CN117251537B - Session processing method, device, equipment and storage medium for artificial intelligence - Google Patents

Abstract

The embodiments of the present application provide a conversation processing method, device, equipment and storage medium for artificial intelligence. In the embodiments of the present application, an artificial intelligence aggregation application that aggregates multiple AI models is provided, and a user conducts a conversation with a first AI model on a first conversation interface provided by the application, generates a first conversation content, converts the first conversation content into a second conversation content adapted to a second AI model, and displays it on a second conversation interface provided by the second AI model, so that the user can continue to have a conversation with the second AI model based on the second conversation content, and realizes free switching between multiple AI models for the same conversation. Furthermore, since the user's conversation is continuous and will not be interrupted due to the change of AI models, the user experience is good.

Description

Conversation processing method, device, equipment and storage medium for artificial intelligence

Technical Field

The present application relates to the field of artificial intelligence technology, and in particular to a conversation processing method, device, equipment and storage medium for artificial intelligence.

Background Art

With the development of artificial intelligence (AI) technology, some artificial intelligence models (AI models for short) have also emerged. These AI models can learn and understand human language, allowing them to chat and communicate with users like humans, and even complete tasks such as copywriting, translation, code writing, and image drawing according to human needs.

When using an AI model, users can start a conversation in the AI model and communicate with the AI model about a problem in the conversation. However, currently, the same conversation can only appear in the same AI model, and users cannot switch freely between multiple AI models for the same conversation, which affects the user experience.

Summary of the invention

Multiple aspects of the present application provide a conversation processing method, apparatus, device and storage medium for artificial intelligence, which are used to achieve free switching between multiple AI models for the same conversation.

An embodiment of the present application provides a conversation processing method for artificial intelligence, which is applicable to artificial intelligence aggregation applications, in which multiple AI models are aggregated. The method includes: when a first user needs to have a conversation with the first AI model, running the first AI model to display a first conversation interface of the first AI model, and in response to a dialogue operation between the first user and the first AI model, displaying a first conversation content between the first user and the first AI model on the first conversation interface; during the conversation between the first user and the first AI model, in response to a trigger operation of switching from the first AI model to the second AI model, using a conversion middleware to convert the format of the first conversation content to obtain a second conversation content supported by the second AI model, and providing the second conversation content to the second AI model; running the second AI model to switch the first conversation interface to a second conversation interface corresponding to the second AI model, and displaying the second conversation content on the second conversation interface, so that the first user can continue to have a conversation with the second AI model based on the second conversation content.

The embodiment of the present application also provides a conversation processing device for artificial intelligence, an artificial intelligence aggregation application running on the device, and multiple AI models are aggregated in the application. The device includes: an operation module, a display module, a conversion middleware module, a providing module, and a switching module; the operation module is used to run the first AI model when the first user needs to have a conversation with the first AI model, the display module is used to display the first conversation interface of the first AI model, and the display module is also used to display the first conversation content between the first user and the first AI model on the first conversation interface in response to the dialogue operation between the first user and the first AI model; the conversion middleware module is used to convert the format of the first conversation content in response to the trigger operation of switching from the first AI model to the second AI model during the conversation between the first user and the first AI model to obtain the second conversation content supported by the second AI model, and the providing module is used to provide the second conversation content to the second AI model; the operation module is used to run the second AI model, the switching module is used to switch the first conversation interface to the second conversation interface corresponding to the second AI model, and the display module is also used to display the second conversation content on the second conversation interface, so that the first user can continue to have a conversation with the second AI model based on the second conversation content.

An embodiment of the present application also provides a conversation processing device for artificial intelligence, including: a memory and a processor; the memory is used to store a computer program; the processor is coupled to the memory and is used to execute the computer program to implement each step in the conversation processing device for artificial intelligence provided in the embodiment of the present application.

The embodiment of the present application also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor implements each step in the conversation processing device for artificial intelligence provided by the embodiment of the present application.

In an embodiment of the present application, an artificial intelligence aggregation application that aggregates multiple AI models is provided. A user conducts a conversation with a first AI model on a first conversation interface provided by the application, generates first conversation content, converts the first conversation content into second conversation content adapted to a second AI model, and displays it on a second conversation interface provided by the second AI model, so that the user can continue to have a conversation with the second AI model based on the second conversation content, thereby achieving free switching between multiple AI models for the same conversation. Furthermore, since the user's session is continuous and will not be interrupted due to the change of AI models, the user experience is good.

BRIEF DESCRIPTION OF THE DRAWINGS

The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

FIG1a is a flow chart of a conversation processing method for artificial intelligence provided by an exemplary embodiment of the present application;

FIG1b is a schematic diagram of a first conversation interface provided by an exemplary embodiment of the present application;

FIG2 is a schematic diagram of the structure of a conversation processing device for artificial intelligence provided by an exemplary embodiment of the present application;

FIG3 is a schematic diagram of the structure of a conversation processing device for artificial intelligence provided by an exemplary embodiment of the present application.

DETAILED DESCRIPTION

In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

In response to the above-mentioned technical problem that users cannot freely switch between multiple AI models for the same conversation, in an embodiment of the present application, an artificial intelligence aggregation application that aggregates multiple AI models is provided. The user has a conversation with a first AI model on a first conversation interface provided by the application, generates first conversation content, converts the first conversation content into second conversation content adapted to the second AI model, and displays it on a second conversation interface provided by the second AI model, so that the user can continue to have a conversation with the second AI model based on the second conversation content, thereby achieving free switching between multiple AI models for the same conversation. Furthermore, since the user's session is continuous and will not be interrupted due to the change of AI models, the user experience is good.

A solution provided by an embodiment of the present application is described in detail below in conjunction with the accompanying drawings.

FIG. 1a is a flow chart of a method for artificial intelligence session processing provided by an exemplary embodiment of the present application. The method is applicable to an artificial intelligence aggregation application in which multiple AI models are aggregated. As shown in FIG. 1a , the method includes:

101. When a first user needs to have a conversation with a first AI model, run the first AI model to display a first conversation interface of the first AI model, and in response to a conversation operation between the first user and the first AI model, display a first conversation content between the first user and the first AI model on the first conversation interface;

102. During a conversation between the first user and the first AI model, in response to a trigger operation of switching from the first AI model to the second AI model, the first conversation content is formatted using the conversion middleware to obtain second conversation content supported by the second AI model, and the second conversation content is provided to the second AI model;

103. Run the second AI model to switch the first conversation interface to a second conversation interface corresponding to the second AI model, and display the second conversation content on the second conversation interface so that the first user can continue the conversation with the second AI model based on the second conversation content.

In this embodiment, the multiple AI models aggregated in the artificial intelligence aggregation application are not limited. For example, multiple AI models can be models of different types. For example, from the perspective of the technology used by the AI model, the types of AI models include but are not limited to: supervised learning models, semi-supervised learning models, unsupervised learning models, reinforcement learning models, convolutional neural networks (CNN) models, recurrent neural networks (RNN) models, generative models or anomaly detection models, etc. For another example, multiple AI models can be implemented as AI models of the same type but belonging to different manufacturers. From the perspective of the functions implemented by the AI model, the AI model includes but is not limited to: natural language understanding models, image or video understanding models, chat models, AI programming models, intelligent customer service models, AI office models, AI search engines, and multimodal image generation models, etc.

In this embodiment, the artificial intelligence aggregation application is used to aggregate multiple AI models, and the application contains program codes corresponding to multiple AIs, or contains download addresses of program codes corresponding to multiple AI models, so that the program code of the required AI model can be loaded through the download address during the operation of the artificial intelligence aggregation application. In addition, the artificial intelligence aggregation application has an interactive function with the user, and can respond to the operation or natural language instruction of the user selecting the AI model mentioned in the following embodiment to select the AI model for the user. The interactive function is implemented as an interactive interface and a first functional module that responds to and processes the operation or instruction on the interactive interface. In addition, the artificial intelligence aggregation application also has an interactive function for interacting with each AI model, which is implemented as an interactive interface (such as an API interface) with each AI model and a second functional module for data transmission and processing through the interactive interface, so as to send a scheduling request to the AI model, transmit the session information input by the user on the interactive interface to the AI model for the AI model to respond and obtain the response information of the AI model for the session information feedback (the response information is also a session information), and display the response information on the interactive interface through the first functional module to realize the session interaction between the called AI model and the user. Of course, the first functional module and the second functional module can also be integrated into the same functional module, which is not limited to this. The relevant description of the session information processing can be found in the subsequent embodiments, which will not be described in detail here.

Optionally, the AI aggregation application will provide an AI model selection portal for users to select the AI model they want to use from multiple AI models. For example, the AI aggregation application can provide a home page that includes a list of AI models, which can be a drop-down list or a flat list. In addition to the home page, the session page generated by running a certain AI model can also include an AI model selection portal, so that users can switch from one AI model to another at any time.

Optionally, the artificial intelligence aggregation application can also adaptively match the appropriate AI model for the user according to the natural language instructions input by the user. Specifically, the artificial intelligence aggregation application aggregates multiple AI models, and adopts an independent task scheduling model to provide the natural language instructions input by the user to the independent task scheduling model. The model receives the user's natural language instructions, estimates the user's motivation for the natural language instructions, and outputs the AI model that needs to be called, so as to realize the free switching of multiple AI models according to the natural language instructions input by the user. Among them, the independent task scheduling model can estimate the user's motivation for the natural language instructions for each AI model, and determine the AI model that is adapted to the natural language instructions input by the user. For example, the service quality of the AI model processing the natural language instructions input by the user can be evaluated. The service quality can be measured by the time of processing the natural language instructions or the attention of the output content, etc. The AI model with the best service quality is used as the AI model adapted to the natural language instructions input by the user, that is, the AI model that needs to be called. Of course, the selection operation initiated by the user through the AI model selection entrance can also be converted into a special case of the natural language instructions input by the user after being converted into a selection instruction, and then the independent task scheduling model schedules the adapted AI model for the user.

Among them, the artificial intelligence aggregation application can run on the server-side device, and the server-side device provides a web page to the client of the artificial intelligence aggregation application, and the user conducts a conversation with the AI model based on the web page. In addition, when the computing power of the terminal device is strong enough, for example, the processor of the terminal device is implemented with a dedicated chip, the artificial intelligence aggregation application can also be implemented as: an application (app) or a small program running locally on the terminal device. Among them, no matter which deployment implementation method the artificial intelligence aggregation application adopts, the multiple AI models aggregated in the artificial intelligence aggregation application can be deployed on the same device or distributed on multiple devices, without limitation. Regardless of whether multiple AI models are deployed on the same device or on multiple different devices, each AI model can be accessed and run through the artificial intelligence aggregation application.

In this embodiment, the first user can select the first AI model and have a conversation with the first AI model to achieve chatting or communication with the first AI model. In the case where the first user needs to have a conversation with the first AI model, if the artificial intelligence aggregation application runs on the terminal device, the artificial intelligence aggregation application can load the program code of the first AI model from the terminal device locally or remotely, and then run the first AI model; if the artificial intelligence aggregation application runs on the server device, the artificial intelligence aggregation application can load the program code of the first AI model from the server device locally or on other devices, and then run the first AI model. In the process of running the first AI model, the first conversation interface of the first AI model is displayed, and the first user can have a conversation operation with the first AI model on the first conversation interface. Accordingly, the artificial intelligence aggregation application can respond to the conversation operation between the first user and the first AI model and display the first conversation content between the first user and the first AI model on the first conversation interface. For example, the first user inputs text, pictures, videos or files on the first session page, and the first AI model outputs text, pictures, videos or files. The session content refers to the text, pictures, videos or files input by the first user, and the text, pictures, videos or files output by the AI model. The same session content has globally unique identification information, such as the identity document (ID) of the first session content.

During a conversation between a first user and a first AI model, the first user wishes to continue the conversation using a second AI model. The first user initiates a trigger operation to switch from the first AI model to the second AI model. Accordingly, the artificial intelligence aggregation application responds to the trigger operation and uses the conversion middleware to convert the format of the first conversation content to obtain the second conversation content supported by the second AI model.

Among them, there are differences in the input parameter format or output parameter format of different AI models, and the formats of different conversation contents are also different. For example, the conversation content input into the first AI model needs to meet the input parameter format of the first AI model, and the conversation content output from the first AI model has the output parameter format of the first AI model; the conversation content input into the second AI model needs to meet the input parameter format of the second AI model, and the conversation content output from the second AI model has the output parameter format of the second AI model. Based on this, a conversion middleware (converter) is provided in an embodiment of the present application. The conversion middleware is an independent system software service system that can provide format conversion services for each AI model, mainly used for format conversion of conversation content to adapt to the corresponding AI model. Specifically, the data input by the first user in the first conversation content conforms to the input parameter format of the first AI model, and the data output by the first AI model conforms to the output parameter format of the first AI model. The first conversation content can be converted into the input parameter format supported by the second AI model through the conversion middleware, so as to facilitate the subsequent input of the data as the input parameter of the second AI model into the second AI model, so as to conduct subsequent conversations based on the second AI model.

Optionally, the first session content is saved in a universal format supported by the aggregation application to obtain the first session content in a universal format. Further optionally, the first session content and its source, status and identification information are encapsulated and stored in a database, and the encapsulated first session content is saved in the database based on the identification information of the first session content. When the content in the first session needs to be used, the encapsulated first session content is obtained from the database based on the identification information, and is unpacked and converted into the first session content in a universal format. Among them, different AI models have different input parameter formats, and the input parameter format at least includes: context length information (context length) and input type information. The context length information is a parameter used to describe how many tokens (tokens) can be used by the AI model in a single request. For example, the context length can be 4k, 8k or 32k, etc., and the input type information can be pictures, text or videos, etc. When it is necessary to use the second AI model to continue the session, the input parameter format supported by the second AI model is used to convert the first session content in the universal format into the second session content adapted to the input parameter format. In this embodiment, the second conversation content can be provided to the second AI model. If the artificial intelligence aggregation application runs on the terminal device, the artificial intelligence aggregation application can load the program code of the second AI model from the terminal device locally or remotely, and then run the second AI model. Alternatively, if the artificial intelligence aggregation application runs on the server device, the artificial intelligence aggregation application can load the program code of the second AI model from the server device locally or on other devices, and then run the second AI model. When the second AI model is running, the first conversation interface is switched to the second conversation interface corresponding to the second AI model, and the second conversation content is displayed on the second conversation interface, so that the first user can continue the conversation with the second AI model based on the second conversation content. For example, when the second conversation content is provided to the second AI model, the monitoring module of the artificial intelligence aggregation application monitors that the second conversation content is sent, and then triggers the operation of the second AI model, and switches the currently displayed first conversation interface to the second conversation interface. Optionally, the display styles of the first conversation interface and the second conversation interface can be consistent, for example, the background color, text style, layout style, etc. are consistent, so that the first conversation interface is smoothly switched to the second conversation interface without the user's perception.

As shown in Figure 1b, the present application exemplarily provides a schematic diagram of a first conversation interface, and the schematic diagram of the second conversation interface can be the same or similar to the first conversation interface, without limitation. Among them, the first conversation interface includes the question information of the first user and the answer information of the first AI model, and the question information and the answer information constitute the conversation content between the first user and the first AI model. The question information can be text, pictures, videos, files, etc., and the answer information can be text, pictures, videos, files, etc.

In an embodiment of the present application, an artificial intelligence aggregation application that aggregates multiple AI models is provided. A user conducts a conversation with a first AI model on a first conversation interface provided by the application, generates first conversation content, converts the first conversation content into second conversation content adapted to a second AI model, and displays it on a second conversation interface provided by the second AI model, so that the user can continue to have a conversation with the second AI model based on the second conversation content, thereby achieving free switching between multiple AI models for the same conversation. Furthermore, since the user's session is continuous and will not be interrupted due to the change of AI models, the user experience is good.

In an optional embodiment, an AI model list is displayed on the first conversation interface in a pop-up window, floating layer, or drop-down list, and the AI model list includes at least one AI model, and the at least one AI model includes at least: a first AI model and a second AI model; in response to the selection operation of the second AI model, it is determined that the user needs to switch from the first AI model to the second AI model, and then the conversion middleware is used to convert the format of the first conversation content to obtain the second conversation content supported by the second AI model. Among them, FIG1b exemplarily shows an AI model list, and the AI model list includes the first AI model, the second AI model to the Nth AI model, where N is a positive integer and N is greater than or equal to 3.

The triggering operation for switching from the first AI model to the second AI model is not limited. For example, the first conversation interface includes an AI model selection control, which is implemented as a "switch control" in FIG. 1b, and the trigger operation of switching from the first AI model to the second AI model is implemented as a click operation on the AI model selection control; for another example, the artificial intelligence aggregation application supports switching of AI models through gesture operations, and pre-maintains gesture operations for switching between different AI models, wherein the gesture operations for switching between different AI models are not exactly the same, so as to distinguish the switching between different AI models, based on this, the trigger operation of switching from the first AI model to the second AI model is implemented as a gesture operation performed on the first conversation interface, such as an S-shaped gesture operation, a Z-shaped gesture operation, or a W-shaped gesture operation, etc.; for another example, the artificial intelligence aggregation application has a voice recognition function (for example, includes an AI model with voice recognition) and supports switching of AI models by voice, and maintains voice instructions for switching AI models, based on this, the trigger operation of switching from the first AI model to the second AI model is implemented as a voice instruction for switching AI models received on the first conversation interface, and the voice instruction contains identification information of the second AI model to be switched to, such as the name or system ID of the second AI model.

Optionally, in some embodiments, the AI model includes two states: unlocked and unlocked; an AI model in an unlocked state indicates that the user does not have the permission to use it temporarily, and the user needs to unlock the AI model before using the AI model in an unlocked state. In addition to displaying the identification information of the AI model, the unlocking state of the AI model, such as unlocked or unlocked, is also displayed in the AI model list, so that the user can understand the unlocking state of each AI model and better select the AI model to switch. In an optional embodiment, the AI model list includes at least two AI models, and the at least two AI models include an unlocked AI model and an unlocked AI model, wherein the unlocked AI model is an AI model that the user currently has the permission to use, and the unlocked AI model is an AI model that the user currently does not have the permission to use. Accordingly, when the AI model list is displayed on the first session interface in the form of a pop-up window, a floating layer or a drop-down list, the unlocking status of each AI model can be displayed in the AI model list. For example, the unlocking status of an unlocked AI model can be represented by an "open lock", and the unlocking status of an unlocked AI model can be represented by a "closed lock"; for another example, the unlocking status of an unlocked AI model can be represented by a word such as "free" or "paid" or "unlocked", and the unlocking status of an unlocked AI model can be represented by a word such as "payment required" or "need to unlock".

After the user chooses to switch to the second AI model and before the conversion middleware is used to convert the format of the first session content, it can be determined whether the second AI model is an unlocked AI model; if the second AI model is an unlocked AI model, an unlocking interface is displayed, and the unlocking interface includes: unlocking task description information, and the unlocking task description information is used to assist the user to perform the unlocking task for the AI model. For example, the unlocking task description information is implemented as a payment address. When the payment address is triggered, a payment page is displayed, and the first user can pay for the second AI model on the payment page to complete the unlocking task; for another example, the unlocking page maintains preset verification information, and the unlocking task description information indicates the verification information input on the unlocking interface, so that the input verification information can be obtained in response to the input operation on the unlocking interface, and it is determined whether the verification information is consistent with the preset verification information. If it is consistent, it is considered that the unlocking task is completed; for another example, the unlocking task description information is a prompt information indicating watching an advertisement on the unlocking interface, and the unlocking page also includes a playback control for triggering the playback of the advertisement. In response to the triggering operation on the playback control, the advertisement information is played, and after the advertisement is played, it is considered that the unlocking task is completed.

When the unlocking task is completed, the second AI model is unlocked so that the first user obtains the right to use the second AI model. On this basis, the first conversation content is further converted using the conversion middleware to obtain the second conversation content supported by the second AI model, and the second conversation content is provided to the second AI model so that the first user can continue the conversation with the second AI model based on the second conversation content. Specifically, there is an aggregation interface between the artificial intelligence aggregation application and each AI model. The second conversation content is provided to the second AI model through the aggregation interface corresponding to the second AI model. The second AI model uses the second conversation content as part of the AI model. When the second AI model is running, the second conversation interface is presented, and the second conversation content is presented on the second conversation interface so that the user can continue the conversation based on the second conversation content.

For example, if a user starts a session on the intelligent chat model to communicate with the intelligent chat model, and then starts a session on the intelligent question-and-answer model, this solution allows the user to continue the previous session on the intelligent chat model on the intelligent question-and-answer model. In this way, the user's session is continuous and will not be interrupted due to the change of AI models. The user experience is good, which is conducive to solving all user needs, helping users switch between different AI models, and meeting users' complex application needs.

In an optional embodiment, a user can perform a session publishing operation on a session interface provided by any AI model to publish the session content on the session interface to other users, so that other users can understand the user's session content. Further, optionally, it can also be determined whether to continue the session based on the session content to achieve the purpose of communicating between different users in the same session. For the convenience of description and distinction, the user who conducts a session on the first session interface and the second session interface mentioned in this embodiment is referred to as the first user. Specifically, the first user can perform a session publishing operation on the second session interface to publish the second session content and the new session content generated on the second session interface to other users of the aggregated application. Among them, the implementation method of the session publishing operation is not limited. For example, the second session interface includes a session publishing control, and the session publishing operation is implemented as a triggering operation of the session publishing control, wherein the publishing control is shown in Figure 1b. Of course, the first user can perform a session publishing operation on the first session interface to publish the first session content to other users of the aggregated application. The following is an example of the first user performing a session publishing operation on the second session interface.

Specifically, the aggregate application responds to the first user's session publishing operation on the second session interface, generates the first shared session content of the first user according to the second session content and the new session content generated on the second session interface; for example, the second session content and the new session content generated on the second session interface are encoded and packaged to obtain the first shared session content; and the first shared session content is published to other users of the aggregate application.

Among them, the first shared session content is published to other users of the aggregation application, and other users can view the first shared session content; optionally, the conversation can also be continued based on the first shared session content to achieve the purpose of communication between different users in the same conversation. Specifically, the first shared session content is published to other users of the aggregation application, so that other users can continue to have a conversation with the third AI model based on the first shared session content on the third conversation interface of the third AI model. The third AI model is the AI model used by other users to have a conversation using the artificial intelligence aggregation application. Among them, the artificial intelligence aggregation application used by other users can be run on the server device, and the third AI model is run on the server device. The server device provides the third conversation interface corresponding to the third AI model to the client of other users, and the conversation with the third AI model is continued on the third conversation interface based on the first shared session content. The terminal device used by other users uses a dedicated chip with strong computing power. The artificial intelligence aggregation application and the third AI model can be run on the terminal device, and the third conversation interface can be displayed. On the third conversation interface, the conversation with the third AI model can be continued based on the first shared session content.

The implementation method of publishing the first shared session content to other users of the aggregated application is not limited. In an optional embodiment, publishing the first shared session content to other users of the aggregated application is implemented by publishing the first shared session content to a shared page provided by the aggregated application, and the shared session content on the shared page is visible to a specific user of the aggregated application. The specific user may be some or all users of the aggregated application. Further, the specific user is implemented as a user with a certain user level, for example, a feature user is implemented as a user with an account level exceeding level 5; or the specific user is implemented as a user with membership. Further, the shared page can be provided to other platforms corresponding to the aggregated application, for example, a social platform. In another optional embodiment, in order to maintain information security and protect user privacy, the first shared session content can be shared in a point-to-point or end-to-end manner, for example, the first shared session content can be shared through a private message sharing method, a text message sharing method, or an email sharing method on a social platform.

Further optionally, the first user can publish or share his own session content with other users, and can also receive session content published or shared by other users, and continue the session based on the session content published or shared by other users. Based on this, the first user can select the second shared session content that needs to be continued from the sharing page, so as to continue the session based on the second shared session content. Accordingly, the artificial intelligence aggregation application can respond to the first user's selection operation on the second shared session content published by the second user on the sharing page, and determine whether the fourth AI model currently used by the first user and the AI model corresponding to the second shared session content are the same. The AI model corresponding to the second shared session content refers to the AI model used when generating the second shared session content; if the fourth AI model currently used by the first user and the AI model corresponding to the second shared session content are not the same, the second shared session content is converted into a format using the conversion middleware to obtain the third session content supported by the fourth AI model; wherein, the detailed implementation method of format conversion using the conversion middleware can be referred to above, and will not be repeated here; the third session content is displayed on the fourth session interface of the fourth AI model, so that the first user can continue the session with the fourth AI model based on the third session content. In an optional embodiment, the fourth AI model can be implemented as the first AI model, the third conversation content can be implemented as the first conversation content, and the process of the first user continuing the conversation with the fourth AI model based on the third conversation content can be implemented according to the method shown in Figure 1a.

Further optionally, if the fourth AI model currently used by the first user is the same as the AI model corresponding to the second shared session content, the second shared session content is directly displayed on the fourth session interface of the fourth AI model, so that the first user can continue the conversation with the fourth AI model based on the second shared session content. It should be noted that the fourth AI model can be implemented as the first AI model, the second shared session content can be implemented as the first session content, and the process of the first user continuing the conversation with the fourth AI model based on the second shared session content can be implemented according to the method shown in Figure 1a.

For example, user x1 has a conversation with AI model y1 to generate conversation content z1; user x2 is interested in conversation content z1, and user x2 is accustomed to using AI model y2, so the conversation content z1 is converted into conversation content z1’ supported by AI model y2. Based on conversation content z1’, user x2 has a conversation with AI model y2 to generate conversation content z2; user x3 is interested in conversation content z2, and user x3 is accustomed to using AI model y3, so the conversation content z2 is converted into conversation content z2’ supported by AI model y3. Based on conversation content z2’, user x3 has a conversation with AI model y3 to generate conversation content z3; user x4 is also interested in conversation content z2, and user x4 is accustomed to using AI model y1, so the conversation content z2 is converted into conversation content z2” supported by AI model y1. Based on conversation content z2”, user x4 has a conversation with AI model y1 to generate conversation content z4.

The embodiments of the present application assist users in freely switching between multiple AI models and having continuous conversations in the same session, and can also realize multiple user role switching, multiple AI model switching and continuous conversations in the same session.

In an optional embodiment, when the shared session content on the shared page is accessed, the access popularity information of each shared session content can also be recorded, and based on the access popularity information, multiple AI models aggregated by the aggregated application are optimized and updated. Specifically: in response to access operations on each shared session content on a shared page, access heat information of each shared session content is generated, the access operations at least include: interactive behaviors such as query, comment, like or dislike, each interactive behavior corresponds to a weight and an interaction number, the access heat information may be a heat score, and the heat score may be obtained by weighted summation/weighted average based on the respective weights and interaction numbers for each interactive behavior, and normalized to a specified interval (such as 0-1); shared session content whose access heat information is higher than a set access heat threshold is regarded as a hot shared session content; the access heat threshold may be 0.8, 0.9 or 0.95, etc., and the number of hot shared session content is multiple; based on the hot shared session content, multiple AI models aggregated in the application are optimized and updated, for example, the hot shared session content is input into each AI model, and each AI model is trained, so that the AI model builds a heat system for session content popular with users.

Optionally, the shared session content may be a single-round session, for example, the shared session content includes the first session content, or the shared session content may be a multi-round session, for example, the shared session content includes the second session content and the new session content generated on the second session interface; based on this, an implementation method of generating access heat information of each shared session content in response to an access operation to each shared session content on a shared page includes: for any shared session content, in the case where the shared session content includes multiple rounds of sessions, in response to an access operation to each round of sessions, generating access heat information of each round of sessions; based on the access heat information of each round of sessions, generating access heat information of each shared session content. For example, weighted averaging/weighted summing is performed on the access heat information of each round of sessions to generate the access heat of the session content. Among them, the user's interaction (or judgment) is fully accepted for each round of sessions to discover better session content, and the AI model can be optimized and updated to generate session content that is more in line with user preferences.

In an optional embodiment, during a conversation between a user and an AI model, an auxiliary AI model is also provided. The auxiliary AI model serves as a third party outside the user and the AI model and is used to observe the conversation content between the user and the AI model and provide the user with better suggestions for conversation techniques, so that the user can better describe the problem and obtain more valuable answers provided by the AI model.

Specifically, during a conversation between a first user and a first AI model, at least part of the conversation content in the first conversation interface is provided to the auxiliary AI model so that the auxiliary AI model can generate speech suggestion information; wherein, at least part of the conversation content can be all the content of the user on the first conversation interface, for example, all of the first conversation content, or it can also be part of the first conversation content, such as the first conversation content includes three questions of the user and three answers of the first AI model, and part of the conversation content can be three questions, or two questions and two answers, without limitation; the speech suggestion information generated by the auxiliary AI model is displayed to guide the first user to input subsequent conversation content according to the speech suggestion information, for example, the speech suggestion information can be displayed on the first conversation interface in a floating manner, or it can be displayed in the form of a list; in response to the input operation of the first user on the first conversation interface, the question information input by the first user is obtained, the question information is displayed in the first conversation interface, and the question information is provided to the first AI model, and the first AI model returns answer information based on the question information to form the conversation content between the first user and the first AI model. The question information input by the first user may be the question information edited by the user based on the speech suggestion information when the speech suggestion information is displayed, and the question information may also be the speech suggestion information clicked by the user.

Alternatively, during a conversation between the first user and the second AI model, at least part of the conversation content in the second conversation interface is provided to the auxiliary AI model so that the auxiliary AI model can generate speech suggestion information, and at least part of the conversation content can be the second conversation content, or the second conversation content and new conversation content generated in the second conversation interface; the speech suggestion information generated by the auxiliary AI model is displayed to guide the first user to input subsequent conversation content according to the speech suggestion information; and in response to the input operation of the first user on the second conversation interface, the question information input by the first user is obtained, the question information is displayed in the second conversation interface, and the question information is provided to the second AI model. For details, please refer to the above and will not be repeated here.

In an optional embodiment, before the first user interacts with the AI model, the AI can pre-learn the language features of the first user so that the AI model interacts with the first user based on the learned language features. Specifically, before the first user uses the first AI model or the second AI model to conduct a conversation, the first user's historical conversation content can also be obtained. The historical conversation content can be the conversation content between the first user and other AI models. Before the first user interacts with the second AI model, the historical conversation content can be the second conversation content; based on the first user's historical conversation content, the first user's language features are learned. The language features can be dialect features, children's features, elderly features, or celebrity features, etc.; based on the first user's language features, the language features of the first AI model or the second AI model are updated so that the first AI model or the second AI model can simulate the language features of the first user to conduct a conversation with the first user, so as to optimize the conversation content and enhance the user experience. It should be noted that when the first user conducts a conversation with the AI model for the first time, there is no historical conversation information of the first user. This is because the AI model uses the default language features to conduct a conversation with the first user.

It should be noted that the execution subject of each step of the method provided in the above embodiment can be the same device, or the method can be executed by different devices. For example, the execution subject of steps 101 to 103 can be device A; for another example, the execution subject of steps 101 and 102 can be device A, and the execution subject of step 103 can be device B; and so on.

In addition, in some of the processes described in the above embodiments and the accompanying drawings, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel, and the sequence numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the sequence numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

In addition to providing a method embodiment, the embodiment of the present application also provides a conversation processing device for artificial intelligence. The process of the conversation processing device for artificial intelligence provided by the embodiment of the present application is described below.

Figure 2 is a structural diagram of a conversation processing device for artificial intelligence provided by an exemplary embodiment of the present application. The artificial intelligence aggregation application running on the device aggregates multiple AI models in the application. As shown in Figure 2, the device includes: an operation module 21, a display module 22, a conversion middleware module 23, a providing module 24 and a switching module 25.

An operation module 21 is used to operate the first AI model when the first user needs to have a conversation with the first AI model, a display module 22 is used to display a first conversation interface of the first AI model, and the display module 22 is also used to display the first conversation content between the first user and the first AI model on the first conversation interface in response to the conversation operation between the first user and the first AI model;

A conversion middleware module 23 is used to convert the format of the first conversation content to obtain the second conversation content supported by the second AI model in response to a trigger operation of switching from the first AI model to the second AI model during the conversation between the first user and the first AI model, and a providing module 24 is used to provide the second conversation content to the second AI model;

The running module 21 is used to run the second AI model, the switching module 25 is used to switch the first conversation interface to the second conversation interface corresponding to the second AI model, and the display module 22 is also used to display the second conversation content on the second conversation interface so that the first user can continue the conversation with the second AI model based on the second conversation content.

In an optional embodiment, the conversion middleware module is specifically used to: in response to a trigger operation of switching from a first AI model to a second AI model, display an AI model list in a pop-up window on the first conversation interface, wherein the AI model list includes at least one AI model including the second AI model; in response to a selection operation of the second AI model, perform format conversion on the first conversation content to obtain second conversation content supported by the second AI model; wherein the trigger operation is implemented as a click operation on the AI model selection control on the first conversation interface, or as a voice command recognition operation or gesture operation on the first conversation interface for displaying the AI model list.

In an optional embodiment, the device also includes an unlocking module, and when the AI model list includes at least two AI models, the AI model list includes unlocked AI models and unlocked AI models; before using the conversion middleware to convert the format of the first session content, a display module is used to display an unlocking interface if the second AI model is an unlocked AI model, and the unlocking interface includes unlocking task description information; the unlocking module is used to respond to the first user's operation of performing an unlocking task on the unlocking task description information, and when the unlocking task is completed, unlock the second AI model so that the first user obtains the right to use the second AI model.

In an optional embodiment, the conversion middleware module is specifically used to: save the first session content in a universal format supported by the application to obtain the first session content in the universal format; convert the first session content in the universal format into second session content compatible with the input parameter format using an input parameter format supported by the second AI model, and the input parameter format supported by the second AI model includes at least context length information and input type information.

In an optional embodiment, the device also includes: a generation module and a publishing module; the generation module is used to generate the first shared session content of the first user according to the second session content and the new session content generated on the second session interface in response to the first user's session publishing operation on the second session interface; the publishing module is used to publish the first shared session content to other users of the aggregation application.

Optionally, the publishing module is specifically used to: publish the first shared session content to other users of the aggregation application, so that other users can continue to have a conversation with the third AI model based on the first shared session content on the third conversation interface of the third AI model. The third AI model is an AI model used by other users to conduct conversations using the artificial intelligence aggregation application.

In an optional embodiment, the publishing module is specifically used to: publish the first shared session content to a shared page provided by the aggregate application, and the shared session content on the shared page is visible to a specific user of the aggregate application.

In an optional embodiment, the device also includes: a judgment module; a judgment module, used to judge whether the fourth AI model currently used by the first user and the AI model corresponding to the second shared session content are the same in response to the first user's selection operation on the second shared session content posted by the second user on the shared page; if not, a conversion middleware module, also used to convert the format of the second shared session content to obtain the third session content supported by the fourth AI model; a display module, also used to display the third session content on the fourth session interface of the fourth AI model, so that the first user can continue the conversation with the fourth AI model based on the third session content.

In an optional embodiment, the device also includes: a processing module and an update module; a generation module, which is used to generate access heat information of each shared session content in response to the access operation of each shared session content on the shared page, and the processing module is used to take the shared session content whose access heat information is higher than the set access heat threshold as the hot shared session content; the update module is used to optimize and update the multiple AI models aggregated in the application based on the hot shared session content, and the access operation includes at least: query, comment, like or dislike.

In an optional embodiment, the generation module is specifically used to: for any shared session content, when the shared session content includes multiple rounds of sessions, in response to an access operation to each round of sessions, generate access popularity information of each round of sessions; based on the access popularity information of each round of sessions, generate access popularity information of each shared session content.

In an optional embodiment, the device also includes: an acquisition module; the providing module is also used to: during a conversation between the first user and the first or second AI model, provide at least part of the conversation content in the first or second conversation interface to the auxiliary AI model, so that the auxiliary AI model can generate speech suggestion information; the display module is also used to: display the speech suggestion information generated by the auxiliary AI model to guide the first user to input subsequent conversation content according to the speech suggestion information; and in response to the input operation of the first user on the first or second conversation interface, obtain the question information input by the first user through the acquisition module, display the question information in the first or second conversation interface through the display module, and provide the question information to the first or second AI model through the providing module.

In an optional embodiment, the device also includes: a learning module; before the first user uses the first or second AI model to conduct a conversation, the learning module is used to learn the language characteristics of the first user based on the historical conversation content of the first user; an updating module is used to update the language characteristics of the first or second AI model based on the language characteristics of the first user, so that the first or second AI model can simulate the language characteristics of the first user to conduct a conversation with the first user.

The detailed implementation and beneficial effects of each step in the device shown in Figure 2 provided in the embodiment of the present application have been described in detail in the aforementioned embodiments and will not be elaborated here.

Figure 3 is a structural diagram of a conversation processing device for artificial intelligence provided by an exemplary embodiment of the present application. The artificial intelligence aggregation application running on the device aggregates multiple AI models in the application. As shown in Figure 3, the device includes: a memory 34 and a processor 35.

The memory 34 is used to store computer programs and can be configured to store various other data to support operations on the conversation processing device for artificial intelligence. Examples of such data include instructions for any application or method operating on the conversation processing device for artificial intelligence.

The processor 35 is coupled to the memory 34 and is used to execute the computer program in the memory 34, so as to: when the first user needs to have a conversation with the first AI model, run the first AI model to display the first conversation interface of the first AI model, and in response to the dialogue operation between the first user and the first AI model, display the first conversation content between the first user and the first AI model on the first conversation interface; during the conversation between the first user and the first AI model, in response to the trigger operation of switching from the first AI model to the second AI model, use the conversion middleware to convert the format of the first conversation content to obtain the second conversation content supported by the second AI model, and provide the second conversation content to the second AI model; run the second AI model to switch the first conversation interface to the second conversation interface corresponding to the second AI model, and display the second conversation content on the second conversation interface, so that the first user can continue to have a conversation with the second AI model based on the second conversation content.

In an optional embodiment, when the processor 35 performs format conversion on the first session content using the conversion middleware in response to a trigger operation of switching from the first AI model to the second AI model to obtain the second session content supported by the second AI model, the processor 35 is specifically used to: in response to the trigger operation of switching from the first AI model to the second AI model, display an AI model list in a pop-up window on the first session interface, wherein the AI model list includes at least one AI model including the second AI model; in response to a selection operation of the second AI model, perform format conversion on the first session content using the conversion middleware to obtain the second session content supported by the second AI model; wherein the trigger operation is implemented as a click operation on the AI model selection control on the first session interface, or as a voice command recognition operation or gesture operation on the first session interface for displaying the AI model list.

In an optional embodiment, when the AI model list includes at least two AI models, the AI model list includes unlocked AI models and unlocked AI models; before using the conversion middleware to convert the format of the first session content, the processor 35 is also used to: if the second AI model is an unlocked AI model, display an unlocking interface, which includes unlocking task description information; in response to the first user performing an unlocking task on the unlocking task description information, when the unlocking task is completed, unlock the second AI model so that the first user obtains the right to use the second AI model.

In an optional embodiment, when the processor 35 uses the conversion middleware to convert the format of the first session content to obtain the second session content supported by the second AI model, it is specifically used to: save the first session content in a general format supported by the application to obtain the first session content in the general format; use the input parameter format supported by the second AI model to convert the first session content in the general format into the second session content compatible with the input parameter format, and the input parameter format supported by the second AI model includes at least: context length information and input type information.

In an optional embodiment, the processor 35 is further used to: in response to a session publishing operation of the first user on the second session interface, generate a first shared session content of the first user according to the second session content and new session content generated on the second session interface; and publish the first shared session content to other users of the aggregate application.

Optionally, when the processor 35 publishes the first shared session content to other users of the aggregation application, it is specifically used to: publish the first shared session content to other users of the aggregation application, so that other users can continue to have a conversation with the third AI model based on the first shared session content on the third conversation interface of the third AI model, and the third AI model is an AI model used by other users to conduct conversations using the artificial intelligence aggregation application.

In an optional embodiment, when publishing the first shared session content to other users of the aggregate application, the processor 35 is specifically configured to: publish the first shared session content to a shared page provided by the aggregate application, and the shared session content on the shared page is visible to specific users of the aggregate application.

In an optional embodiment, the processor 35 is also used to: in response to the first user's selection operation on the second shared session content posted by the second user on the shared page, determine whether the fourth AI model currently used by the first user and the AI model corresponding to the second shared session content are the same; if not, use the conversion middleware to convert the format of the second shared session content to obtain the third session content supported by the fourth AI model; display the third session content on the fourth session interface of the fourth AI model, so that the first user can continue the conversation with the fourth AI model based on the third session content.

In an optional embodiment, the processor 35 is also used to: generate access popularity information of each shared session content in response to access operations on each shared session content on the shared page, and use the shared session content whose access popularity information is higher than the set access popularity threshold as the popular shared session content; based on the popular shared session content, optimize and update multiple AI models aggregated in the application, and the access operations include at least: query, comment, like or dislike.

In an optional embodiment, when the processor 35 generates access popularity information of each shared session content in response to an access operation to each shared session content on a shared page, it is specifically used to: for any shared session content, when the shared session content includes multiple rounds of sessions, in response to an access operation to each round of sessions, generate access popularity information of each round of sessions; based on the access popularity information of each round of sessions, generate access popularity information of each shared session content.

In an optional embodiment, the processor 35 is also used to: during a conversation between the first user and the first or second AI model, provide at least part of the conversation content in the first or second conversation interface to the auxiliary AI model, so that the auxiliary AI model can generate speech suggestion information; display the speech suggestion information generated by the auxiliary AI model to guide the first user to input subsequent conversation content according to the speech suggestion information; and in response to the first user's input operation on the first or second conversation interface, obtain the question information input by the first user, display the question information in the first or second conversation interface, and provide the question information to the first or second AI model.

In an optional embodiment, before the first user uses the first or second AI model to conduct a conversation, the processor 35 is further used to: learn the language characteristics of the first user based on the historical conversation content of the first user; and update the language characteristics of the first or second AI model based on the language characteristics of the first user, so that the first or second AI model can simulate the language characteristics of the first user to conduct a conversation with the first user.

The detailed implementation and beneficial effects of each step in the device shown in Figure 3 provided in the embodiment of the present application have been described in detail in the aforementioned embodiments and will not be elaborated here.

Further, as shown in FIG3, the conversation processing device for artificial intelligence also includes: communication component 36, display 33, power supply component 38, audio component 39 and other components. FIG3 only schematically shows some components, which does not mean that the conversation processing device for artificial intelligence only includes the components shown in FIG3. In addition, the components in the dashed box in FIG3 are optional components, not mandatory components, and may depend on the product form of the conversation processing device for artificial intelligence. The conversation processing device for artificial intelligence of this embodiment can be implemented as a terminal device such as a desktop computer, a laptop computer, a smart phone or an IOT device, or a server-side device such as a conventional server, a cloud server or a server array. If the conversation processing device for artificial intelligence of this embodiment is implemented as a terminal device such as a desktop computer, a laptop computer, a smart phone, etc., it may include the components in the dashed box in FIG3; if the conversation processing device for artificial intelligence of this embodiment is implemented as a server-side device such as a conventional server, a cloud server or a server array, it may not include the components in the dashed box in FIG3. Among them, when the conversation processing device for artificial intelligence is implemented as a server-side device, that is, the artificial intelligence aggregation application runs on the server, it can be displayed as a cloud.

In an optional implementation, cloud display refers to an information display method based on cloud computing. In the operation mode of cloud display, the operation subject of the artificial intelligence aggregation application and the presentation subject of the conversation interface are separated. The data storage and operation corresponding to the artificial intelligence aggregation application can be completed on the cloud display server. The role of the cloud display client is to receive and send data and present the picture. For example, the cloud display client can be a display device with data transmission function close to the user side, such as a mobile terminal, a television, a computer, a handheld computer, etc.; but the electronic terminal for information data processing is a cloud display server in the cloud. When browsing information, the user operates the cloud display client to send an operation instruction to the cloud display server. The cloud display server displays the relevant pictures according to the operation instruction, encodes and compresses the data, and returns it to the cloud display client through the network. Finally, the cloud display client decodes and outputs the corresponding picture.

In another optional embodiment, the conversation processing device for artificial intelligence can be a local electronic terminal. The local electronic terminal stores the program code of the artificial intelligence aggregation application and is used to present the application interface. The local electronic terminal is used to interact with the user through a graphical user interface, that is, the application is downloaded and installed and run by the conventional electronic device. The local electronic terminal may provide the graphical user interface to the user in a variety of ways, for example, it can be rendered and displayed on the display screen of the terminal, or provided to the user through a holographic projection. For example, the local electronic terminal may include a display screen and a processor, the display screen is used to present a graphical user interface, the graphical user interface includes an application screen, and the processor is used to run the application, generate a graphical user interface, and control the display of the graphical user interface on the display screen.

The detailed implementation and beneficial effects of the conversation processing device for artificial intelligence provided in the embodiments of the present application have been described in detail in the aforementioned embodiments and will not be elaborated here.

Accordingly, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, which, when executed, can implement the steps that can be performed by the session state synchronization device in the method embodiment shown in FIG. 1a above.

The above-mentioned memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable programmable read only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

The above-mentioned communication component is configured to facilitate wired or wireless communication between the device where the communication component is located and other devices. The device where the communication component is located can access a wireless network based on a communication standard, such as WiFi, 2G, 3G, 4G/LTE, 5G and other mobile communication networks, or a combination thereof. In an exemplary embodiment, the communication component receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be based on Radio Frequency Identification (RFID) technology, Infrared Data Association (IrDA) technology, Ultra Wide Band (UWB) technology, Bluetooth (BT) technology and other technologies.

The above-mentioned display includes a screen, and the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation.

The power supply assembly provides power to various components of the device where the power supply assembly is located. The power supply assembly may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the device where the power supply assembly is located.

The above-mentioned audio component can be configured to output and/or input audio signals. For example, the audio component includes a microphone (Microphone, MIC), and when the device where the audio component is located is in an operating mode, such as a call mode, a recording mode, and a speech recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in a memory or sent via a communication component. In some embodiments, the audio component also includes a speaker for outputting an audio signal.

Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, compact disc read-only memory (Compact Disc Read-Only Memory, CD-ROM), optical storage, etc.) containing computer-usable program code.

The present application is described with reference to the flowchart and/or block diagram of the method, device (system) and computer program product according to the embodiment of the present application. It should be understood that each process and/or box in the flowchart and/or block diagram, and the combination of the process and/or box in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for realizing the function specified in one process or multiple processes in the flowchart and/or one box or multiple boxes in the block diagram.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device that implements the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

In a typical configuration, a computing device includes one or more processors (Central Processing Unit, CPU), input/output interface, network interface and memory.

Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and/or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

Computer readable media include permanent and non-permanent, removable and non-removable media that can be used to store information by any method or technology. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase-change random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims (15)

1. A session handling method for artificial intelligence, adapted for use in an artificial intelligence aggregation application in which a plurality of AI models are aggregated, the method comprising:

In the event that a first user needs to session with a first AI model, running the first AI model to expose a first session interface of the first AI model, and responsive to a dialog operation between the first user and the first AI model, exposing first session content between the first user and the first AI model on the first session interface;

In the process of the conversation between the first user and the first AI model, responding to the triggering operation of switching from the first AI model to a second AI model, carrying out format conversion on the first conversation content by utilizing a conversion middleware so as to obtain second conversation content supported by the second AI model, and providing the second conversation content to the second AI model; the triggering operation is initiated by the first user;

Operating the second AI model to switch the first session interface to a second session interface corresponding to the second AI model, and displaying the second session content on the second session interface so that the first user can continue to perform a session based on the second session content and the second AI model; the first session interface is the same as or similar to the second session interface.

2. The method of claim 1, wherein responsive to a trigger operation to switch from the first AI model to a second AI model, formatting the first session content with a conversion middleware to obtain second session content supported by the second AI model comprises:

Responsive to a trigger operation to switch from the first AI model to a second AI model, displaying a list of AI models on the first session interface in a popup window, at least one AI model of the list of AI models including the second AI model;

Responding to the selection operation of the second AI model, and performing format conversion on the first session content by utilizing conversion middleware to obtain second session content supported by the second AI model; the triggering operation is implemented as a clicking operation of an AI model selection control on the first session interface, or as a voice instruction recognition operation or a gesture operation for displaying an AI model list on the first session interface.

3. The method of claim 2, wherein, in the case where the AI model list includes at least two AI models, the AI model list includes an unlocked AI model and an unlocked AI model; before format conversion of the first session content using the conversion middleware, the method further comprises:

If the second AI model is an unlocked AI model, displaying an unlocking interface, wherein the unlocking interface comprises unlocking task description information;

And responding to the operation of the first user on the unlocking task description information to execute an unlocking task, and under the condition that the unlocking task is completed, carrying out unlocking operation on the second AI model so as to enable the first user to obtain the use authority of the second AI model.

4. The method of claim 1, wherein format converting the first session content with conversion middleware to obtain second session content supported by the second AI model comprises:

The universal format supported by the application is adopted to store the first session content, so that the first session content in the universal format is obtained;

and converting the first session content in the general format into second session content adapted to the access format by adopting the access format supported by a second AI model, wherein the access format supported by the second AI model at least comprises: context length information and input type information.

5. The method according to claim 1, wherein the method further comprises:

Responding to the session release operation of the first user on the second session interface, and generating first shared session content of the first user according to the second session content and new session content generated on the second session interface;

And publishing the first shared session content to other users of the aggregation application.

6. The method of claim 5, wherein publishing the first shared session content to other users of the aggregated application comprises:

And publishing the first shared session content to other users of the aggregation application so that the other users can continue to perform a session with a third AI model on a third session interface of the third AI model based on the first shared session content, wherein the third AI model is an AI model used by the other users for performing the session by adopting the artificial intelligence aggregation application.

7. The method of claim 5, wherein publishing the first shared session content to other users of the aggregated application comprises:

And publishing the first shared session content to a shared page provided by the aggregation application, wherein the shared session content on the shared page is visible to a specific user of the aggregation application.

8. The method as recited in claim 7, further comprising:

Responding to the selection operation of the first user on the second shared session content issued by the second user on the shared page, and judging whether a fourth AI model currently used by the first user is the same as an AI model corresponding to the second shared session content;

If not, format conversion is carried out on the second shared session content by utilizing a conversion middleware so as to obtain third session content supported by the fourth AI model;

and displaying the third session content on a fourth session interface of the fourth AI model for the first user to continue the session based on the third session content and the fourth AI model.

9. The method of claim 7, wherein the method further comprises:

Responding to the access operation of each shared session content on the shared page, generating access heat information of each shared session content, and taking the shared session content with the access heat information higher than the set access heat threshold as hot shared session content;

And based on the content of the hot sharing session, optimizing and updating a plurality of AI models aggregated in the application, wherein the access operation at least comprises: query, comment, praise, or click.

10. The method of claim 7, wherein generating access heat information for each shared session content in response to an access operation for each shared session content on the shared page comprises:

for any shared session content, in the case that the shared session content includes multiple rounds of sessions, generating access heat information of each round of session in response to an access operation for each round of session;

And generating access heat information of the content of each shared session according to the access heat information of each round session.

11. The method as recited in claim 1, further comprising:

providing at least part of session content in the first or second session interface to an auxiliary AI model in the process of the first user and the first or second AI model for the auxiliary AI model to generate speech suggestion information;

Displaying the speaking suggestion information generated by the auxiliary AI model so as to guide the first user to input subsequent speaking contents according to the speaking suggestion information; and

And responding to the input operation of the first user on a first or second session interface, acquiring the question information input by the first user, displaying the question information in the first or second session interface, and providing the question information to the first or second AI model.

12. The method of claim 1, further comprising, prior to the first user conducting a session using the first or second AI model:

Learning language features of a first user according to historical session content of the first user; and updating the language features of the first or second AI model according to the language features of the first user so that the first or second AI model can simulate the language features of the first user to carry out conversation with the first user.

13. A session processing apparatus for artificial intelligence, wherein an artificial intelligence aggregation application running on the apparatus aggregates a plurality of AI models therein, the apparatus comprising: the system comprises an operation module, a display module, a conversion middleware module, a providing module and a switching module;

The operation module is used for operating the first AI model under the condition that a first user needs to conduct a session with the first AI model, the display module is used for displaying a first session interface of the first AI model, and the display module is also used for responding to the dialogue operation between the first user and the first AI model and displaying first session content between the first user and the first AI model on the first session interface;

The conversion middleware module is configured to perform format conversion on the first session content in response to a trigger operation of switching from the first AI model to a second AI model during a session between the first user and the first AI model, so as to obtain second session content supported by the second AI model, and the providing module is configured to provide the second session content to the second AI model; the triggering operation is initiated by the first user;

The operation module is configured to operate the second AI model, the switching module is configured to switch the first session interface to a second session interface corresponding to the second AI model, and the display module is further configured to display the second session content on the second session interface, so that the first user continues to perform a session based on the second session content and the second AI model; the first session interface is the same as or similar to the second session interface.

14. A session handling device for artificial intelligence, comprising: a memory and a processor; the memory is used for storing a computer program; the processor, coupled to the memory, for executing the computer program to implement the steps in the method of any of claims 1-12.

15. A computer readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, causes the processor to carry out the steps of the method according to any one of claims 1-12.