CN117880230A - Group dialogue method and device, electronic equipment and computer readable medium - Google Patents

Abstract

The disclosure provides a group dialogue method and device, relates to the technical field of artificial intelligence, and particularly relates to the technical fields of natural language processing, computer vision, deep learning and the like. The specific implementation scheme is as follows: acquiring dialogue information of a dialogue main body in a group dialogue; determining a virtual main body to be added into the group dialogue and a problem to be processed based on the dialogue message; obtaining reply information of the virtual main body based on the virtual main body and the problem to be processed; and adding a virtual main body in the group dialogue, and sending a reply message in the group dialogue through the virtual main body. This embodiment improves the dialog experience of the dialog body in the group dialog.

Description

Group dialogue method and device, electronic equipment and computer readable medium

Technical Field

The present disclosure relates to the field of artificial intelligence, and in particular, to the technical fields of natural language processing, computer vision, deep learning, and the like, and more particularly, to a group dialogue method and apparatus, an electronic device, a computer readable medium, and a computer program product.

Background

In a common message session, which is generally one-to-one, other real people can be added, and virtual subjects with history roles, such as a multi-la A dream, can be added to create a group session, and the virtual subjects can be driven by a large language model.

For the current group dialogue, the virtual main body is generally directly enabled to respond to the intra-group message, the content possibly responded cannot meet the requirements of each main body in the group dialogue, and the reply of the virtual main body is easy to be disordered and the order is easy to be disordered.

Disclosure of Invention

A group conversation method and apparatus, an electronic device, a computer-readable storage medium, and a computer program product are provided.

According to a first aspect, there is provided a group dialogue method comprising: acquiring dialogue information of a dialogue main body in a group dialogue; determining a virtual main body to be added into the group dialogue and a problem to be processed based on the dialogue message; obtaining reply information of the virtual main body based on the virtual main body and the problem to be processed; and adding a virtual main body in the group dialogue, and sending a reply message in the group dialogue through the virtual main body.

According to a second aspect, there is provided a group dialogue device comprising: an acquisition unit configured to acquire a conversation message of a conversation body in a group conversation; a determining unit configured to determine a virtual body to be added to the group conversation and a problem to be processed based on the conversation message; the obtaining unit is configured to obtain reply information of the virtual main body based on the virtual main body and the problem to be processed; and the sending unit is configured to add a virtual main body in the group dialogue and send a reply message in the group dialogue through the virtual main body.

According to a third aspect, there is provided an electronic device comprising: at least one processor; and a memory communicatively coupled to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method as described in any one of the implementations of the first aspect.

According to a fourth aspect, there is provided a non-transitory computer readable storage medium storing computer instructions for causing a computer to perform a method as described in any implementation of the first aspect.

According to a fifth aspect, there is provided a computer program product comprising a computer program which, when executed by a processor, implements a method as described in any of the implementations of the first aspect.

The embodiment of the disclosure provides a group conversation method and device, firstly, conversation messages of conversation main bodies in a group conversation are obtained; secondly, determining a virtual main body to be added into the group dialogue and a problem to be processed based on the dialogue message; thirdly, based on the virtual main body and the problem to be processed, obtaining reply information of the virtual main body; and finally, adding a virtual main body in the group dialogue, and sending a reply message in the group dialogue through the virtual main body. Thus, based on the dialogue information of the dialogue main body in the group dialogue, the virtual main body is determined, and the obtained virtual main body can be matched with the dialogue information in the group dialogue; based on the dialogue message, the problem to be processed is determined, so that the center thought of the dialogue message and the problem to be solved can be effectively analyzed, and the reliability of obtaining the virtual main body reply information is improved; the virtual main body is added in the group dialogue, and the virtual main body is controlled to send out the reply information, so that the effectiveness and the reliability of dialogue contents in the group dialogue can be ensured, and the dialogue experience of the dialogue main body in the group dialogue is improved.

It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.

Drawings

The drawings are for a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:

FIG. 1 is a flow chart of one embodiment of a group conversation method according to the present disclosure;

FIG. 2 is a flow chart of another embodiment of a group conversation method according to the present disclosure;

FIG. 3 is a schematic diagram of an embodiment of a group conversation device in accordance with the present disclosure;

fig. 4 is a block diagram of an electronic device used to implement a group conversation method of an embodiment of the present disclosure.

Detailed Description

Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

In this embodiment, "first" and "second" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include at least one such feature.

In response to the problem of poor experience of a conversation body in a group conversation in the conventional technology, the present disclosure provides a group conversation method, by which fig. 1 shows a flow 100 according to one embodiment of the group conversation method of the present disclosure, the group conversation method includes the following steps:

step 101, obtaining a dialogue message of a dialogue main body in a group dialogue.

In this embodiment, the group conversation is a conversation having a plurality of conversation subjects, where a conversation subject is a subject that participates in the group conversation and is currently speaking or has been speaking in the group conversation, and the conversation subject may be a real person, a virtual person, or a virtual history role. For example, there are three dialog bodies in a group dialog of a plurality of dialog bodies currently generated through a dialog box, and the three dialog bodies include a person, a duola a dream, and a Qin king.

In this embodiment, the dialogue message may be the current or historical language information of a plurality of dialogue subjects in the group dialogue, for example, the dialogue message is the historical language of all the dialogue subjects in the past five days from the current start to the history. The dialog message may include an image, text, or voice, and the topic of the dialog message may be determined by processing the dialog message. Wherein processing the conversation message includes: and converting the voice in the dialogue message into text or carrying out semantic recognition on the image in the dialogue message to determine the concrete representation meaning of the image in the dialogue message.

Step 102, determining a virtual body to be added to the group conversation and a problem to be processed based on the conversation message.

In this embodiment, the dialogue message is the speaking content of at least one dialogue body in the group dialogue, the subject of the current dialogue message can be determined by the content of the dialogue message, the virtual body that can participate in the group dialogue can be determined by the content of the dialogue message, specifically, the virtual body to be added to the group dialogue can be determined by the keyword with the same name as the virtual body in the content of the dialogue message; optionally, when the virtual body is related to the history role, determining the history role related to the description feature word by the description feature word related to the virtual body in the content of the dialogue message, wherein the description feature word is a word describing the characteristics of a dialogue body, and taking the determined history role as the virtual body.

In this embodiment, the problem to be processed is a problem that all dialogue bodies in the current group dialogue are confused or interested after the group dialogue method summarizes the dialogue messages. The generated virtual main body replies to the problem to be processed, so that the experience of each dialog main body in the group dialog can be improved. In step 102, determining the problem to be processed based on the dialogue message includes: based on the dialog messages, dialog topics for the dialog body discussion are determined, and based on the dialog topics, questions to be processed are determined. The questions to be processed may be all questions related to the dialogue topic.

And step 103, obtaining reply information of the virtual main body based on the virtual main body and the problem to be processed.

In this embodiment, the reply message is reply content of the virtual body in the group dialogue, and the reply message may include text, image and voice. The reply content is also information obtained by feeding back the problem to be processed.

According to the group conversation method provided by the embodiment, the reply information of the virtual main body can be determined according to the context analysis result, the reply information is the speaking content which the virtual main body needs to send in the group conversation, and the virtual main body can keep consistent with the conversation content of each conversation main body in the current group conversation through the reply information.

Step 104, adding a virtual main body in the group dialogue, and sending a reply message in the group dialogue through the virtual main body.

In this embodiment, adding the virtual body to the group session refers to making the virtual body participate in the group session as a session body, and the output information of the virtual body in the group session is reply information, which is information obtained after summarizing the session message.

Optionally, the group dialogue method may further include: sending the information to be confirmed of the reply information to the group dialogue; and responding to the received determination information corresponding to the information to be confirmed, and determining that the reply information is accurate. In this embodiment, the information to be confirmed is information for inquiring about the reply information of the dialog subjects in the group dialog, for example, the reply information of each dialog subject is inquired whether the reply information is accurate by the information to be confirmed, and the determination information corresponding to the information to be confirmed refers to: the reply information is accurate.

Optionally, the group dialogue method may further include: and inputting the virtual main body and the reply information into a verification large language model to obtain a verification result output by the verification large language model, wherein the verification large language model is used for verifying whether the virtual main body and the reply information are consistent or not, namely whether the reply information is consistent with the role setting of the virtual main body or not. Wherein, the verification result comprises: the reply message matches the role of the virtual principal or the reply message does not match the role of the virtual principal, and the reply message needs to be regenerated.

The group dialogue method provided by the embodiment of the disclosure includes the steps of firstly, acquiring dialogue messages of dialogue main bodies in a group dialogue; secondly, determining a virtual main body to be added into the group dialogue and a problem to be processed based on the dialogue message; thirdly, based on the virtual main body and the problem to be processed, obtaining reply information of the virtual main body; and finally, adding a virtual main body in the group dialogue, and sending a reply message in the group dialogue through the virtual main body. Thus, based on the dialogue information of the dialogue main body in the group dialogue, the virtual main body is determined, and the obtained virtual main body can be matched with the dialogue information in the group dialogue; based on the dialogue message, the problem to be processed is determined, so that the center thought of the dialogue message and the problem to be solved can be effectively analyzed, and the reliability of obtaining the virtual main body reply information is improved; the virtual main body is added in the group dialogue, and the virtual main body is controlled to send out the reply information, so that the effectiveness and the reliability of dialogue contents in the group dialogue can be ensured, and the dialogue experience of the dialogue main body in the group dialogue is improved.

In the group dialogue, the group dialogue has real people and virtual people, and the real people can send out the conversation according to the dialogue information of the dialogue main body, so that the problem is solved; but how to mobilize and coordinate virtual characters, start speaking at what time, say what theme, make it appear more like a real person to respond randomly and strongly up and down, otherwise it appears that this virtual character is especially false, and that the virtual character will respond directly to messages in the group very messy, and for each virtual scheduler of content to respond, there will be order confusion, while creating endless loop conversations, in order to solve these problems, the present disclosure provides another embodiment of a group conversation method, fig. 2 shows a flow 200 according to another embodiment of the group conversation method of the present disclosure, the group conversation method including the steps of:

Step 201, a dialogue message of a dialogue body in a group dialogue is acquired.

Step 202, determining virtual subjects to be added to the group conversation and questions to be processed based on the conversation message.

And 203, obtaining reply information of the virtual main body based on the virtual main body and the problem to be processed.

It should be understood that the operations and features in steps 201-203 described above correspond to those in steps 101-103, respectively, and thus the descriptions of the operations and features in steps 101-103 described above are equally applicable to steps 201-203, and are not repeated here.

Step 204, determining the speaking time of the virtual main body based on the dialogue message.

In this embodiment, the speaking opportunity includes speaking time, and when the speaking opportunity is speaking time, as long as the current time of the group chat reaches the speaking time, the virtual main body is directly added into the group dialogue, so that the dialogue in the group dialogue is not abrupt for the addition of the virtual main body, and the experience of the dialogue main body is improved.

Step 205, adding a virtual body in the group dialogue based on the speaking time, and sending a reply message in the group dialogue through the virtual body.

In this embodiment, adding a virtual body to a group session means: a dialog system for enabling a virtual subject to participate in a group dialog as a dialog subject, wherein the dialog system comprises: a dialog box in which a plurality of dialog bodies participate, a plurality of dialog bodies in group chat are provided in the dialog box, and after a virtual body is added in the group dialog, a head portrait corresponding to the virtual body is provided in the dialog box. The above-mentioned sending a reply message in a group session through the virtual body means: the reply information is displayed in the dialog box of the dialog system in the head portrait pair position of the virtual main body.

In this embodiment, the speaking time includes speaking time, and the step 205 includes: detecting whether the current time reaches speaking time; and in response to the current time reaching the speaking time, adding a virtual main body in the group dialogue, and sending a reply message in the group dialogue through the virtual main body.

Optionally, the speaking opportunity includes: keywords associated with the virtual body in the conversation message, step 205 includes: detecting whether a keyword exists in the dialogue message; and in response to the keywords in the dialogue message, adding a virtual main body in the group dialogue, and sending out reply information in the group dialogue through the virtual main body.

For some virtual bodies with history roles, the virtual bodies can be added into the group dialogue first, and when the group dialogue reaches a certain program, the virtual bodies send reply information in the group dialogue according to speaking time. Specifically, the step 205 includes: adding a virtual main body into the group dialogue, and displaying an icon corresponding to the virtual main body in a specific area of a dialogue box of the group dialogue; and responding to the fact that the current time is equal to the speaking time in the speaking opportunity, or the dialogue message has the key words corresponding to the virtual main body in the speaking opportunity, and sending reply information in the group dialogue through the virtual main body.

The group dialogue method provided by the disclosure runs and an execution subject on the group dialogue method is a background virtual dispatcher, the virtual dispatcher can also be driven by a large predictive model, the virtual dispatcher cannot be seen by a real person and a virtual subject in the group dialogue, and the virtual dispatcher can determine whether the round of speaking, speaking content, language and the like of a certain virtual person is achieved according to the context.

The group dialogue method provided by the disclosure can effectively mobilize and coordinate the virtual main body, so that the virtual main body is more similar to a real person to respond randomly and in a strong context, and the speaking time of the virtual main body can be determined according to the analysis result of dialogue information, so that the virtual main body can speak at a proper time.

The group dialogue method provided by the embodiment of the disclosure determines the speaking time of the virtual main body based on dialogue information; and in response to the current time reaching the speaking opportunity, adding a virtual main body in the group dialogue, and sending a reply message in the group dialogue through the virtual main body. Therefore, based on the dialogue message, the obtained speaking time can enable the virtual main body to join the group dialogue at a proper time, thereby ensuring the orderly joining of the virtual main body and further improving the experience of the dialogue main body in the group dialogue.

In some embodiments of the present disclosure, before the adding the virtual body to the group session and sending the reply message in the group session through the virtual body, the method further includes: sending timing query information to the group dialogue; and in response to receiving the confirmation information corresponding to the timing query information, adding a virtual main body in the group dialogue, and sending a reply message in the group dialogue through the virtual main body.

In this embodiment, the timing query information may be used as a message in a group session to inquire whether each session body in the group session can join the current virtual body. The timing query information may include: text, controls, images, etc., for example, the timing query information includes: question text, which is: "whether XX" can be added; for another example, the timing query information is: whether XX can be added, and a yes or no button. The confirmation information corresponding to the timing query information is confirmation information of the reply of the dialogue main body to the timing query information in the group dialogue, for example, the confirmation information is information after whether a button is selected in the timing query information.

In this embodiment, after determining the virtual main body and the reply information, in order to make the virtual main body participate in the group conversation, the opinion of each conversation main body in the group conversation may be queried by sending timing query information to the group conversation; after the confirmation information of the dialogue main body in the group dialogue is obtained, the virtual main body can be directly added into the group dialogue, and the virtual main body is introduced on the basis that the dialogue main body of the group dialogue has the knowledge, so that the experience of the dialogue main body in the group dialogue is improved.

Optionally, when sending the timing query information to the group session, it is also possible to receive rejection information corresponding to the timing query information, where the group session method further includes: and sending the timing query information to the group dialogue, and determining to directly push out without joining the virtual main body in response to receiving the rejection information of the corresponding timing query information.

In some optional implementations of this embodiment, the group dialogue method further includes: determining a designated role to be added into the group dialogue based on the reply information; determining specified setting information based on the specified role; acquiring feedback information of the designated role based on the designated setting information and the problem to be processed; adding a designated role in the group dialogue, and sending feedback information in the group dialogue through the designated role.

In this alternative implementation, based on the reply message of the virtual character, a history role or virtual character related to the reply message content may be determined, where the history role or virtual character is the designated role. The designated character is a conversation body standing at the same position or at a different position from the character of the virtual body, and after the designated character is added to the group conversation, the conversation body in the group conversation can determine the distinction between the virtual body and the designated character.

In this embodiment, the specified setting information is information describing the character content of the specified character, for example, if the specified character is a Caoca, the specified setting information is: the first generation of Xiaohong is a marshal in military and is created in literature, calligraphies and music.

In this embodiment, the feedback information is reply content obtained by analyzing the to-be-processed problem at the angle of the designated role, and the feedback information can be supplemented to knowledge of the invention that cannot be recognized by the virtual body with respect to the reply information.

In this embodiment, the obtaining feedback information of the designated role based on the designated setting information and the to-be-processed problem includes: outputting the appointed setting information and the problem to be processed to a main body large language model to obtain feedback information of an appointed role output by the main body large language model; the main body large language model is used for representing the corresponding relation between the appointed setting information and the to-be-processed problem as well as the feedback information, and specifically, the main body large language model can be a subsequent second large language model. In this embodiment, the subject large language model determines a specified role based on specified setting information; and standing at the angle of the designated role, analyzing the problem to be processed, and determining feedback information corresponding to the problem to be processed, wherein the feedback information is information for replying to another problem to be processed.

The group dialogue method provided by the embodiment of the disclosure determines a designated role to be added into a group dialogue based on reply information; determining specified setting information based on the specified role; acquiring feedback information of the designated role based on the designated setting information and the problem to be processed; adding a designated role in the group dialogue, and sending feedback information in the group dialogue through the designated role. After the virtual main body is added, the appointed role corresponding to the virtual main body is determined based on the reply information of the virtual main body, and the feedback information of the appointed role is obtained based on the appointed information of the appointed role, so that another main body with history synchronous relation with the virtual main body is added in the group dialogue, the appointed role publishes the feedback information related to the role setting of the appointed role in the group dialogue, the dialogue main body in the group dialogue can comprehensively refer to the reply information of the virtual main body and the feedback information of the appointed role, and the deep understanding of the dialogue main body on the dialogue content is improved.

Optionally, the group dialogue method further includes: and comprehensively evaluating the reply information and the feedback information based on the virtual main body and the designated role to obtain an evaluation result of the reply information and the feedback information, wherein the evaluation result comprises the reply information and the feedback information which are accurate, and the reply information and the feedback information are inaccurate, and the reply information or the feedback information is accurate.

In some embodiments of the present disclosure, the obtaining reply information of the virtual main body based on the virtual main body and the to-be-processed problem includes: determining character setting information based on the virtual body; and obtaining the reply information of the virtual main body based on the role setting information and the problem to be processed.

In this alternative implementation manner, the character setting information is information for describing character content of the history character corresponding to the virtual main body. After the character setting information is obtained, a virtual character corresponding to the historical character can be generated, and the virtual character is applied to the virtual main body, so that group conversations of people with various historical characteristics can be guided to be generated.

In this alternative implementation, the reply information may be reply content that replies to the to-be-processed question from the perspective of the historical role corresponding to the virtual main body.

The method for obtaining the reply information of the virtual main body provided by the embodiment of the disclosure determines role setting information based on the virtual main body; based on the character setting information and the problem to be processed, the reply information of the virtual main body is obtained, and the reply information is generated on the basis of the obtained character setting information, so that the reply information has the character characteristics corresponding to the character setting information, and the accuracy and the reliability of the reply information generation of the dialogue main body in the group dialogue are improved.

In some embodiments of the present disclosure, before the adding a virtual body to the group session and sending a reply message in the group session through the virtual body, the group session method further includes: determining the mood of the virtual main body based on the character setting information; adding a virtual body to the group conversation; and controlling the virtual main body to output the reply information in the group dialogue according to the language.

In this optional implementation manner, the mood of the virtual main body is the speaking mood of the history role corresponding to the role setting information in the history period, and each history role has a corresponding speaking mood due to the different status and social relationship of the history role.

The group dialogue method provided by the alternative implementation mode can determine the speaking language of the virtual main body according to the role setting of the virtual main body, so that the virtual main body is more vivid and interesting.

In the alternative implementation manner, the virtual main body is controlled to output the reply information in the group dialogue according to the language corresponding to the character setting information, so that the character characteristics of the virtual main body can be effectively reflected, and the dialogue main body in the group dialogue can have the experience of the dialogue with the history character. The present disclosure may be applied to various virtual character group chat systems, such as games, social software, educational software, and the like. The method and the device can enable the virtual body to be more lifelike and natural in the group conversation, so that experience of the conversation body in the group conversation is improved.

According to the group conversation method provided by the embodiment of the disclosure, conversation messages of conversation main bodies in group conversation are obtained; determining a virtual main body to be added into the group dialogue and a problem to be processed based on the dialogue message; determining character setting information based on the virtual body; obtaining reply information of the virtual main body based on the role setting information and the problem to be processed; determining the mood of the virtual main body based on the character setting information; adding a virtual main body into the group dialogue, and controlling the virtual main body to output reply information in the group dialogue according to the language.

According to the group conversation method provided by the embodiment of the disclosure, based on the role setting information, the semantics of the virtual main body are determined, after the virtual main body is added into the group conversation, the virtual main body is controlled to output the reply information in the group conversation according to the language, so that the reply information carries the historical character characteristics of the virtual main body, and the experience of the conversation main body in the group conversation is improved.

Optionally, before adding the virtual body in the group session and sending the reply message in the group session through the virtual body, the method further includes: determining a reply voice of the virtual main body based on the role setting information; adding a virtual main body into the group dialogue, and controlling the virtual main body to reply the tone of the voice and output reply information.

In some optional implementations of the disclosure, determining the role setting information based on the virtual body includes: acquiring a preset role setting relation table, wherein the role setting relation table is used for representing the corresponding relation between the table roles and the role setting information; matching the virtual main body with the table roles in the role setting relation table; and selecting the character setting information corresponding to the table character from the character setting relation table in response to the virtual main body being matched with the table character.

In this alternative implementation manner, the role setting relation table is a table for recording a correspondence between table roles and role setting information, where the table roles may be a role text obtained based on a group dialogue.

In this optional implementation manner, the executing body may set a table role for each dialog body in the group dialog in advance based on the gist of the group dialog, where the table role may be a history role that appears in history, or a role that is set in advance in a scenario. When the table character is a history character, the character setting information of the history character can be determined by crawling text and image information related to the history character from the Internet, and a corresponding relation between the history character and the character setting information of the history character is established, so that a character setting relation table is obtained.

Optionally, when the table character is a character on the script, since the script generally has description information of various characters, text or image content related to the character can be directly intercepted from the script, character setting information of the character is determined based on the intercepted content, and a corresponding relationship between the character and the character setting information of the character is obtained, so as to obtain the character setting relationship table.

In this alternative implementation manner, the role setting relation table may be stored in a data table in the large language model, when the large language model determines a problem to be processed, a virtual main body may be determined first, a table role is obtained by matching from the role setting relation table based on the similarity between the determined virtual main body and the table role, corresponding role setting information is obtained based on the matched table role, an avatar conforming to the description of the role setting information is generated through the role setting information, and the dialogue message is analyzed in the standpoint of the avatar to determine the problem to be processed that can be asked to the avatar.

According to the method for determining the role setting information provided by the alternative implementation mode, the role setting information corresponding to the current virtual main body is queried through the preset role setting relation table, so that the reliability of obtaining the role setting information is improved.

In some optional implementations of the disclosure, the above role setting relationship table is obtained by: determining historical text or characteristic text for describing different historical roles; inputting the historical text or the characteristic text into a first large language model to obtain character setting information output by the first large language model, wherein the first large language model is used for representing the corresponding relation between the historical text or the characteristic text and the character setting information; and storing the historical roles as table roles corresponding to the role setting information to obtain a role setting relation table.

In this alternative implementation, the different history roles are history characters that have actually appeared in history, and the history text is a text describing the history roles, for example, self-transmission documents and public documents of the history characters; the feature text may be a text obtained by describing the features of the history character, for example, the feature text of the history character XX is: taking XX as an original role for shooting a movie, please summarize the mode of XX speaking, the multi-Chinese movie requires: accurate and outstanding character. If you were to define a virtual character of "XX", you would generate a piece of character set information for this character of XX.

In the alternative implementation manner, the history text or the characteristic text is input into the first large language model, so that the character setting information output by the first large language model can be obtained. For example, after the above feature text is input into the first large language model, character setting information output from the first large language model may be obtained: XX, a young female born in ancient China, has remarkable beauty and intelligence. She receives good education from a small scale and is well-skilled in poetry gifts. The characteristics of softness, intelligence, toughness and delicateness are combined in the character of XX. Her talking about presents a unique elegance and wisdom while protecting himself and her loved ones ingeniously. XX always exhibits inflexible spirit and positive attitude towards the future when it encounters challenges and difficulties. Her story fills up the fluctuations in emotion, wisdom, and the complexity of humanization, making her a profound and memorable character.

According to the method for constructing the role setting relation table, the obtained historical text or the characteristic text is input into the first large language model, so that the role setting information output by the first large language model is obtained, and an accurate and reliable implementation mode is provided for obtaining the role setting relation table by means of the understanding capability of the large language model.

Optionally, the above role setting relationship table may be further obtained by: determining script roles aiming at a preset script; based on the script characters, selecting character description text related to the script characters from the script; inputting the character description text into the first large language model to obtain character setting information output by the first large language model; and taking the script characters as table characters, establishing a corresponding relation between the table characters and the character setting information, and obtaining a character setting relation table.

Optionally, the above-mentioned role setting relationship table may also be generated manually by an operator, specifically, the operator may receive role setting information set by the operator for different table roles, and establish a correspondence between the table roles and the role setting information to obtain the role setting relationship table, for example, the operator manually sets that the hole speaking should be said to say the text and the mory speaking should be filled with the philosophy.

In some optional implementations of the disclosure, the obtaining the reply information of the virtual main body based on the role setting information and the to-be-processed problem includes: inputting the character setting information and the problem to be processed into a second large language model to obtain reply information output by the second large language model, wherein the second large language model is used for representing the corresponding relation between the character setting information and the problem to be processed and the reply information.

In the alternative implementation manner, the second large language model is a large language model different from the first large language model, and the second large language model is a large language model for analyzing the problem to be processed on the basis of determining the role setting information to obtain the reply information; the first large language model is a large language model for inputting character text and obtaining character setting information.

According to the method for obtaining the reply information of the virtual main body, provided by the alternative implementation mode, the character setting information and the problem to be processed are input into the second large language model, so that the second large language model replies the reply information to the problem to be processed on the basis of the character described by the character setting information, and the reliability and accuracy of obtaining the reply information are improved.

In some optional implementations of the disclosure, determining the virtual body to be added to the group conversation and the problem to be processed based on the conversation message includes: determining context information for the group conversation based on the conversation message; determining a virtual body to be added to the group conversation based on the context information; based on the context information, a problem to be processed is determined.

In this alternative implementation, the dialogue message includes: the output message of each dialog body has a corresponding output timing and content for the output message of each dialog body. The above-mentioned context information includes: the timing encoding of each outgoing message, the semantic information of the outgoing message, and the relationship between adjacent outgoing messages. Optionally, the above context information may further include: output message sender, output message receiver, output message content, output message time.

In this embodiment, the dialog body output message may include: the "hyperlinked" message, the execution on which the group conversation method operates, by accessing the "hyperlinks" in the outgoing message, can access the linked content, and generate a summary or initiate a new topic from the web page content, etc.

In this optional implementation manner, determining, based on the context information, the virtual body to be added to the group session includes: extracting keywords of the entity from the output message, and determining the keywords of the entity; and responding to the keywords of the entity including the historic character names, and obtaining virtual subjects to be added into the group dialogue and corresponding to the historic characters.

Optionally, determining the virtual body to be added to the group session based on the context information includes: extracting main body characteristics of the output message, and determining main body characteristics corresponding to the output message; and responding to the matching of the main body characteristics and the preset characteristics of the historical main body, and obtaining the virtual main body which is to be added into the group dialogue and corresponds to the historical main body. The history subject may be a history character, a history action, a history character, or the like, for example, the history subject is a machine cat.

The determining the problem to be processed based on the context information includes: determining a topic of the conversation message based on the timing and content of the outgoing message; determining a question direction of the topic based on the relation between each adjacent output message in the context information; based on the subject and the problem direction, the problem to be processed is obtained. For example, the subject is a bamboo reclining chair, the problem direction is the manufacturing step of the reclining chair, and the problem to be treated is: and (3) manufacturing the bamboo deck chair, wherein each manufacturing step has different manufacturing results. The to-be-processed question is a question asking a bamboo artist, and the execution subject can generate a virtual subject related to the bamboo artist at this time to reply to the to-be-processed question.

The virtual main body and the problem to be processed in the group dialogue are determined, and the virtual main body and the problem to be processed are determined based on the context information of the dialogue message, so that the reliability and the accuracy of determining the virtual main body and the problem to be processed are improved.

With further reference to fig. 3, as an implementation of the method shown in the foregoing figures, the present disclosure provides an embodiment of a group dialogue device, which corresponds to the method embodiment shown in fig. 1, and which is particularly applicable in various electronic apparatuses.

As shown in fig. 3, the group dialogue device 300 provided in this embodiment includes: an acquisition unit 301, a determination unit 302, an acquisition unit 303, an issuing unit 304. The obtaining unit 301 may be configured to obtain a conversation message of a conversation body in a group conversation. The determining unit 302 may be configured to determine the virtual body to be added to the group conversation and the problem to be processed based on the conversation message. The obtaining unit 303 may be configured to obtain reply information of the virtual main body based on the virtual main body and the problem to be processed. The sending unit 304 may be configured to join a virtual body in the group session, and send a reply message in the group session through the virtual body.

In the present embodiment, in the group dialogue device 300: the specific processing of the obtaining unit 301, the determining unit 302, the obtaining unit 303, the sending unit 304 and the technical effects thereof may refer to the descriptions related to step 101, step 102, step 103 and step 104 in the corresponding embodiment of fig. 1, and are not described herein.

In some optional implementations of this embodiment, the apparatus further includes: a timing unit (not shown in the figure), a first joining unit (not shown in the figure), which may be configured to determine a speaking timing of the virtual body based on the dialog message. The first joining unit may be configured to join a virtual body in the group conversation based on the speaking opportunity, and send out a reply message in the group conversation through the virtual body.

In some optional implementations of this embodiment, the apparatus further includes: a query unit (not shown), a second joining unit (not shown). Wherein the above-mentioned query unit may be configured to send the timing query information to the group dialogue. The second joining unit may be configured to join the virtual body in the group conversation in response to receiving the acknowledgement information corresponding to the timing query information, and send the reply information in the group conversation through the virtual body.

In some optional implementations of this embodiment, the apparatus further includes: a specification unit (not shown in the figure), a setting unit (not shown in the figure), a feedback unit (not shown in the figure), a third joining unit (not shown in the figure). Wherein the specifying unit may be configured to determine, based on the reply information, a specified role to be added to the group conversation. The setting unit may be configured to determine the specified setting information based on the specified character. The feedback unit may be configured to obtain feedback information of the specified character based on the specified setting information and the problem to be processed. The third joining unit may be configured to join a specified role in the group conversation and send feedback information in the group conversation through the specified role.

In some optional implementations of this embodiment, the obtaining unit 303 may be configured to: determining character setting information based on the virtual body; and obtaining the reply information of the virtual main body based on the role setting information and the problem to be processed.

In some optional implementations of the disclosure, the apparatus 400 further includes: a mood unit (not shown), wherein the mood unit may be configured to: determining the mood of the virtual main body based on the character setting information; adding a virtual body to the group conversation; and controlling the virtual main body to output the reply information in the group dialogue according to the language.

In some optional implementations of the present disclosure, the obtaining unit 303 may be configured to: acquiring a preset role setting relation table, wherein the role setting relation table is used for representing the corresponding relation between the table roles and the role setting information; matching the virtual main body with the table roles in the role setting relation table; and selecting the character setting information corresponding to the table character from the character setting relation table in response to the virtual main body being matched with the table character.

In some optional implementations of the disclosure, the role setting relationship table may be determined by a relationship unit (not shown in the figure) that may be configured to: determining historical text or characteristic text for describing different historical roles; inputting the historical text or the characteristic text into a first large language model to obtain character setting information output by the first large language model, wherein the first large language model is used for representing the corresponding relation between the historical text or the characteristic text and the character setting information; and storing the historical roles as table roles corresponding to the role setting information to obtain a role setting relation table.

In some optional implementations of the present disclosure, the obtaining unit 303 may be configured to: inputting the character setting information and the problem to be processed into a second large language model to obtain reply information output by the second large language model, wherein the second large language model is used for representing the corresponding relation between the character setting information and the problem to be processed and the reply information.

In some optional implementations of the present disclosure, the determining unit 302 may be configured to: determining context information for the group conversation based on the conversation message; determining a virtual body to be added to the group conversation based on the context information; based on the context information, a problem to be processed is determined.

The group conversation device provided in the embodiment of the present disclosure, first, the acquisition unit 301 acquires a conversation message of a conversation body in a group conversation; next, the determining unit 302 determines a virtual body to be added to the group conversation and a problem to be processed based on the conversation message; again, the obtaining unit 303 obtains reply information of the virtual main body based on the virtual main body and the problem to be processed; finally, the issuing unit 304 joins the virtual body in the group conversation, and issues a reply message in the group conversation through the virtual body. Thus, based on the dialogue information of the dialogue main body in the group dialogue, the virtual main body is determined, and the obtained virtual main body can be matched with the dialogue information in the group dialogue; based on the dialogue message, the problem to be processed is determined, so that the center thought of the dialogue message and the problem to be solved can be effectively analyzed, and the reliability of obtaining the virtual main body reply information is improved; the virtual main body is added in the group dialogue, and the virtual main body is controlled to send out the reply information, so that the effectiveness and the reliability of dialogue contents in the group dialogue can be ensured, and the dialogue experience of the dialogue main body in the group dialogue is improved.

In the technical scheme of the disclosure, the related processes of collecting, storing, using, processing, transmitting, providing, disclosing and the like of the personal information of the user accord with the regulations of related laws and regulations, and the public order colloquial is not violated.

According to embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

Fig. 4 illustrates a schematic block diagram of an example electronic device 400 that may be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 4, the apparatus 400 includes a computing unit 401 that can perform various suitable actions and processes according to a computer program stored in a Read Only Memory (ROM) 402 or a computer program loaded from a storage unit 408 into a Random Access Memory (RAM) 403. In RAM403, various programs and data required for the operation of device 400 may also be stored. The computing unit 401, ROM 402, and RAM403 are connected to each other by a bus 404. An input/output (I/O) interface 405 is also connected to bus 404.

Various components in device 400 are connected to I/O interface 405, including: an input unit 406 such as a keyboard, a mouse, etc.; an output unit 407 such as various types of displays, speakers, and the like; a storage unit 408, such as a magnetic disk, optical disk, etc.; and a communication unit 409 such as a network card, modem, wireless communication transceiver, etc. The communication unit 409 allows the device 400 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.

The computing unit 401 may be a variety of general purpose and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 401 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 401 performs the various methods and processes described above, such as a group conversation method. For example, in some embodiments, the group conversation method may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as the storage unit 408. In some embodiments, part or all of the computer program may be loaded and/or installed onto the device 400 via the ROM402 and/or the communication unit 409. When the computer program is loaded into RAM 403 and executed by computing unit 401, one or more steps of the group dialogue method described above may be performed. Alternatively, in other embodiments, the computing unit 401 may be configured to perform the group dialogue method by any other suitable means (e.g. by means of firmware).

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable group dialogue device such that the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), and the internet.

The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.

It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel, sequentially, or in a different order, provided that the desired results of the disclosed aspects are achieved, and are not limited herein.

The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.

Claims (23)

1. A group dialogue method, the method comprising:

acquiring dialogue information of a dialogue main body in a group dialogue;

determining a virtual body to be added into the group dialogue and a problem to be processed based on the dialogue message;

obtaining reply information of the virtual main body based on the virtual main body and the to-be-processed problem;

and adding the virtual main body in the group dialogue, and sending out the reply information in the group dialogue through the virtual main body.

2. The method of claim 1, the joining the virtual body in the group conversation and prior to issuing the reply message in the group conversation by the virtual body, the method further comprising:

determining a speaking opportunity of the virtual subject based on the dialogue message;

And adding the virtual main body into the group dialogue based on the speaking time, and sending out the reply information in the group dialogue through the virtual main body.

3. The method of claim 1, the joining the virtual body in the group conversation and prior to issuing the reply message in the group conversation by the virtual body, the method further comprising:

sending timing query information to the group session;

and responding to the received confirmation information corresponding to the timing query information, adding the virtual main body in the group dialogue, and sending out the reply information in the group dialogue through the virtual main body.

4. A method according to one of claims 1-3, the method further comprising:

determining a designated role to be added into the group dialogue based on the reply information;

determining specified setting information based on the specified role;

based on the appointed setting information and the to-be-processed problem, obtaining feedback information of the appointed role;

and adding the designated role in the group dialogue, and sending the feedback information in the group dialogue through the designated role.

5. The method of claim 1, wherein the obtaining reply information of the virtual body based on the virtual body and the to-be-processed problem comprises:

determining character setting information based on the virtual body;

and obtaining the reply information of the virtual main body based on the role setting information and the to-be-processed problem.

6. The method of claim 5, the joining the virtual body in the group conversation and prior to issuing the reply message in the group conversation by the virtual body, the method further comprising:

determining the mood of the virtual main body based on the character setting information;

adding the virtual body to the group conversation;

and controlling the virtual main body to output the reply information in the group dialogue according to the language.

7. The method of claim 5, wherein the determining character setting information based on the virtual body comprises:

acquiring a preset role setting relation table, wherein the role setting relation table is used for representing the corresponding relation between the table roles and the role setting information;

matching the virtual main body with table roles in the role setting relation table;

And selecting character setting information corresponding to the table character from the character setting relation table in response to the virtual main body being matched with the table character.

8. The method of claim 7, wherein the role setting relationship table is obtained by:

determining historical text or characteristic text for describing different historical roles;

inputting the historical text or the characteristic text into a first large language model to obtain character setting information output by the first large language model, wherein the first large language model is used for representing the corresponding relation between the historical text or the characteristic text and the character setting information;

and storing the historical roles as table roles corresponding to the role setting information to obtain a role setting relation table.

9. The method of claim 5, wherein the obtaining reply information of the virtual body based on the character setting information and the to-be-processed question comprises:

inputting the character setting information and the problem to be processed into a second large language model to obtain reply information output by the second large language model, wherein the second large language model is used for representing the corresponding relation between the character setting information and the problem to be processed and the reply information.

10. The method of claim 1, wherein the determining, based on the conversation message, a virtual body to join in the group conversation and a problem to process comprises:

determining context information for the group conversation based on the conversation message;

determining a virtual body to be added to the group conversation based on the context information;

and determining a problem to be processed based on the context information.

11. A group dialogue device, the device comprising:

an acquisition unit configured to acquire a conversation message of a conversation body in a group conversation;

a determining unit configured to determine a virtual body to be added to the group conversation and a problem to be processed based on the conversation message;

the obtaining unit is configured to obtain reply information of the virtual main body based on the virtual main body and the to-be-processed problem;

and the sending unit is configured to join the virtual main body in the group dialogue and send the reply information in the group dialogue through the virtual main body.

12. The apparatus of claim 11, the apparatus further comprising:

a timing unit configured to determine a speaking opportunity of the virtual body based on the dialog message;

And a first joining unit configured to join the virtual body in the group conversation based on the speaking opportunity, and to issue the reply message in the group conversation through the virtual body.

13. The apparatus of claim 11, the apparatus further comprising:

a query unit configured to send timing query information to the group conversation;

and a second joining unit configured to join the virtual body in the group dialogue in response to receiving acknowledgement information corresponding to the timing query information, and to issue the reply information in the group dialogue through the virtual body.

14. The apparatus according to one of claims 11-13, the apparatus further comprising:

a specifying unit configured to determine a specified character to be added to the group conversation based on the reply information;

a setting unit configured to determine specified setting information based on the specified character;

a feedback unit configured to obtain feedback information of the specified character based on the specified setting information and the question to be processed;

and a third joining unit configured to join the specified role in the group conversation and to issue the feedback information in the group conversation through the specified role.

15. The apparatus of claim 11, wherein the deriving unit is configured to: determining character setting information based on the virtual body; and obtaining the reply information of the virtual main body based on the role setting information and the to-be-processed problem.

16. The apparatus of claim 15, the apparatus further comprising: a mood unit configured to: determining the mood of the virtual main body based on the character setting information; adding the virtual body to the group conversation; and controlling the virtual main body to output the reply information in the group dialogue according to the language.

17. The apparatus of claim 15, wherein the deriving unit is configured to: acquiring a preset role setting relation table, wherein the role setting relation table is used for representing the corresponding relation between the table roles and the role setting information; matching the virtual main body with table roles in the role setting relation table; and selecting character setting information corresponding to the table character from the character setting relation table in response to the virtual main body being matched with the table character.

18. The apparatus of claim 17, wherein the role setting relationship table is determined by a relationship unit configured to: determining historical text or characteristic text for describing different historical roles; inputting the historical text or the characteristic text into a first large language model to obtain character setting information output by the first large language model, wherein the first large language model is used for representing the corresponding relation between the historical text or the characteristic text and the character setting information; and storing the historical roles as table roles corresponding to the role setting information to obtain a role setting relation table.

19. The apparatus of claim 15, wherein the deriving unit is configured to: inputting the character setting information and the problem to be processed into a second large language model to obtain reply information output by the second large language model, wherein the second large language model is used for representing the corresponding relation between the character setting information and the problem to be processed and the reply information.

20. The apparatus of claim 11, wherein the determining unit is configured to: determining context information for the group conversation based on the conversation message; determining a virtual body to be added to the group conversation based on the context information; and determining a problem to be processed based on the context information.

21. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-10.

22. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-10.

23. A computer program product comprising a computer program which, when executed by a processor, implements the method of any of claims 1-10.