobile phone number to retrieve my password." and so on. The knowledge base is preset based on the preset affairs of the instant messaging group. The knowledge base can be supplemented, modified and changed. The knowledge base may be stored in the server 200 or in the database 150.

The candidate list may be a number of knowledge point titles related to the input sentence generated by the server 200 according to the input sentence and the knowledge base. The candidate list may include at least one recommended knowledge point related to the input sentence. Specifically, the candidate list may include at least one recommended knowledge point title related to the input sentence.

No matter what the input sentence is, in order to ensure that the robot can quickly respond to questions related to the preset affairs of the instant messaging group without disturbing the normal communication between the users 110 in the instant messaging group , The robot needs to recognize the input sentence for the transaction problem, and judge whether it needs to intervene in the dialogue to provide an answer. Fig. 4 shows a flowchart for implementing step S230 according to an embodiment of the specification, that is, a flowchart for generating the candidate list. Specifically, step S230 may include: S232: identifying whether the input sentence includes a target question related to a preset transaction.

After the target user 111 inputs the input sentence through the instant messaging group on the target client 301, the server 200 needs to perform transaction problem recognition on the input sentence to identify whether the input sentence is related to the preset transaction Question related goals. When the input sentence is not a question related to the preset transaction, the server 200 does not respond to the input sentence; when the input sentence is a question related to the preset transaction, the server 200 asks all questions. Autonomous response to the input sentence. The target question related to the predetermined transaction may be a question related to multiple knowledge points in the knowledge base. The target question may also be at least one of multiple knowledge point titles in the knowledge base.

The server 200 may perform transaction problem recognition on the input sentence through the question recognition module 282 and the transaction recognition module 284. The question recognition module 282 may receive the input sentence and be configured to recognize whether the input sentence is a question sentence. Wherein, the question recognition module 282 is obtained by training based on a plurality of preset question templates. The preset multiple question sentence templates may be created based on multiple question sentence format information. The question recognition module 282 can use the preset multiple question template to match the input sentence to determine whether a template matches the input sentence, thereby identifying whether the input sentence is a question sentence Statement. In this specification, question format information may include general question format information and declarative question format information. The general question format refers to the question format in the grammatical sense. For example, the obvious question format, such as "How is the weather?", "What is Dingding?", "Why the income of Yu'ebao is getting lower?" and so on. The question format of the statement refers to the statement format used to describe a certain fact in order to seek a response (for example, hope that others can help to deal with it), such as "attendance check-in failed", "help me approve", "Taobao can't Login" and so on.

The transaction recognition module 284 may receive the input sentence and be configured to classify the input sentence, that is, transaction question recognition, and identify whether the input sentence includes a target question related to the preset transaction. The transaction recognition module 284 is trained based on the preset knowledge base and preset non-transaction corpus information. The transaction recognition module 284 may be a text classification model, and the classification result is a two-class classification. The transaction recognition module 284 uses the preset knowledge base as a positive sample for training, and uses the preset non-transaction corpus information as a negative sample for training, thereby forming a training sample set for model training. The ratio of the data amount of the positive sample to the data amount of the negative sample may be 1:1, or other ratios, such as 1.2:1, and so on. The server 200 inputs the input sentence into the transaction recognition module 284, and the output result of the transaction recognition module 284 is that the input sentence belongs to the knowledge base type or the input sentence belongs to the non-transaction corpus type. The preset non-business corpus information may be general small chat information, for example, "Hello", "How is the weather today" and so on. When the output result of the transaction recognition module 284 is that the input sentence belongs to the knowledge base category, the server 200 determines that the input sentence includes a target question related to the preset transaction, and the server 200 can perform autonomously on the target question. Response; when the output result of the transaction recognition module 284 is that the input sentence belongs to a non-transaction corpus, the server 200 determines that the input sentence does not include a target question related to the preset transaction, and therefore can choose not to answer. Through such a setting, the users 110 can be disturbed as little as possible during normal chats, and the user experience is enhanced.

The transaction recognition module 284 may include one of a FastText model, a CNN model, and an LSTM model. The type of text classification model used by the transaction recognition module 284 may be determined based on the data volume of the sample used in the training of the text classification model. For example, when the sample used in the text classification model training is a large amount of data, a CNN model or an LSTM model can be used. When the sample used in the text classification model training is a small amount of data, the FastText model can be used. The transaction recognition module 284 provided in this specification can determine the type of text classification model used based on the amount of sample data used during text classification model training, and can select suitable text classification models for classification based on different application scenarios, that is, transaction Problem identification, thereby improving the efficiency of transaction problem identification.

The question recognition module 282 and the transaction recognition module 284 can perform question recognition on the input sentence first, and perform transaction problem recognition on the input sentence when the input sentence is a question sentence. When sentence, no transaction problem identification is performed, which can effectively save the amount of calculation and save resources.

Specifically, step S232 may include: S232-2: matching the input sentence with the plurality of preset question template to identify whether the input sentence is a question sentence.

The server 200 performs question format matching on the input sentence through the question recognition module 282, and uses the input sentence to traverse all the question sentence formats in the plurality of preset question sentence formats. If there is a matching question format, the server 200 determines that the input sentence is a question sentence. Otherwise, the server 200 determines that the input sentence is not a question sentence. If the input sentence is determined by the server 200 as a question sentence, the server 200 regards the input sentence as a potential transaction problem and performs transaction problem identification.

S232-4: When the input sentence is a question sentence, classify the input sentence, and identify whether the input sentence includes a target question related to the preset transaction.

When the input sentence is not a question sentence, the server 200 determines that the input sentence does not include a target question related to the preset transaction. The server 200 considers that the input sentence is not a question about the predetermined transaction, and does not respond to the input sentence.

When the input sentence is a question sentence, the server 200 needs to further classify the input sentence, that is, transaction problem identification, and determine whether the input sentence includes a target question related to the preset transaction. Specifically, the server 200 classifies the input sentence through the transaction recognition module 284, and recognizes whether the input sentence includes a target question related to the preset transaction. When classifying the input sentence, the transaction recognition module 284 may first perform word segmentation processing on the input sentence. For example, the input sentence recognized as a question sentence "How to activate Huabei?" is segmented as "Huabe", "How" and "Open"; then, perform text classification processing on the word segmentation obtained after the word segmentation processing. For example, in the case where the transaction recognition module 284 is the FastText model, first, the transaction recognition module 284 expresses the word segmentation "Huabe", "How" and "Kaitong" as three word vectors, for example, using Word2Vec to segment the word "Hua "Bai", "How" and "Open" are represented as 3 word vectors; then, the transaction recognition module 284 uses convolution processing to convert the obtained 3 word vectors into a semantic matrix; then, the transaction recognition module 284 uses pooling Calculation to convert the semantic matrix into a semantic vector; finally, the transaction recognition module 284 uses the Softmax layer to classify to obtain the scores classified into the knowledge base class and the non-transaction corpus class, and compare the scores Large categories are output as text classification results of the transaction recognition module 284. When the output result of the transaction recognition module 284 is that the input sentence belongs to the knowledge base class, the server 200 determines that the input sentence includes a target question related to the preset transaction; when the output result of the transaction recognition module 284 is all When the input sentence belongs to a non-transactional corpus, the server 200 determines that the input sentence does not include a target question related to the predetermined transaction.

The group chat robot autonomous response method P200 and system 100 provided in this specification uses the question recognition module 282 and the transaction recognition module 284 to first perform question recognition 282 on the input sentence, only when the input sentence is a question sentence In the case of, the input sentences are classified, that is, transaction problem recognition, and it is not necessary to perform transaction problem recognition on all input sentences, thereby reducing the amount of calculation and improving work efficiency.

The server 200 can also recognize whether the input sentence includes a target question related to a preset transaction through the semantic recognition module. The semantic recognition module may be trained based on historical input sentences. The historical input sentence may include a question related to the preset transaction, and may also include an exchange dialogue between the users 110. The semantic recognition module may perform semantic analysis on the input sentence, and determine whether the semantics of the input sentence includes a target question related to the preset transaction according to the result of the semantic analysis. For example, the historical input sentences related to the preset transaction are "What is Dingding?", "Why the income of Yu'ebao is getting low?", "Check attendance failed", "Taobao can't log in", etc. For example, the historical input sentence that is not related to the preset transaction may be "how is the weather?" and so on. For example, the input sentence is "My Yu'E Bao income was 100 last month, but only 50 this month". The server 200 can obtain the semantic analysis result "Yuebao's income has decreased" by performing semantic analysis on the input sentence. Therefore, it is judged that the input sentence may include a target question related to the preset transaction. The server 200 can perform semantic recognition on the input sentence through the semantic recognition module, analyze the semantics intended to be expressed by the input sentence, and thereby recognize whether the input sentence includes a target question related to the predetermined transaction. The semantic recognition module can perform semantic analysis on input sentences whose question expressions are not obvious, and dig out the potential expressions of the input sentences, so as to make more accurate judgments.

When the input sentence does not include a target question related to the preset transaction, the server 200 considers that the input sentence is not a question about the preset transaction, and does not respond to the input sentence. When the input sentence includes a target question related to the preset transaction, the server 200 may actively respond to the input sentence. This can ensure that the server 200 (ie the robot) can actively respond to the transaction questions raised by the user 110 in the instant messaging group when the user 110 does not take the initiative to Disturbing the dialogue and communication between the users 110 in the instant messaging group, improving the user experience, and at the same time alleviating the work pressure of manual customer service, reducing the workload of manual customer service, increasing the utilization rate of the robot, and improving customer service ( Including the work efficiency of the manual customer service and the robot). Specifically, after recognizing that the input sentence includes a target question related to the preset transaction, the server 200 may perform: S234: Determine that the input sentence includes a target question related to the preset transaction.

S236: Generate the candidate list based on the input sentence and the preset knowledge base.

As described above, the candidate list may include at least one recommended knowledge point title related to the input sentence. In step S236, the server 200 may generate the candidate list to reply to the user by finding the knowledge point title most similar to the input sentence. In step S236, generating a candidate list based on the input sentence and the preset knowledge base may include executing by at least one processor 220 of the server 200: S236-2: performing the input sentence Word segmentation processing to extract keywords corresponding to the input sentence.

For example, the server 200 divides the input sentence "How to open Huabei?" into "Huabe", "How" and "Open"; then, extract the keywords in the word segmentation obtained after word segmentation processing, for example, the keyword is "Huabei" and "kaitong".

S236-4: Match the keyword corresponding to the input sentence with the preset knowledge base to obtain at least one recalled knowledge point, and the multiple knowledge points include the at least one recalled knowledge point.

After the server 200 extracts the keywords of the input sentence, it matches the keywords of the input sentence with multiple knowledge points in the knowledge base. In fact, it matches the keywords of the input sentence with the knowledge points. Multiple knowledge point titles in the library are matched, and at least one knowledge point matching the keyword of the input sentence is recalled from the knowledge base as the recalled knowledge point. The recall is to select knowledge points matching the keywords of the input sentence from the knowledge base. The recalled knowledge points are the selected candidate knowledge points. For example, the keywords of the input sentence are "huabei" and "kaitong". The server 200 searches the knowledge base for whether there is a knowledge point title consistent with "Huabe" and "Open". For example, the titles of the recalled knowledge points may be "How to activate Huabei" and "How to adjust the quota after Huabei is activated", and so on.

S236-6: Calculate the correlation value between the input sentence and the at least one recalled knowledge point.

After the server 200 recalls the at least one knowledge point, it may calculate the associated value of each recalled knowledge point in the input sentence and the at least one recalled knowledge point. The correlation value may be a similarity value between the input sentence and each recalled knowledge point. The server 200 performs semantic analysis on the input sentence to calculate the semantic similarity value between the input sentence and the title of each recalled knowledge point.

S236-8: Based on the correlation value, select at least one recommended knowledge point from the at least one recalled knowledge point, and generate the candidate list based on the title of the recommended knowledge point.

After the server 200 calculates the associated value between the input sentence and each recalled knowledge point, it sorts the at least one recalled knowledge point based on the associated value. The ranking may be based on ranking the at least one recalled knowledge point from large to small based on the associated value. The server 200 may select the at least one recalled knowledge point with the largest correlation value from the at least one recalled knowledge point as the recommended knowledge point, and generate the candidate list according to the title of the recommended knowledge point. The candidate list may include at least one recommended knowledge point title.

The server 200 selects the recommended knowledge points with the highest similarity to the input sentence from the knowledge base in the above-mentioned manner as the candidate list and sends to the instant messaging group, which improves the accuracy of autonomous response of the customer service group chat robot and improves User experience. After the candidate list is generated, the server 200 continues to execute: S240: Send the candidate list to the instant messaging group.

After the server 200 generates the candidate list, it may send the candidate list to the instant messaging group. In order not to disturb the users 110 in the instant messaging group, the candidate list is marked as visible only to the first group of clients 310 in the instant messaging group, and the first group of clients 310 includes the target Some clients including client 301. The first group of clients 310 may be the target client 301 in the instant messaging group, or the client of a user with special rights, for example, the client of the group owner, the client of the administrator, etc., and However, the client selected by the target client 301, for example, after receiving the candidate list, the target client 301 can choose to display the candidate list to the designated users in the instant messaging group. The first group of clients 310 may also be clients corresponding to historical input sentences related to the input sentence in the instant messaging group. For example, the input sentence may be "How to adjust Huabei quota?", and the historical input sentence related to the input sentence may be "My Huabei quota is not enough", then the client corresponding to the historical input sentence may be said Part of the first group of clients 310.

There are many ways to make the candidate list be marked as visible only to the first group of clients 310 in the instant messaging group. For example, it can be realized by defining the attribute data of the candidate list. The attribute data can be label data, data symbols, etc., any data that can be marked. The attribute data of the candidate list is marked as visible only to the first group of clients 310 in the instant messaging group. Specifically, the label data of the candidate list is associated with the devices of the first group of clients 310. It is also possible that the attribute data and the first group of clients 310 have the same mark, so that only the first group of clients 310 can listen to the candidate list in the server 200, and therefore, only the first group of clients 310 can receive all the candidates. The candidate list. In addition, it is also possible to reserve specific display position flags in the data structure of the candidate list, and then assign values to the display flags to define which clients are visible in the instant messaging group. When the client 300 monitors the information in the instant messaging group running on the server 200, the client 300 first detects whether the value of the display flag allows it to be displayed by itself. For the first group of clients 310, the value of the display flag allows itself to display the content of the candidate list. For other clients 300, the value of the display flag does not allow itself to display the content of the candidate list. . Therefore, the display flag allows only the first group of clients 310 to monitor the candidate list in the server 200.

The candidate list can be directly sent and displayed in the instant messaging group of the first group of clients 310, or a low-invasive expression method can be used. For example, the server 200 displays a list display request to the interface of the instant messaging group of the first group of clients 310. When the user chooses to agree to display the candidate list, the server 200 may display the candidate list. When the user chooses not to agree to display the candidate list, the server 200 does not display the candidate list. The list display request may be an operable icon displayed on the instant messaging group interface of the first group of clients 310. For example, the server 200 displays a red dot icon beside the input sentence on the instant messaging group interface of the target client 301 to indicate that there is a list of candidates recommended by the robot. The candidate list is displayed only when the target user 111 touches or clicks this red dot icon. Alternatively, other part or all of the users of the first group of clients 310 can also display the candidate list by touching or clicking this red dot icon. In this way, the server 200 can interrupt the normal communication between the target user 111 and other users 110 as little as possible on the target communication interface.

In order to realize the above function, the server 200 sending the candidate list to the instant messaging group may include the following steps: the server 200 sends the first icon data to the instant messaging group, and the attribute of the first icon data is marked as only The first group of clients 310 is visible, and the first icon data includes the list display request. For example, the first icon data includes a link to the list; the server 200 receives the first group of clients 310 Any one of the clients triggers the signal of the list display request (for example, the icon is triggered on any one of the first group of clients 310), the link is activated; the server 200 sends to the instant messaging group The candidate list, the candidate list is marked as visible to the client that triggers the list display request.

The server 200 sending the candidate list to the instant messaging group may also be: the server 200 simultaneously sending the candidate list and the first icon data to the instant messaging group, wherein the first icon data includes a list display request, so The attribute of the first icon data is marked as visible only to the first group of clients 310, and the candidate list and the first icon data are defined as: first display the first icon data on the first group of clients 310, When the first icon data is triggered by any client in the first group of clients 310, the candidate list will be displayed on the client that triggered the first icon data. When any one of the clients in the first group 310 receives the candidate list and the first icon data, the client only displays the first icon data, and the first icon data may be an option. The operation icon, the client will display the candidate list on the interface only when the user triggers the operable icon.

S250: Receive a selection of the candidate list by a management client in the first group of clients 310.

As described above, the candidate list may include at least one recommended knowledge point title related to the input sentence. After the target user 111 receives the candidate list at the target client 301, he can select one or more recommended knowledge point titles that are most relevant to his question or that he cares about or want to view from the candidate list to view The content of the answer corresponding to the selected title. At this time, the target client 301 is the management client. The server 200 may receive the selection of the target user 111 on the target client 301. Of course, the management client may also be other clients in the first group of clients 310. For example, the management client may be some clients (such as an administrator) with special permissions in the first group of clients 310, and the management client may also be any client in the first group of clients 310. At this time, all users in the first group of clients 310 can make a selection on the candidate list, and the server 200 can receive a selection of the users in the first group of clients 310 on the corresponding client.

S270: Send a target answer matching the selection to the instant messaging group.

The target answer may be a target knowledge point answer corresponding to the knowledge point title selected by the target client 301. After the server 200 receives the selection of the target user 111 on the target client 301, it may send the target answer to the instant messaging group. Wherein, the target answer is marked as visible to the second group of clients 320 in the instant messaging group. The second group of clients 320 may include at least a part of the clients in the instant messaging group, and may also include the target client 301. The second group of clients 320 may be the clients corresponding to the users selected by the target user 111 from the instant messaging group, or the clients of the first group of clients 310, or the clients of the first group of clients 310. The clients selected in the candidate list may also be all clients 300. Of course, the second group of clients 320 may also be the clients corresponding to the users selected by the management client in the first group of clients 310 from the instant messaging group.

The server 200 can send the target answer to the instant messaging group in many ways. The target answer can be directly sent and displayed in the instant messaging group of the second group of clients 320, or the above-mentioned low-intrusive expression can also be used. For example, the server 200 displays an answer display request to the interface of the instant messaging group of the second group of clients 320. When the user chooses to agree to display the target answer, the server 200 may display the target answer, and when the user chooses not to agree to display the target answer In the case of the target answer, the server 200 does not display the target answer. The answer display request may be an operable icon displayed on the instant messaging group interface of the second group of clients 320. For example, the server 200 displays a red dot icon next to the candidate list on the IM group interface of the target client 301 or the target user's 111 selection of the candidate list, indicating that there is a target answer recommended by the robot. The target answer is displayed only when the target user 111 touches or clicks this red dot icon. Alternatively, other part or all of the users of the second group of clients 320 can also display the target answer by touching or clicking this red dot icon. In this way, the server 200 can interrupt the normal communication between the target user 111 and other users 110 as little as possible on the target communication interface.

The server 200 sending the target answer to the instant messaging group may include: the server 200 sending second icon data to the instant messaging group, and the attributes of the second icon data are marked as visible only to the second group of clients 320 , The second icon data includes the answer display request, for example, the first icon data includes a link to the list; the server 200 receives any client trigger in the second group of clients 320 The answer displays the requested signal (for example, the icon is triggered on any client of the second group of clients 320), the link is activated; the server 200 sends the target answer to the instant messaging group, the The target answer is marked as visible to the client that triggered the answer display request.

The server 200 sending the target answer to the instant messaging group may also be: the server 200 simultaneously sending the candidate list and the second icon data to the instant messaging group, wherein the second icon data includes an answer display request, so The attribute of the second icon data is marked as visible only to the second group of clients 320, and the target answer and the second icon data are defined as: first display the second icon data on the second group of clients 320, When the second icon data is triggered by any client in the second group of clients 320, the target answer will be displayed on the client that triggered the second icon data. When any one client in the second group of clients 320 receives the target answer and the second icon data, the client only displays the second icon data, and the second icon data may be an option The operation icon, the client will display the target answer on the interface only when the user triggers the operable icon.

When the server 200 sends the target answer to the client 300, it can perform semantic analysis according to the semantics of the input sentence or in combination with the semantics of the historical input sentence, and provide the target answer in the form of dialogue and communication according to the result of the semantic analysis, so as to improve User experience. For example, the input sentence is "What's going on? My Yu'E Bao income last month was 100, but this month only has 50!". The server 200 can answer "Customers don't need to worry, it is normal for Yu'E Bao's income to fluctuate, this is because...", and then give the target answer. This kind of intelligent dialogue response can make the user 110 more acceptable than the mechanical response, answer the user 110's doubts in a relaxed and happy atmosphere, and reduce the user 110's sensitivity and sensitivity to artificial intelligence to a certain extent. A sense of rejection, and at the same time improve the adhesion between the user 110 and the robot, and increase the utilization rate of the robot.

As shown in FIG. 3, the method P200 may further include executing by at least one processor 220 of the server 200: S280: Send an answer evaluation request to the target client 301.

S290: Receive the answer evaluation of the target client 301.

After the server 200 sends the target answer to the target client 301, the server 200 may also send an answer evaluation request to the target client 301, asking the target user 111 to rate the displayed answer content and receive the answer from the target client 301 The evaluation result is stored in the server 200 or the database 150 to facilitate the upgrade and improvement of the server 200.

In summary, the group chat robot autonomous response method P200 provided in this specification. After the server 200 receives the input sentence sent by the target user 111 in the instant messaging group, it is based on the input sentence and the preset knowledge base Generate a candidate list related to the input sentence, and send the candidate list to the target client 301 corresponding to the target user 111 or the first group of clients 310 including the target client 301 to reduce the impact on other users ; After the target user 111 makes a selection on the candidate list, the target answer matching the selection is sent to the target client 301 corresponding to the target user 111 or the second group of clients 320 including the target client 301. The method P200 can not only actively answer the questions raised by the user, but also will not cause interference to other users in the group chat. While improving work efficiency, it can also improve user experience.

Fig. 5 shows a flowchart of a method P300 for autonomous response of a group chat robot. As mentioned above, the client 300 can execute the group chat robot autonomous response method P300 provided in this specification. Specifically, at least one processor in the target client 301 can read the instruction set stored in its local storage medium and/or database 150, and then execute the group chat robot autonomously provided in this specification according to the provisions of the instruction set. Method of response P300. The method P300 may include execution by at least one processor of the target client 301: S310: Receive the input sentence and send it to the server 200.

The input sentence is input by the target user 111 on the interface of the instant messaging group.

S330: Receive the candidate list sent by the server 200 and display it on the interface of the instant messaging group.

After receiving the candidate list, the target client 301 can display the candidate list on the interface of the instant messaging group. The candidate list can be directly displayed on the instant messaging group interface, or can be displayed on the instant messaging group interface in a low-invasive manner. For example, a list display request is displayed on the instant messaging group interface, and the target user 111 can make a selection operation on the list display request. When the target user 111 chooses to agree to display the candidate list, the target client 301 displays the candidate list on the interface of the instant messaging group. When the target user 111 chooses not to agree to display the candidate list, the target client 301 does not display the candidate list on the instant messaging group interface.

Step S330 may include: receiving the candidate list and the first icon data sent by the server 200, where the first icon data includes the list display request, and the first icon data may be displayed as one on the client 300 An operable icon, and the first icon data is defined as: when the first icon data is triggered by a user, the candidate list is displayed on the client that triggered the first icon data; The triggering operation of the first icon data; and displaying the candidate list in the instant messaging group.

Fig. 6 shows a schematic diagram 600 of the instant messaging group interface provided according to an embodiment of the present specification. As shown in Figure 6, the first icon data may be an icon 620 displayed on the instant messaging group interface of the target client 301. Checking or clicking on the icon 620 means agreeing to display, and not checking or clicking on the icon. The icon represents disagreement. For example, the server 200 displays a red dot next to the input sentence 640 on the IM group interface of the target client 301 to indicate that there is a list of candidates recommended by the robot. The answer content is displayed only when the target user 111 touches or clicks on this red dot.

S350: Receive the target user 111's selection of the candidate list, and send it to the server 200.

The candidate list may include at least one recommended knowledge point. The target user 111's selection of the candidate list includes selecting from the at least one recommended knowledge point, one or more recommended knowledge points that are most relevant to their own question or that they are most concerned about or most want to view. The target client 301 may send the selection to the server 200.

S370: Send the selection to the instant messaging group.

The target client 301 may also send the selection to the instant messaging group. The selection may be marked as visible to all users or some users in the instant messaging group. The target user 111 can select visible users from the instant messaging group.

S380: Receive a target answer that matches the selection sent by the server 200.

After receiving the target answer, the target client 301 can display the target answer on the interface of the instant messaging group. The target answer may be marked as visible to all users or some users in the instant messaging group. The target user 111 can make a selection of visible users from the instant messaging group. The target answer can be directly displayed on the instant messaging group interface, or can be displayed on the instant messaging group interface in a low-invasive manner. For example, the answer display request is displayed on the instant messaging group interface, and the target user 111 can make a selection operation on the answer display request. When the target user 111 chooses to agree to display the target answer, the target client 301 displays the target answer on the instant messaging group interface. When the target user 111 chooses not to agree to display the target answer, the target client 301 does not display the target answer on the instant messaging group interface.

Step S380 may include: receiving the target answer and the second icon data sent by the server 200, where the second icon data includes the answer display request, and the second icon data may be displayed as one on the client 300 An operable icon, and the second icon data is defined as: when the second icon data is triggered by a user, the target answer is displayed on the client that triggered the second icon data; The triggering operation of the second icon data; and displaying the target answer in the instant messaging group.

In summary, the group chat robot autonomous response method P200, method P300, and system 100 provided in this specification. After the server 200 receives the input sentence sent by the target user 111 in the instant messaging group, it is based on the input sentence and the prediction. It is assumed that the knowledge base generates a candidate list related to the input sentence, and sends the candidate list to the target client 301 corresponding to the target user 111 or the first group of clients 310 including the target client 301 to Reduce the impact on other users; after the target user 111 or the management client in the first group of clients 310 makes a selection on the candidate list, the target answer that matches the selection is sent to the second group of clients 320. The method P200, the method P300, and the system 100 can not only actively answer the questions raised by the user, but also will not cause interference to other users in the group chat. While improving work efficiency, it can also improve user experience.

The foregoing describes specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims can be performed in a different order than in the embodiments and still achieve desired results. In addition, the processes depicted in the drawings do not necessarily require a specific order or sequential order to be shown in order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

In summary, after reading this detailed disclosure, those skilled in the art can understand that the foregoing detailed disclosure may be presented by way of example only, and is not restrictive. Although there is no clear description here, those skilled in the art can understand that the requirements of this specification include various reasonable changes, improvements and modifications to the embodiments. These changes, improvements and modifications are intended to be proposed by this specification and are within the spirit and scope of the exemplary embodiments of this specification.

Certain terms in this specification have been used to describe the embodiments of this specification. For example, "one embodiment", "an embodiment" or "some embodiments" means that a specific feature, structure, or characteristic described in conjunction with the embodiment may be included in at least one embodiment in this specification. Therefore, it can be emphasized and should be understood that two or more references to "an embodiment" or "one embodiment" or "alternative embodiment" in various parts of this specification do not necessarily all refer to the same embodiment. In addition, specific features, structures, or characteristics can be appropriately combined in one or more embodiments of this specification.

It should be understood that in the foregoing description of the embodiments of this specification, in order to help understand one feature, for the purpose of simplifying this specification, this specification combines various features into a single embodiment, drawings, or descriptions thereof. However, this does not mean that the combination of these features is necessary. It is entirely possible for those skilled in the art to extract some of the features as a separate embodiment to understand when reading this specification. In other words, the embodiments in this specification can also be understood as an integration of multiple sub-embodiments. It is also true when the content of each sub-embodiment is less than all the features of a single aforementioned disclosed embodiment.

Each patent, patent application, patent application publication and other materials cited herein, such as articles, books, specifications, publications, documents, articles, etc., may be incorporated herein by reference. The entire content used for all purposes, except for any related file history of the prosecution, may be inconsistent or conflict with this document, or any identical prosecution file that may have restrictive effects on the broadest scope of the claims. history. Associate with this document now or in the future. For example, if there is any inconsistency or conflict between the description, definition, or use of terms related to any contained materials, the terms in this document shall prevail. .

Finally, it should be understood that the embodiment of the application disclosed herein is an explanation of the principle of the embodiment of this specification. Other modified embodiments are also within the scope of this specification. Therefore, the embodiments disclosed in this specification are merely examples rather than limitations. Those skilled in the art can adopt alternative configurations according to the embodiments in this specification to implement the application in this specification. Therefore, the embodiments of this specification are not limited to the embodiments precisely described in the application.

Claims

A method for group chat robots to respond autonomously, including:

Receiving an input sentence, the input sentence being input by the target client on the instant messaging group interface;

Generating a candidate list based on the input sentence and a preset knowledge base, the candidate list including at least one recommended knowledge point related to the input sentence;

The candidate list is sent to the instant messaging group, and the candidate list is marked as visible only to the first group of clients in the instant messaging group, and the first group of clients includes the target client in Some clients within;

Receiving the selection of the candidate list by the management client in the first group of clients; and

Send a target answer matching the selection to the instant messaging group, where the target answer is marked as visible to a second group of clients in the instant messaging group, and the second group of clients includes the At least some clients in the instant messaging group.
8. The method for autonomous response of group chat robots according to claim 1, wherein said sending said candidate list to said instant messaging group comprises:

Sending first icon data to the instant messaging group, where the first icon data is marked as visible only to the first group of clients, and the first icon data includes a list display request;

Receiving a signal for triggering the list display request by the first group of clients;

The candidate list is sent to the instant messaging group, and the candidate list is marked as visible to the client that triggers the list display request.
8. The method for autonomous response of group chat robots according to claim 1, wherein said sending said candidate list to said instant messaging group comprises:

Sending the candidate list and the first icon data to the instant messaging group,

Wherein, the first icon data is marked as visible only to the first group of clients, the first icon data includes a list display request, and when the first icon data is triggered by the client, the The client of the first icon data displays the candidate list.
3. The method for autonomous answering of group chat robots according to claim 1, wherein said sending a target answer matching said selection to said instant messaging group comprises:

Sending the target answer and second icon data to the instant messaging group,

Wherein, the second icon data is marked as visible to the second group of clients, the second icon data includes an answer display request, and when the second icon data is triggered by the client, the first icon data is triggered 2. The client of the icon data displays the target answer.
The method for autonomous response of group chat robots according to claim 1, wherein said generating a candidate list based on said input sentence and a preset knowledge base comprises:

Identifying whether the input sentence includes a target question related to a preset transaction, the preset transaction being customized for the instant messaging group;

Determining that the input sentence includes the target question; and

Based on the input sentence and the preset knowledge base, the candidate list is generated.
The method for autonomous response of group chat robots according to claim 5, wherein said identifying whether said input sentence includes a target question related to a preset transaction comprises:

Matching the input sentence with a plurality of preset question templates to identify whether the input sentence is a question sentence; and

When the input sentence is a question sentence, classify the input sentence and identify whether the input sentence includes a target question related to the preset transaction.
The method for autonomous response of group chat robots according to claim 1, further comprising:

Sending an answer evaluation request to the instant messaging group, the answer evaluation request being marked as visible only to the target client; and

Receive the answer evaluation of the target client.
The method for autonomous response of group chat robots according to claim 1, wherein the management client includes the target client.
The method for autonomous response of a group chat robot according to claim 1, wherein the second group of clients comprises: a client selected by the management client from the instant messaging group or a client in the instant messaging group All clients.
A system for autonomous response of group chat robots, including:

At least one storage medium, including at least one instruction set, for autonomous response of group chat robots; and

At least one processor, in communication connection with the at least one storage medium,

Wherein, when the system is running, the at least one processor reads the at least one instruction set, and executes the group chat robot according to any one of claims 1-9 according to the instructions of the at least one instruction set The method of autonomous response.
A method for group chat robots to respond autonomously, including:

Receiving an input sentence and sending it to the server, the input sentence being input by the target user on the interface of the instant messaging group;

The candidate list sent by the server is received and displayed on the instant messaging group interface. The candidate list is generated based on the input sentence and the preset knowledge base, and the candidate list includes At least one recommended knowledge point related to the input sentence;

Receiving the target user's selection of the candidate list, and sending it to the server;

Sending the selection to the instant messaging group; and

Receiving a target answer that matches the choice sent by the server.
The method for group chat robot autonomous response according to claim 11, wherein said receiving the candidate list sent by the server side comprises:

Receive the candidate list and the first icon data sent by the server, where the first icon data includes a list display request, and when the first icon data is triggered by the client, the first icon data is triggered The client terminal displays the candidate list;

Receiving a trigger operation of the target user on the first icon data; and

The candidate list is displayed in the instant messaging group.
A system for autonomous response of group chat robots, including:

At least one storage medium, including at least one instruction set, for autonomous response of group chat robots; and

At least one processor, in communication connection with the at least one storage medium,

Wherein, when the system is running, the at least one processor reads the at least one instruction set, and executes the group chat robot according to any one of claims 11-12 according to the instructions of the at least one instruction set The method of autonomous response.