CN114676235A - Response method and device based on robot, computer equipment and storage medium - Google Patents
Response method and device based on robot, computer equipment and storage medium Download PDFInfo
- Publication number
- CN114676235A CN114676235A CN202210202355.3A CN202210202355A CN114676235A CN 114676235 A CN114676235 A CN 114676235A CN 202210202355 A CN202210202355 A CN 202210202355A CN 114676235 A CN114676235 A CN 114676235A
- Authority
- CN
- China
- Prior art keywords
- robot
- candidate
- robots
- intention
- content
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
- G06F16/353—Clustering; Classification into predefined classes
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
- G06F40/35—Discourse or dialogue representation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Manipulator (AREA)
Abstract
The application relates to a response method and device based on a robot, computer equipment and a storage medium. The method comprises the following steps: acquiring current conversation content of a conversation object in a current conversation; performing intention identification on the current conversation content to obtain the intention type of the current conversation content; selecting robots with the types consistent with the intention types from robots to be selected of different types to obtain a first candidate robot; determining a robot which responds to the above conversation content of the current conversation content and supports multiple rounds of responses from the robots to be selected to obtain a second candidate robot; determining at least one target candidate robot based on the first candidate robot and the second candidate robot; scheduling at least one target candidate robot to answer the current conversation content to obtain at least one answer; a target answer for the current dialog content is determined based on the at least one answer. The method can improve the response accuracy.
Description
Technical Field
The present application relates to the field of software technologies, and in particular, to a robot-based response method, apparatus, computer device, and storage medium.
Background
With the development of artificial intelligence technology, many different types of question-answering robots have appeared, such as high-frequency question-answering robots, knowledge-graph question-answering robots, chat robots, and the like. When selecting a robot to perform question answering, a corresponding robot is generally selected to perform question answering in accordance with the intention of the current conversation contents of the conversation object.
However, in the conventional method, the intention of only using the current dialog content is used for selecting the robot to respond too restrictively, and in some cases, a completely accurate response cannot be obtained, so that the response accuracy is low.
Disclosure of Invention
In view of the above, it is necessary to provide a robot-based response method, apparatus, computer device, storage medium, and computer program product capable of improving response accuracy in view of the above technical problems.
In a first aspect, the present application provides a robot-based response method. The method comprises the following steps:
acquiring current conversation content of a conversation object in a current conversation;
performing intention identification on the current conversation content to obtain an intention type of the current conversation content;
selecting robots with the types consistent with the intention types from robots to be selected of different types to obtain first candidate robots;
Determining robots which respond to the above conversation contents of the current conversation contents and support multiple rounds of responses from the robots to be selected to obtain second candidate robots;
determining at least one target candidate robot based on the first candidate robot and the second candidate robot;
scheduling the at least one target candidate robot to answer the current conversation content to obtain at least one answer;
determining a target answer for the current dialog content based on the at least one answer.
In a second aspect, the application also provides a robot-based answering machine. The device comprises:
the acquisition module is used for acquiring the current conversation content of the conversation object in the current conversation;
the first screening module is used for identifying the intention of the current conversation content to obtain the intention type of the current conversation content; selecting robots with the types consistent with the intention types from robots to be selected of different types to obtain first candidate robots;
the second screening module is used for determining a robot which responds to the above conversation contents of the current conversation contents and supports multiple rounds of responses from the to-be-selected robot to obtain a second candidate robot;
A confirmation module to determine at least one target candidate robot based on the first candidate robot and the second candidate robot;
the answer generating module is used for scheduling the at least one target candidate robot to answer the current conversation content to obtain at least one answer; determining a target answer for the current dialog content based on the at least one answer.
In one embodiment, the second candidate robot comprises a candidate robot in an incomplete state; the second screening module is further configured to determine, from the robots to be selected, a robot that needs to make multiple rounds of responses for the above conversation content of the current conversation content but does not complete the multiple rounds of responses, and obtain a candidate robot in an uncompleted state.
In one embodiment, the second candidate robot comprises the candidate robot of the most recent round of response; the second screening module is further configured to determine a robot that responds to a previous dialog content of the current dialog content and supports multiple rounds of responses from the candidate robots, and obtain a candidate robot that has responded in the latest round.
In one embodiment, the confirmation module is further configured to remove the same candidate robot from the first candidate robot and the second candidate robot to obtain the at least one target candidate robot.
In one embodiment, the intention type is a type to which an intention obtained by performing intention recognition on the current dialog content belongs; the second candidate robots comprise the candidate robot which answers in the latest round; the confirmation module is further to determine a set of candidate robots based on the first candidate robot and the second candidate robot; determining an intention matching score corresponding to the identified intention; the intent matching score is used to characterize a degree of match between the identified intent and the current dialog content; if the intention matching score is larger than a preset first threshold value, removing the candidate robot of the latest round of answer from the candidate robot set, and determining the remaining candidate robots in the candidate robot set as target candidate robots.
In one embodiment, the first screening module is further configured to perform intent recognition on the current dialog content, and determine an intent of the current dialog content; and obtaining the intention type corresponding to the determined intention according to the preset corresponding relation between the intention and the intention type.
In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor performing the steps of the robot-based answering method described above.
In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium has stored thereon a computer program which is executed by a processor for performing the steps of the robot-based answering method described above.
In a fifth aspect, the present application further provides a computer program product. The computer program product comprises a computer program which is executed by a processor for the steps of the robot-based answering method described above.
The robot-based response method, the robot-based response device, the robot-based response computer equipment, the robot-based storage medium and the robot-based response computer program product are used for acquiring the current conversation content of the conversation object in the current conversation; and performing intention identification on the current conversation content to obtain an intention type. And selecting robots with the types consistent with the intention types from the robots to be selected of different types to obtain a first candidate robot. And determining the robots which respond to the above conversation contents of the current conversation contents and support multiple rounds of responses from the robots to be selected to obtain second candidate robots, so that the candidate robots which have multiple rounds of response capabilities and store the above information of the current conversation contents are selected. Determining at least one target candidate robot based on the first candidate robot and the second candidate robot. And scheduling the at least one target candidate robot to answer the current conversation content to obtain at least one answer, wherein the obtained at least one answer can be related to the intention type and the above information of the current conversation content. Determining a target answer for the current dialog content based on the at least one answer. Therefore, a first candidate robot is selected based on the intention type, a second candidate robot is selected based on the above dialogue content of the current dialogue content, and at least one target candidate robot is screened again from the candidate robots to answer, so that an accurate target answer is obtained.
Drawings
FIG. 1 is a diagram of an environment in which a robot-based response method may be implemented in one embodiment;
FIG. 2 is a schematic flow chart diagram of a robot-based response method in one embodiment;
FIG. 3 is a schematic flow chart diagram of a robot-based response method in one embodiment;
FIG. 4 is a block diagram of a robot-based answering machine in one embodiment;
FIG. 5 is a block diagram of a robot-based answering machine in one embodiment;
FIG. 6 is a diagram of the internal structure of a computer device in one embodiment;
FIG. 7 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The robot-based response method provided by the application can be applied to the application environment shown in fig. 1. Wherein the terminal 110 communicates with the server 120 through a network. The data storage system may store data that the server 120 needs to process. The data storage system may be integrated on the server 120, or may be placed on the cloud or other network server. The terminal 110 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices, and the server 120 may be implemented by an independent server or a server cluster formed by a plurality of servers. The terminal 110 may collect the session content of the session object in the current session and send the session content to the server 120; the server 120 acquires the current conversation content; and performing intention identification on the current conversation content to obtain the intention type of the current conversation content. The server 120 selects robots with the types consistent with the intention types from the robots to be selected of different types, and obtains a first candidate robot. The server 120 determines a robot that responds to the above conversation content of the current conversation content and supports multiple rounds of responses from the candidate robots, and obtains a second candidate robot. The server 120 determines at least one target candidate robot based on the first candidate robot and the second candidate robot. The server 120 schedules at least one target candidate robot to answer the current dialog content, resulting in at least one answer. The server 120 determines a target answer for the current dialog content based on the at least one answer. The server 120 transmits the generated target to the terminal 110.
In one embodiment, as shown in fig. 2, a robot-based response method is provided, and this embodiment is illustrated by applying the method to a server, it is to be understood that the method may also be applied to a terminal, and may also be applied to a system including a terminal and a server, and is implemented by interaction between the terminal and the server. In this embodiment, the method includes the steps of:
s202, acquiring the current conversation content of a conversation object in the current conversation; performing intention identification on the current conversation content to obtain the intention type of the current conversation content; and selecting robots with the types consistent with the intention types from the robots to be selected of different types to obtain a first candidate robot.
The dialog object is a user who has performed a dialog using the robot. The current conversation content is the latest conversation content published by the conversation object and can be characters, voice or pictures and the like. The intent type is a classification of intent. It is to be understood that one intent type may correspond to a plurality of intents, since the intent type is a classification of intents.
Specifically, the terminal sends the current conversation content of the conversation object in the current conversation to the server, the server receives the current conversation content, then performs intention identification on the current conversation content, and obtains the intention type of the current conversation content according to the intention identification result. The server selects a robot with the type consistent with the intention type from a plurality of different robot types to be selected, and a first candidate robot is obtained.
In one embodiment, in the process of classifying the intentions, a domain corresponding to the intentions is determined, and then the intentions are further classified in the corresponding domain to obtain the intention type. For example, for the intention of ' why the fish fry is sticky, the field corresponding to the intention is ' food ', and then in the field of ' food ', the intention of ' why the fish fry is sticky ' is classified to obtain the corresponding intention type such as ' consultation of fish fry sticky '.
In one embodiment, the type of the robot to be selected may be at least one of a chatting robot, a graph question and answer robot, a task robot, and the like. The task robot is a robot that provides information or services under specific conditions. Usually for the purpose of satisfying a user with a specific purpose, such as traffic volume, call charges, ordering, booking tickets, consulting etc. It is understood that the type of the first candidate robot coincides with the intention type, for example, for an intention type of "consultation fry pan sticking", a question-and-answer robot may coincide with the intention type, and may be the first candidate robot of the intention type.
In one embodiment, the correspondence between the intent and the intent type is configurable.
In one embodiment, the type of candidate robot may match at least one intent type. For example, for a gourmet question-and-answer robot, two intention types of 'making gourmet' and 'buying gourmet' can be matched.
And S204, determining robots which respond to the above conversation contents of the current conversation contents and support multiple rounds of responses from the robots to be selected, and obtaining second candidate robots.
Wherein the above dialog content is at least one dialog content preceding the current dialog content. It will be appreciated that, as with the current dialog content, the above dialog content is also the dialog content that was posted by the dialog object. A single round of response refers to a question-and-answer that generally does not involve context, reference, omission, or hidden information in the dialog content. It is understood that a robot supporting a single round of response is a robot supporting a question-and-answer. A multi-round response is a multi-question multi-response as opposed to a single-round response. The robot supporting multi-turn response is a robot which recognizes multi-sentence conversation contents and responds after context matching.
Specifically, the server selects a robot that supports multiple rounds of responses to the above conversation content of the current conversation content, and takes the selected robot as a second candidate robot.
In one embodiment, the server divides the candidate robots into candidate robots supporting a single round of response and candidate robots supporting multiple rounds of response.
In one embodiment, in the process of selecting the second candidate robot, the server confirms that the robot which needs to make multiple rounds of responses to the above conversation contents of the current conversation contents but does not complete multiple rounds of responses is the candidate robot in the incomplete state.
In one embodiment, in the process of selecting the second candidate robot, the server determines that the robot which responds to the last conversation content of the current conversation content and supports multiple rounds of responses is the candidate robot which responds in the latest round.
In one embodiment, in step S208, if the target candidate robot supports multiple rounds of responses, after the target candidate robot responds to the current dialog content in the current session, the server puts the target candidate robot into the response record of the current session. It is understood that the server may select the robot responding to the above dialogue content of the current dialogue content from the response records.
S206, determining at least one target candidate robot based on the first candidate robot and the second candidate robot.
Specifically, the server screens out at least one target candidate robot from a first candidate robot and a second candidate robot according to a preset screening rule.
In one embodiment, the server may perform intent recognition on the current dialog content to obtain an intent. If the intention matching score between the intention and the current conversation content is larger than a preset first threshold value, selecting a first candidate robot and a candidate robot in an unfinished state as target candidate robots; and if the intention matching score between the intention and the current conversation content is less than or equal to a preset first threshold value, selecting a first candidate robot, a candidate robot in an uncompleted state and a candidate robot of the latest round of answer as target candidate robots.
In one embodiment, the server fails to select the first candidate robot and the second candidate robot for the current conversation content, and the server stores the current conversation content and the operation log, so that later optimization is facilitated.
In one embodiment, the server may optimize the correspondence between the intention and the intention type according to the selection result of the target candidate robot.
S208, the scheduling target candidate robot answers the current conversation content to obtain at least one answer; a target answer for the current dialog content is determined based on the at least one answer.
Specifically, the server schedules the target candidate robot to answer the current conversation content to obtain at least one answer. And after processing at least one answer, the server generates a target answer of the current conversation content.
In one embodiment, the completion status of the target candidate robot is updated if the target candidate robot supports multiple rounds of answer capabilities. Specifically, if the target candidate robot fails to answer, setting the completion state of the target candidate robot to be a completed state; and if the target candidate robot answers successfully, acquiring the target completion state of the target candidate robot for the current question, and updating the completion state of the target candidate robot for the current question to be the target completion state.
In one embodiment, the server may perform answer fusion on at least one answer to generate a target answer of the current dialog content. Specifically, the server may extract a key semantic feature from each of the at least one answer, and regenerate a new answer based on the extracted key semantic feature, where the new answer is the target answer generated after the fusion. It will be appreciated that the final generated target answer is equivalent to an answer generated in a semantic reorganization language that incorporates the individual answers.
In another embodiment, the server may schedule the target candidate robots to answer the current conversation content, resulting in at least one answer and a corresponding match score. The server may rank the at least one answer according to the matching score, thereby picking out the best ranked answer as the target answer of the current dialog content.
In one embodiment, the server may optimize the correspondence between the intent and the intent type if a first candidate robot of the target candidate robots fails to answer.
The robot-based response method acquires the current conversation content of the conversation object in the current conversation; and performing intention identification on the current conversation content to obtain an intention type. And selecting robots with the types consistent with the intention types from the robots to be selected of different types to obtain a first candidate robot. And determining the robots which respond to the above conversation contents of the current conversation contents and support multiple rounds of responses from the robots to be selected to obtain second candidate robots, so that the candidate robots which have multiple rounds of response capabilities and store the above information of the current conversation contents are selected. Determining at least one target candidate robot based on the first candidate robot and the second candidate robot. And scheduling the at least one target candidate robot to answer the current conversation content to obtain at least one answer, wherein the obtained at least one answer can be related to the intention type and the above information of the current conversation content. Determining a target answer for the current dialog content based on the at least one answer. Therefore, a first candidate robot is selected based on the intention type, a second candidate robot is selected based on the above dialogue content of the current dialogue content, and at least one target candidate robot is screened again from the candidate robots to answer, so that an accurate target answer is obtained. Moreover, on the basis of ensuring the question-answering effect, the robot-based answering method does not need all robots to answer, so that unnecessary robot calling is reduced, the resource expenditure is saved, and the answer fusion complexity can be reduced.
In one embodiment, the second candidate robot comprises a candidate robot in an incomplete state; determining a robot which responds to the above conversation content of the current conversation content and supports multiple rounds of responses from the candidate robots, and obtaining a second candidate robot comprises: and determining the robots which need to make multiple rounds of responses aiming at the above conversation contents of the current conversation contents but do not finish multiple rounds of responses from the robots to be selected, and obtaining the candidate robots in the uncompleted state.
The incomplete state refers to a state that multiple rounds of responses need to be made for the above conversation content of the current conversation content, but multiple rounds of responses are not completed.
Specifically, the server determines a robot which needs to make multiple rounds of responses to the above conversation content of the current conversation content but does not complete multiple rounds of responses from the robots to be selected, and obtains a candidate robot in an uncompleted state.
In one embodiment, in the process of executing step S208, if the target candidate robot is a robot supporting multiple rounds of responses, the server records the completion status of the target candidate robot and puts the target candidate robot into the response record. It is understood that the server may determine, from the response record of the above dialog content of the current dialog content, a robot that needs to make multiple rounds of responses to the above dialog content of the current dialog content, but does not complete multiple rounds of responses. It can be understood that by recording the completion status information of the target candidate robot, the robot which does not complete the multiple rounds of responses is guaranteed to have the opportunity to continue to respond in the next round, thereby guaranteeing the continuity of the responses to the current conversation content.
In one implementation, the session management module of the server is responsible for updating the completion status of the target candidate robot.
In this embodiment, a robot that needs to make multiple rounds of responses for the above dialog content of the current dialog content but does not complete multiple rounds of responses is determined, so as to obtain a candidate robot in an uncompleted state, and the candidate robot in the uncompleted state is determined as a second candidate robot. In this way, the server picks out robots that support multiple rounds of responses but do not complete responses to respond to the current dialog content, so that the generated answers are generated based on the above dialog content and the current dialog content, thereby laying the basis for the accuracy of the responses for the generation and screening of answers in step S208.
In one embodiment, the second candidate robot comprises the candidate robot of the most recent round of answer; determining a robot which responds to the above conversation content of the current conversation content and supports multiple rounds of responses from the robots to be selected, and obtaining a second candidate robot comprises: and determining the robots which respond to the last conversation content of the current conversation content and support multiple rounds of responses from the robots to be selected, and obtaining the candidate robots which respond to the latest round.
Specifically, the server determines a robot which responds to the last conversation content of the current conversation content and supports multiple rounds of responses from the robots to be selected, and obtains a candidate robot which responds in the latest round.
In one embodiment, in the process of executing step S208, if the target candidate robot is a robot supporting multiple rounds of responses, the server puts the target candidate robot and the time point of the dialog content of the response into the response record. It is understood that the server can confirm the robot responding to the dialog content last to the current dialog content by responding to the time point of the dialog content of the response recorded in the record.
In this embodiment, a robot that responds to a previous dialog content of the current dialog content and supports multiple rounds of responses is determined to obtain a candidate robot that has responded in the latest round, and the candidate robot that has responded in the latest round is determined as a second candidate robot. In this way, the server picks out a robot that responded to the last dialog content of the current dialog content and supports multiple rounds of responses to respond to the current dialog content, so that the generated answer is generated based on the last dialog content and the current dialog content, thereby preparing for improving the accuracy of the response for the generation and screening of the answer in step S208.
In one embodiment, determining the target candidate robot based on the first candidate robot and the second candidate robot comprises: and removing the same candidate robot from the first candidate robot and the second candidate robot to obtain the target candidate robot.
Specifically, after determining the first candidate robot and the second candidate robot, if the first candidate robot and the second candidate robot have the same candidate robot, the server removes the same candidate robot to obtain the target candidate robot. It will be appreciated that the removal of the same robot may avoid the same robot making the same answer.
In this embodiment, the target candidate robot is obtained by removing the same candidate robot from the first candidate robot and the second candidate robot, and the same robot is prevented from making the same answer, so that the waste of computer resources is avoided, and the utilization rate of the computer resources is improved.
In one embodiment, the intention type is a type to which an intention obtained by intention recognition of the current dialog content belongs; the second candidate robots comprise the candidate robot which answers in the latest round; determining the target candidate robot based on the first candidate robot and the second candidate robot includes: determining a set of candidate robots based on the first candidate robot and the second candidate robot; determining an intention matching score corresponding to the identified intention; the intention matching score is used for representing the matching degree between the recognized intention and the current dialogue content; and if the intention matching score is larger than a preset first threshold value, removing the candidate robot of the latest round of answer from the candidate robot set, and determining the remaining candidate robots in the candidate robot set as target candidate robots.
Specifically, the server determines a set of a first candidate robot and a second candidate robot as a candidate robot set. The server determines an intent matching score corresponding to the identified intent. And the service judges the size of the intention matching score, if the intention matching score is larger than a preset first threshold value, the candidate robot which has the latest answer in the round is removed from the candidate robot set, and the remaining candidate robots in the candidate robot set are determined as target candidate robots.
In this embodiment, if the intention matching score is greater than a preset first threshold, the candidate robot of the latest round of answer is removed from the candidate robot set, and the remaining candidate robots in the candidate robot set are determined as the target candidate robots. In this way, the server may select only the first candidate robot and the unfinished candidate robot as the target candidate robot to respond when the matching degree between the identified intention and the current conversation content reaches the preset value, thereby avoiding waste of computer resources.
In one embodiment, intention recognition is carried out on the current conversation content, and the intention type of the current conversation content is obtained; performing intention identification on the current conversation content, and determining the intention of the current conversation content; and obtaining an intention type corresponding to the determined intention according to the preset intention and the corresponding relation between the intention types.
Specifically, the server pre-configures the correspondence between the intentions and the intention types. The server identifies the intention of the current conversation content, determines the intention of the current conversation content, and obtains the intention type corresponding to the determined intention from the corresponding relation.
In one embodiment, the server may determine at least one intent of the current dialog content and derive a corresponding at least one intent type from the correspondence.
In one embodiment, the server performs intent recognition on the current dialog content, and may determine a plurality of intents and respective intent matching scores for the current dialog content. The server ranks the plurality of intentions based on intention matching scores of the plurality of intentions, and obtains at least one intention of which the matching score is larger than a preset second threshold. The server obtains at least one corresponding intention type from the corresponding relation. The server correspondingly selects the first candidate robot which is consistent with each intention type. It is to be understood that the first threshold is greater than a second threshold, the second threshold being used to screen out the first candidate robot, the first threshold being used to remove the candidate robot for the most recent round of response. Specifically, after the first candidate robot is screened out, the server judges whether intentions with intention matching scores larger than a preset first threshold exist, and if the intentions exist, the server removes the candidate robot with the latest round of response from the candidate robot set.
In one embodiment, intent recognition is performed on the current dialog content by an intent recognition module of the server, resulting in a corresponding intent and intent type.
In this embodiment, an intention type corresponding to the intention of the current dialog content is obtained according to a pre-configured correspondence between the intention and the intention type, and then the first candidate robot is selected according to the intention type, so that the answer generated by the first candidate robot is related to the intention type, thereby preparing for the generation of the answer in step S208 to improve the accuracy of the response.
In one embodiment, as shown in FIG. 3, a flow chart of a robot-based answering method is presented. Specifically, the server obtains current conversation content of a conversation object in the current conversation. In the selection stage, the server performs intention identification on the current conversation content to obtain the intention type of the current conversation content. And the server selects robots with the types consistent with the intention types from the robots to be selected of different types according to the corresponding relation between the intention and the intention types to obtain a first candidate robot. And the server determines the robots which respond to the above conversation contents of the current conversation contents and support multiple rounds of responses from the robots to be selected to obtain second candidate robots. In the process of obtaining the second candidate robot, the server determines the robot which needs to make multiple rounds of responses to the above conversation content of the current conversation content but does not complete multiple rounds of responses, and obtains the candidate robot in an uncompleted state. And the server determines the robot which responds to the last conversation content of the current conversation content and supports multiple rounds of responses, and obtains the candidate robot which responds in the latest round. In the cutting stage, the server cuts the candidate robot in an incomplete state and the candidate robot in the latest round of response in the first candidate robot and the second candidate robot, the same robot is removed, and when the matching score corresponding to the intention is larger than a preset first threshold value, the candidate robot in the latest round of response is not selected. And the server cuts the candidate robot to obtain the target candidate robot. In the calling phase, the server calls the target candidate robot to respond to the current conversation content to generate at least one answer. In the generation stage, the server selects and fuses at least one answer to obtain a final target answer. It can be understood that, in the robot-based response method according to this embodiment, all robots are not directly made to respond, but candidate robots are selected in the selection stage and the clipping stage in advance, so that complexity of processing target answers in the generation stage is reduced, workload of the candidate robots is reduced, and processing performance of the entire candidate robot cluster is improved.
It should be understood that, although the steps in the flowcharts in the embodiments of the present application are shown in sequence as indicated by the arrows, the steps are not necessarily performed in sequence as indicated by the arrows. The steps are not limited to being performed in the exact order illustrated and, unless explicitly stated herein, may be performed in other orders. Moreover, at least a part of the steps in the flowcharts may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of performing the steps or stages is not necessarily performed in sequence, but may be performed alternately or alternately with other steps or at least a part of the steps or stages in other steps.
Based on the same inventive concept, the embodiment of the application also provides a robot-based response device for realizing the robot-based response method. The solution to the problem provided by the device is similar to the solution described in the above method, so the specific limitations in one or more embodiments of the robot-based response device provided below may refer to the limitations on the robot-based response method described above, and are not described herein again.
In one embodiment, as shown in fig. 4, there is provided a robot-based answering machine 400 comprising: an obtaining module 402, a first screening module 404, a second screening module 406, a confirming module 408, and an answer generating module 410, wherein:
an obtaining module 402, configured to obtain current dialog contents of a dialog object in a current session.
A first screening module 404, configured to perform intent recognition on current dialog content to obtain an intent type of the current dialog content; and selecting robots with the types consistent with the intention types from the robots to be selected of different types to obtain a first candidate robot.
The second screening module 406 is configured to determine a robot that responds to the above dialog content of the current dialog content and supports multiple rounds of responses from the robots to be selected, so as to obtain a second candidate robot.
A validation module 408 for determining at least one target candidate robot based on the first candidate robot and the second candidate robot.
An answer generation module 410, configured to schedule at least one target candidate robot to answer the current conversation content, so as to obtain at least one answer; a target answer for the current dialog content is determined based on the at least one answer.
In one embodiment, the second candidate robot comprises a candidate robot in an incomplete state; the second screening module 406 is further configured to determine, from the robots to be selected, a robot that needs to make multiple rounds of responses to the above conversation content of the current conversation content but does not complete multiple rounds of responses, and obtain a candidate robot in an uncompleted state.
In one embodiment, the second candidate robot comprises the candidate robot of the most recent round of response; the second screening module 406 is further configured to determine a robot that responds to a previous dialog content of the current dialog content and supports multiple rounds of responses from the robots to be selected, so as to obtain a candidate robot that has responded in the latest round.
In one embodiment, the validation module 408 is further configured to remove the same candidate robot from the first candidate robot and the second candidate robot to obtain at least one target candidate robot.
In one embodiment, the intention type is a type to which an intention obtained by intention recognition of the current dialog content belongs; the second candidate robots include the candidate robot which has responded in the latest round.
In one embodiment, as shown in fig. 5, a robot-based answering machine 500 is provided that includes a third screening module 412, wherein: a third screening module 412 to determine a set of candidate robots based on the first candidate robot and the second candidate robot; determining an intention matching score corresponding to the identified intention; the intent-matching score is used to characterize the degree of match between the identified intent and the current dialog content. And if the intention matching score is larger than a preset first threshold value, removing the candidate robot of the latest round of answer from the candidate robot set, and determining the remaining candidate robots in the candidate robot set as target candidate robots.
In one embodiment, the first filtering module 404 is further configured to perform intent recognition on the current dialog content, and determine an intent of the current dialog content; and obtaining an intention type corresponding to the determined intention according to the preset intention and the corresponding relation between the intention types.
The robot-based response device acquires the current conversation content of the conversation object in the current conversation; and performing intention identification on the current conversation content to obtain an intention type. And selecting robots with the types consistent with the intention types from the robots to be selected of different types to obtain a first candidate robot. And determining the robots which respond to the above conversation contents of the current conversation contents and support multiple rounds of responses from the robots to be selected to obtain second candidate robots, so that the candidate robots which have multiple rounds of response capabilities and store the above information of the current conversation contents are selected. Determining at least one target candidate robot based on the first candidate robot and the second candidate robot. And scheduling the at least one target candidate robot to answer the current conversation content to obtain at least one answer, wherein the obtained at least one answer can be related to the intention type and the above information of the current conversation content. A target answer for the current dialog content is determined based on at least one answer. Therefore, a first candidate robot is selected based on the intention type, a second candidate robot is selected based on the above dialogue content of the current dialogue content, and at least one target candidate robot is screened again from the candidate robots to answer, so that an accurate target answer is obtained. Moreover, on the basis of ensuring the question-answering effect, the robot-based answering device does not need all robots to answer, thereby reducing unnecessary robot calls, saving resource expenditure and reducing the complexity of answer fusion.
For the specific limitations of the robot-based answering device, reference may be made to the limitations of the robot-based answering method described above, which are not described herein again. The various modules in the robot-based answering machine described above may be implemented in whole or in part by software, hardware, and combinations thereof. The modules can be embedded in a hardware form or independent of a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in fig. 6. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing the robot response information data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a robot-based answering method.
In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 7. The computer apparatus includes a processor, a memory, an input/output interface, a communication interface, a display unit, and an input device. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input/output interface. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The input/output interface of the computer device is used for exchanging information between the processor and an external device. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a robot-based answering method. The display unit of the computer equipment is used for forming a visual and visible picture, and can be a display screen, a projection device or a virtual reality imaging device, the display screen can be a liquid crystal display screen or an electronic ink display screen, the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, and can also be an external keyboard, a touch pad or a mouse and the like.
Those skilled in the art will appreciate that the configurations shown in fig. 6 and 7 are merely block diagrams of portions of configurations related to aspects of the present application, and do not constitute limitations on the computing devices to which aspects of the present application may be applied, as particular computing devices may include more or less components than shown, or combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is further provided, which includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the above method embodiments when executing the computer program.
In an embodiment, a computer-readable storage medium is provided, on which a computer program is stored which, when being executed by a processor, carries out the steps of the above-mentioned method embodiments.
In an embodiment, a computer program product is provided, comprising a computer program which, when being executed by a processor, carries out the steps of the above-mentioned method embodiments.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware related to instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above examples only express several embodiments of the present application, and the description thereof is more specific and detailed, but not to be construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A robot-based response method, the method comprising:
acquiring current conversation content of a conversation object in a current conversation;
performing intention identification on the current conversation content to obtain an intention type of the current conversation content;
selecting robots with the types consistent with the intention types from robots to be selected of different types to obtain first candidate robots;
Determining a robot which responds to the above conversation content of the current conversation content and supports multiple rounds of responses from the to-be-selected robots to obtain a second candidate robot;
determining at least one target candidate robot based on the first candidate robot and the second candidate robot;
scheduling the at least one target candidate robot to answer the current conversation content to obtain at least one answer;
determining a target answer for the current dialog content based on the at least one answer.
2. The method of claim 1, wherein the second candidate robot comprises a candidate robot in an incomplete state;
the step of determining a robot which responds to the above conversation contents of the current conversation contents and supports multiple rounds of responses from the candidate robots to obtain a second candidate robot comprises:
and determining the robots which need to make multiple rounds of responses aiming at the above conversation contents of the current conversation contents but do not finish the multiple rounds of responses from the robots to be selected, and obtaining the candidate robots in an uncompleted state.
3. The method of claim 1, wherein the second candidate robot comprises a candidate robot of a most recent round of response;
The determining, from the robots to be selected, a robot that responds to the above dialog content of the current dialog content and supports multiple rounds of responses, and obtaining a second candidate robot includes:
and determining the robot which responds to the last conversation content of the current conversation content and supports multiple rounds of responses from the candidate robots to obtain the candidate robot which responds to the latest round of responses.
4. The method of claim 1, wherein the determining at least one target candidate robot based on the first candidate robot and the second candidate robot comprises:
removing the same candidate robot from the first candidate robot and the second candidate robot to obtain the at least one target candidate robot.
5. The method according to claim 1, wherein the intent type is a type to which an intent obtained by intent recognition of the current dialog content belongs; the second candidate robots comprise the candidate robot which answers in the latest round; the determining at least one target candidate robot based on the first candidate robot and the second candidate robot comprises:
Determining a set of candidate robots based on the first candidate robot and the second candidate robot;
determining an intention matching score corresponding to the identified intention; the intent matching score is used to characterize a degree of match between the identified intent and the current dialog content;
if the intention matching score is larger than a preset first threshold value, removing the candidate robot of the latest round of answer from the candidate robot set, and determining the remaining candidate robots in the candidate robot set as target candidate robots.
6. The method of claim 1, wherein the performing intent recognition on the current dialog content, and obtaining an intent type of the current dialog content comprises;
performing intention identification on the current conversation content, and determining the intention of the current conversation content;
and obtaining the intention type corresponding to the determined intention according to the preset corresponding relation between the intention and the intention type.
7. A robot-based answering device, characterized in that the device comprises:
the acquisition module is used for acquiring the current conversation content of the conversation object in the current conversation;
The first screening module is used for performing intention identification on the current conversation content to obtain the intention type of the current conversation content; selecting robots with the types consistent with the intention types from robots to be selected of different types to obtain first candidate robots;
the second screening module is used for determining robots which respond to the above conversation contents of the current conversation contents and support multiple rounds of responses from the robots to be selected to obtain second candidate robots;
a confirmation module to determine at least one target candidate robot based on the first candidate robot and the second candidate robot;
the answer generation module is used for scheduling the at least one target candidate robot to answer the current conversation content to obtain at least one answer; determining a target answer for the current dialog content based on the at least one answer.
8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 6.
9. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 6.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210202355.3A CN114676235A (en) | 2022-03-02 | 2022-03-02 | Response method and device based on robot, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210202355.3A CN114676235A (en) | 2022-03-02 | 2022-03-02 | Response method and device based on robot, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114676235A true CN114676235A (en) | 2022-06-28 |
Family
ID=82072064
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210202355.3A Pending CN114676235A (en) | 2022-03-02 | 2022-03-02 | Response method and device based on robot, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114676235A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108319599A (en) * | 2017-01-17 | 2018-07-24 | 华为技术有限公司 | Method and device for man-machine dialogue |
| CN111694939A (en) * | 2020-04-28 | 2020-09-22 | 平安科技（深圳）有限公司 | Method, device and equipment for intelligently calling robot and storage medium |
| CN111901220A (en) * | 2019-05-06 | 2020-11-06 | 华为技术有限公司 | Method for determining chat robot and response system |
| CN112365892A (en) * | 2020-11-10 | 2021-02-12 | 杭州大搜车汽车服务有限公司 | Man-machine interaction method, device, electronic device and storage medium |
| CN114117023A (en) * | 2022-01-26 | 2022-03-01 | 深圳追一科技有限公司 | Interaction method, interaction device, electronic equipment and storage medium |
-
2022
- 2022-03-02 CN CN202210202355.3A patent/CN114676235A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108319599A (en) * | 2017-01-17 | 2018-07-24 | 华为技术有限公司 | Method and device for man-machine dialogue |
| CN111901220A (en) * | 2019-05-06 | 2020-11-06 | 华为技术有限公司 | Method for determining chat robot and response system |
| CN111694939A (en) * | 2020-04-28 | 2020-09-22 | 平安科技（深圳）有限公司 | Method, device and equipment for intelligently calling robot and storage medium |
| CN112365892A (en) * | 2020-11-10 | 2021-02-12 | 杭州大搜车汽车服务有限公司 | Man-machine interaction method, device, electronic device and storage medium |
| CN114117023A (en) * | 2022-01-26 | 2022-03-01 | 深圳追一科技有限公司 | Interaction method, interaction device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111814089B (en) | Page rendering method and device, rendering server and storage medium | |
| CN113392158A (en) | Service data processing method and device and data center | |
| US20170228107A1 (en) | Generating predictive action buttons within a graphical user interface | |
| CN113886563B (en) | Question recommending method and device, electronic equipment and storage medium | |
| CN106815254A (en) | A kind of data processing method and device | |
| CN104461710A (en) | Method and device for processing tasks | |
| CN105763420A (en) | Automatic information reply method and device | |
| CN111343348A (en) | Application method, system, device and storage medium of interactive voice response menu | |
| CN113836405B (en) | Information query method, device and computer-readable storage medium | |
| CN111651989A (en) | Named entity recognition method and device, storage medium and electronic device | |
| CN113807912A (en) | Method and computer program product for dynamically matching user and broker connections | |
| CN107633080B (en) | User task processing method and device | |
| US11562126B2 (en) | Coaching system and coaching method | |
| CN114676235A (en) | Response method and device based on robot, computer equipment and storage medium | |
| US12356042B2 (en) | System and method for integration of automatic response generating systems with non-API applications | |
| CN119128098A (en) | Multi-source fusion question-answering method and device, storage medium, and electronic device | |
| CN118245525A (en) | Content searching method, device, electronic equipment and computer readable storage medium | |
| CN111163237B (en) | Call service flow control method and related device | |
| CN110413916A (en) | The method and apparatus of the topic page for rendering | |
| CN117806951A (en) | Intelligent scheduling method and related equipment applied to distributed operation of test cases | |
| CN116795256A (en) | Task processing method and device, electronic equipment and storage medium | |
| CN111897942B (en) | Dialogue robot problem processing method, device and equipment combining RPA and AI | |
| EP3468150B1 (en) | Handling questions of conversational services | |
| CN119396982B (en) | Identification method and electronic device for multi-round multi-intention mixed dialogue | |
| US12147788B2 (en) | System and method for intelligently generating code for use in integrating automated response generating systems with non-API applications |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |