Interaction method and device, electronic equipment and storage medium

Abstract

The invention discloses an interaction method and device, electronic equipment and a storage medium, and relates to the technical field of computers, in particular to the field of artificial intelligence. According to the specific implementation scheme, selection of an object on an intelligent agent in a social data interaction platform is received; in response to the selection, displaying an interaction interface corresponding to an agent deployed in a target function unit of the social data interaction platform; and inputting an interaction request input by the object in the interaction interface and information representing the content characteristics of the target function unit into the large model so as to present an interaction response matched with the content characteristics on the interaction interface.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technologies, and more particularly to the field of artificial intelligence, and specifically to an interaction method, apparatus, electronic device, and storage medium. Background Art

[0002] Currently, the widely used dialogue robots mainly focus on interactive and auxiliary functions, and their dialogue capabilities are usually built based on the Q&A form. These robots analyze the information input by users, quickly understand the intention of the questions, and provide accurate answers, thereby helping users obtain the required information efficiently. Summary of the Invention

[0003] The present disclosure provides an interaction method, apparatus, electronic device, and storage medium.

[0004] According to one aspect of the present disclosure, there is provided an interaction method, including: receiving a selection of an agent by an object in a social data interaction platform; in response to the above selection, displaying an interaction interface corresponding to the above agent, where the above agent is deployed in a target functional unit of the above social data interaction platform; inputting an interaction request input by the above object in the above interaction interface and information characterizing the content features of the above target functional unit into a large model, so as to present an interaction response matching the above content features on the above interaction interface.

[0005] According to another aspect of the present disclosure, there is provided an interaction apparatus, including: a selection receiving module, configured to receive a selection of an agent by an object in a social data interaction platform; an interface display module, configured to display an interaction interface corresponding to the above agent in response to the above selection, where the above agent is deployed in a target functional unit of the above social data interaction platform; a request response module, configured to input an interaction request input by the above object in the above interaction interface and information characterizing the content features of the above target functional unit into a large model, so as to present an interaction response matching the above content features on the above interaction interface.

[0006] According to another aspect of the present disclosure, there is provided an electronic device, including: at least one processor; and a memory communicatively connected to the above at least one processor; wherein, the above memory stores instructions executable by the above at least one processor, and the above instructions are executed by the above at least one processor so that the above at least one processor can execute the method as described above.

[0007] According to another aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions, wherein the above computer instructions are used to cause the above computer to execute the method as described above.

[0008] According to another aspect of the present disclosure, there is provided a computer program product including a computer program which, when executed by a processor, implements the method as described above.

[0009] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The drawings are used to better understand the solution and do not constitute a limitation to the present disclosure. Among them:

[0011] Figure 1 Schematically shows an exemplary system architecture to which the interaction method and apparatus according to an embodiment of the present disclosure can be applied;

[0012] Figure 2 Schematically shows a flowchart of the interaction method according to an embodiment of the present disclosure;

[0013] Figure 3A Schematically shows a schematic diagram of an interface of guiding information according to an embodiment of the present disclosure;

[0014] Figure 3B Schematically shows a schematic diagram of an interface of guiding information according to another embodiment of the present disclosure;

[0015] Figure 4 Schematically shows a flowchart of generating guiding information according to an embodiment of the present disclosure;

[0016] Figure 5A Schematically shows a schematic diagram of an interface of an intelligent sharing post according to an embodiment of the present disclosure;

[0017] Figure 5B Schematically shows a flowchart of using historical interaction information according to an embodiment of the present disclosure;

[0018] Figure 5C Schematically shows a schematic diagram of an interface of an interaction Q&A according to an embodiment of the present disclosure;

[0019] Figure 6 Schematically shows a flowchart of risk detection according to an embodiment of the present disclosure;

[0020] Figure 7 Schematically shows a schematic diagram of an interface of an agent creation according to an embodiment of the present disclosure;

[0021] Figure 8 Schematically shows a block diagram of an interaction apparatus according to an embodiment of the present disclosure; and

[0022] Figure 9A block diagram of an electronic device suitable for implementing an interaction method according to an embodiment of the present disclosure is schematically shown. Detailed implementation manners

[0023] The following describes exemplary embodiments of the present disclosure with reference to the accompanying drawings. Various details of the embodiments of the present disclosure are included to facilitate understanding, and they should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0024] Figure 1 An exemplary system architecture to which an interaction method and apparatus can be applied according to an embodiment of the present disclosure is schematically shown.

[0025] It should be noted that Figure 1 The example shown is only an example of the system architecture to which the embodiments of the present disclosure can be applied, to help those skilled in the art understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios. For example, in another embodiment, the exemplary system architecture to which the interaction method and apparatus can be applied may include terminal devices, but the terminal devices may implement the interaction method and apparatus provided by the embodiments of the present disclosure without interacting with the server.

[0026] As Figure 1 shown, the system architecture 100 according to this embodiment may include terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 is used as a medium to provide a communication link between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0027] Users can use the terminal devices 101, 102, 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications may be installed on the terminal devices 101, 102, 103, such as knowledge reading applications, web browser applications, search applications, instant messaging tools, email clients, and / or social platform software, etc. (only as examples).

[0028] The terminal devices 101, 102, 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop portable computers, and desktop computers, etc.

[0029] Server 105 may be a server that provides various services, such as a background management server (for example only) that supports the content browsed by users using terminal devices 101, 102, and 103. The background management server may analyze and process data such as user requests received, and feedback the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0030] It should be noted that the interaction method provided by the embodiments of the present disclosure can generally be executed by terminal devices 101, 102, or 103. Correspondingly, the interaction device provided by the embodiments of the present disclosure can also be set in terminal devices 101, 102, or 103.

[0031] Alternatively, the interaction method provided by the embodiments of the present disclosure can generally also be executed by server 105. Correspondingly, the interaction device provided by the embodiments of the present disclosure can generally be set in server 105. The interaction method provided by the embodiments of the present disclosure can also be executed by a server or a server cluster different from server 105 and capable of communicating with terminal devices 101, 102, 103 and / or server 105. Correspondingly, the interaction device provided by the embodiments of the present disclosure can also be set in a server or a server cluster different from server 105 and capable of communicating with terminal devices 101, 102, 103 and / or server 105.

[0032] It should be understood that Figure 1 the numbers of terminal devices, networks, and servers in

[0033] In the technical solution of the present disclosure, the processing of collection, storage, use, processing, transmission, provision, disclosure, and application of user personal information involved all comply with the provisions of relevant laws and regulations, take necessary confidentiality measures, and do not violate public order and good customs.

[0034] In the technical solution of the present disclosure, before obtaining or collecting user personal information, the authorization or consent of the user is obtained.

[0035] Figure 2 Schematically shows a flowchart of an interaction method according to an embodiment of the present disclosure.

[0036] As Figure 2 shown, the method includes operations S210 and S220.

[0037] In operation S210, receive the selection of an agent by an object in the social data interaction platform.

[0038] In operation S220, in response to the selection, an interaction interface corresponding to the agent is displayed, where the agent is deployed in a target functional unit of the social data interaction platform.

[0039] In operation S230, the interaction request input by the object in the interaction interface and the information characterizing the content features of the target functional unit are input into the large model, so as to present an interaction response matching the content features on the interaction interface.

[0040] According to the embodiments of the present disclosure, most current dialogue robots adopt preset roles as their built-in dialogue personas. However, this preset role mode has limitations in terms of flexibility and personalization, and it is difficult to meet the diverse user needs in different community environments.

[0041] According to the embodiments of the present disclosure, to solve the above problems, in a social data interaction platform (such as Tieba), a dedicated agent (i.e., "AI role") is customized for each functional unit (such as the interest theme module in Tieba), and content features that fit the interest theme are injected into it, so that the AI role can naturally integrate into the community environment. For example, in a specific interest theme module, multiple agents may be deployed, and each agent corresponds to a different aspect or sub-theme of the interest theme.

[0042] According to the embodiments of the present disclosure, when receiving the user's selection of an agent in the social data interaction platform, the interaction interface corresponding to the agent is automatically loaded. Among them, the user's selection can be triggered by a touch operation on the client interface. After the interaction interface is loaded, the interaction request input by the user through the interaction interface is collected, and at the same time, the information used to characterize the content features of the target functional unit is obtained. The obtained information and the interaction request are input into the large model for processing to generate an interaction response matching the content features.

[0043] For example, if the user interacts with a certain agent (such as "landscape photography expert") in an interest theme module about "photography", the system will generate a targeted interaction response based on the user's interaction request and the content features of this module (such as landscape photography skills, equipment recommendations, etc.).

[0044] According to the embodiments of the present disclosure, by deploying agents in the social data interaction platform, users can directly experience the artificial intelligence interaction function in the community. Compared with the need to specially download a separate application to use the artificial intelligence generation content function, the user path is shorter and the operation is more convenient. During the process of the user interacting with the agent, the system will simultaneously input the user's interaction request and the information characterizing the content features of the target functional unit into the large model, so that the output interaction response is more personalized and can provide more intelligent, accurate and relevant responses and services for users.

[0045] The following refers to FIGS. 3 to Figure 7 , and further describes the Figure 2 method shown with reference to specific embodiments.

[0046] According to an embodiment of the present disclosure, the interaction method further includes: determining an interaction record between an object and an intelligent agent in a target functional unit; dynamically generating guiding information based on the interaction record, where the guiding information is used to guide the object to input an interaction request; and displaying the guiding information when the interaction interface is loaded.

[0047] According to an embodiment of the present disclosure, an interaction record between a user and an intelligent agent in a target functional unit is obtained. The interaction record covers all interaction behaviors between the user and the intelligent agent, including interaction requests input by the user, interaction responses of the intelligent agent, and feedback of the user on the interaction responses, etc.

[0048] According to an embodiment of the present disclosure, dynamically generating guiding information based on the interaction record involves identifying the user's interaction patterns and preferences. For example, if the user often asks questions about a specific interest topic in the history, the system may generate guiding information related to the interest topic to enable the user to further explore or deepen related issues. The purpose of the guiding information is to help the user input interaction requests more effectively, thereby improving the efficiency and quality of the interaction.

[0049] As Figure 3A shown, when the user enters the interaction interface to interact with the target intelligent agent 301, the intelligent agent generates guiding information 302 and displays it in the interaction interface. The guiding information 302 includes multiple information items for the user to select, such as Information 1, Information 2, and Information 3, and may also be a complete piece of information. The user can input a corresponding interaction request through the request input box 304 located at the bottom of the interaction interface, or directly click on any option in the guiding information 302 to make a selection.

[0050] According to an embodiment of the present disclosure, by displaying the guiding information in the interaction interface, the user can more quickly understand how to communicate effectively with the intelligent agent, reducing interaction barriers caused by improper operations or misunderstandings. In addition, the guiding information can also help the user explore the functions of the intelligent agent more deeply, discover interaction options that may not have been noticed, and thus make full use of various services provided by the platform.

[0051] As Figure 3A shown, the interaction interface also includes multiple candidate intelligent agents 303. These candidate intelligent agents 303 are displayed side by side with the target intelligent agent 301 and are all deployed in the target functional unit, but their themes may be different, providing diverse interaction options for the user. This design enables the user to select the most suitable intelligent agent for interaction according to their own needs and preferences, thereby improving the flexibility and personalization of the interaction.

[0052] According to an embodiment of the present disclosure, considering that the deployment of agents in the social data interaction platform combines two dimensions, horizontal and vertical, the ways for users to select agent interactions are relatively diverse. In the horizontal dimension, agents are classified according to functional units of different interest themes, enabling users to locate relevant agents in each functional unit. This classification method helps users quickly find the corresponding agents according to their interests, thereby improving the convenience and pertinence of interactions.

[0053] According to an embodiment of the present disclosure, in the vertical dimension, all agents in the social data interaction platform will be carefully divided according to their categories. This division method enables users to quickly find the agents they want to interact with based on specific classification criteria in multiple categories in the agent square of the social data interaction platform, and then conduct more in-depth communication.

[0054] As Figure 3B shown, for the classification in the vertical dimension, users can select the agents they want to interact with in the agent square of the social data interaction platform by browsing different agent categories 305. These categories may include multiple options such as Category 1, Category 2, Category 3, and Category 4. After the user selects the target agent, the system will load the interaction interface 306 that matches the target agent. In the interaction interface 306, corresponding guiding information will be displayed to guide the user on how to interact effectively with the agent. Such a design enables users to quickly find and access the interaction functions they are interested in based on the classification of agents, thereby enhancing the convenience and pertinence of the user experience.

[0055] According to an embodiment of the present disclosure, after it is determined that the user has completed the selection of an agent from the agent square, if it is detected that the agent clicked by the user belongs to the tool type, then the tool-type agent can be used to make further agent recommendations for the user. Specifically, semantic analysis is performed on the interaction request input by the user through natural language processing technology, so as to extract the core keywords and operation intentions. This process can adopt a keyword extraction algorithm enhanced by an attention mechanism to ensure accurate capture of the essential needs of the user's query.

[0056] According to an embodiment of the present disclosure, after semantic analysis, the historical behavior records of the user in the social data interaction platform are obtained, specifically including the access path, stay duration, and interaction frequency of functional units. At the same time, the behavior feature sets of similar user groups are screened out through a portrait clustering algorithm. Based on the collaborative filtering algorithm, a user-functional unit association matrix is constructed, and the similarity index of the target user and the similar user groups in the access mode of interest theme functional units is calculated.

[0057] According to an embodiment of the present disclosure, semantic features and similarity metrics are weighted and fused, and a sorting learning algorithm is used to prioritize candidate interest topic functional units, and finally the functional unit with the highest matching degree is selected to display the recommended agent on the interaction interface.

[0058] According to an embodiment of the present disclosure, by analyzing the keywords of the user interaction request, historical behavior, and the behavior patterns of similar user groups, it is possible to accurately identify the interest topics related to the user interaction request from the social data interaction platform, and display multiple agents under the relevant functional units on the interaction interface, thereby providing a more customized and interesting interaction experience for the user.

[0059] Figure 4 Schematically shows a flowchart of generating guiding information according to an embodiment of the present disclosure.

[0060] As Figure 4 shown, the method includes operations S410 to S460.

[0061] In operation S410, determine the agent selected by the object in the social data interaction platform.

[0062] In operation S420, determine the interaction record between the object and the agent in the target functional unit.

[0063] In operation S430, determine whether the interaction record indicates that there is a historical interaction between the object and the agent.

[0064] In operation S440, determine that the guiding information is the basic guidance for the agent.

[0065] In operation S450, generate guiding information matching the target functional unit according to the interaction data generated by the historical interaction.

[0066] In operation S460, display the guiding information on the interaction interface.

[0067] According to an embodiment of the present disclosure, the obtained interaction record is analyzed to determine whether the user and the agent belong to the target interaction. Among them, the target interaction refers to the initial interaction scenario that meets the preset criteria. When it is confirmed as the target interaction, the corresponding guiding information can be selected from the preset basic guidance template library. These templates are designed for various common initial interaction scenarios and contain general guiding content such as the basic functions of the agent, operation methods, and types of information that can be input, aiming to help users quickly establish a basic understanding of the agent and thus successfully complete the first interaction. The guiding information provided at this time is the basic guidance for the agent.

[0068] According to an embodiment of the present disclosure, if the analysis result of the interaction record indicates that there is a historical interaction between the user and the intelligent agent, the interaction data generated by the historical interaction between the user and the intelligent agent is obtained. Among them, the interaction data includes the specific content of the interaction requests historically input by the user, the response details of the intelligent agent, and the subsequent feedback evaluations of the user, etc.

[0069] According to an embodiment of the present disclosure, the interaction data is input into a processing module with a data analysis algorithm and a machine learning model, so as to dynamically generate matching guiding information through natural language generation technology. The guiding information generated in this way is more targeted and can accurately guide the user to input more interaction requests that meet their own needs and are more efficient in the current interaction, thereby improving the overall interaction experience.

[0070] According to an embodiment of the present disclosure, based on the interaction data generated by the historical interaction, guiding information matching the target functional unit is dynamically generated, including: extracting the interaction features of the object from the interaction data, where the interaction features include at least one of semantic features, context features, and behavior pattern features; fusing the interaction features with the real-time dynamic information of the target functional unit to generate guiding information, where the real-time dynamic information includes the real-time behavior data of the object in the target functional unit and the heat change information in the target functional unit.

[0071] According to an embodiment of the present disclosure, after obtaining the interaction data, lexical analysis, syntactic analysis, and semantic understanding algorithms in natural language processing technology can be used to extract and analyze the semantic features of the user during the historical interaction process. Among them, the semantic features include the professional terms that the user prefers to use, the common sentence patterns, and the positive or negative semantic attitudes that the user tends to express when asking questions, etc.

[0072] According to an embodiment of the present disclosure, by identifying the associated information of the object in different interaction rounds, the context features are further extracted. The composition of the context features not only includes the logical connections between adjacent interaction statements, such as causal relationships, progressive relationships, and transitional relationships, but also includes the overall logical structure between the initial question and the subsequent follow-up questions and feedbacks during the entire interaction process. Extracting the context features helps the intelligent agent to more accurately understand the complete intention of the user.

[0073] According to an embodiment of the present disclosure, combining the user's operation behaviors on the interaction interface, such as click behaviors, input duration, modification times, and the time interval and frequency of interacting with the intelligent agent and other multi-dimensional factors, the user's behavior pattern is determined. For example, the system can determine whether the user tends to quickly input short questions and then ask frequent follow-up questions, or raise complex questions at one time. By extracting at least one of the above features to construct interaction features, a basis is provided for generating targeted guiding information subsequently.

[0074] According to the embodiments of the present disclosure, after completing the interactive feature extraction, the real-time dynamic information of the target functional unit is obtained. Specifically, the user's operating actions, location, browsing path, etc. can be captured in real time through the data monitoring points in the target functional unit. For example, the content that the user is currently browsing and the key operations performed in the past are recorded. In addition, the popularity of various types of information in the target functional unit is also monitored at the same time, including the popularity of the topic, the frequency of use of the function, etc., in order to understand the real-time changes in the popularity.

[0075] According to an embodiment of the present disclosure, real-time dynamic information is fused with interactive features to generate guidance information. Specifically, the fusion algorithm in the machine learning model can be used for fusion, such as a multi-source information fusion model based on an attention mechanism. After the fusion is completed, the fusion result is converted into guidance information using natural language generation technology so as to be displayed in the interactive interface. By fusing real-time dynamic information and interactive features to generate guidance information, the generated guidance information meets both the needs of the user and the actual situation of the target functional unit.

[0076] According to the embodiments of the present disclosure, when users browse content on a social data interaction platform, they can interact with an agent in the following two ways: one is to directly click on the agent entrance in the platform to initiate a conversation; the other is to quickly enter the interaction interface with the agent by clicking on the interaction entrance in the agent sharing post shared by other users. This dual-path interaction design not only retains the accessibility of the platform's native agents, but also makes full use of the social sharing mechanism to expand the reach of the agent.

[0077] like Figure 5A As shown, when entering the interactive interface, the user can not only enter by clicking the agent option in the title drop-down box 501, but also choose to click the agent sharing post 502 in the target functional unit to access it.

[0078] According to the embodiments of the present disclosure, the smart sharing post can effectively gather users to participate in the discussion of specific topics through a precise theme guidance mechanism. This directional interaction mode not only strengthens the topic focus within the social data interaction platform, but also significantly improves the content popularity and user activity level of the platform.

[0079] Figure 5B The flowchart of using historical interaction information according to an embodiment of the present disclosure is schematically shown.

[0080] like Figure 5B As shown, the method includes operations S510~S590.

[0081] In operation S510, an operation of an object on an agent in a social data interaction platform is received.

[0082] In operation S520, it is determined whether the operation is triggered by an object clicking on an agent sharing post in the social data interaction platform.

[0083] In operation S530, the agent entry component built into the agent sharing post is called to read the historical interaction data created by the object sharing the agent post.

[0084] In operation S540, the usage instructions for the historical interaction data are displayed on the interaction interface.

[0085] In operation S550, it is determined whether the object agrees to use the historical interaction data.

[0086] In operation S560, according to the interaction request input by the object, a response query is performed in the historical interaction information.

[0087] In operation S570, it is determined whether the historical interaction information includes an interaction response corresponding to the interaction request.

[0088] In operation S580, the queried interaction response is output.

[0089] In operation S590, the interaction request input by the object on the interaction interface and the information characterizing the content features of the target functional unit are input into the large model, and an interaction response matching the content features is output.

[0090] According to an embodiment of the present disclosure, after receiving an operation instruction executed by a user on an agent in the social data interaction platform, by analyzing the operation trigger path, it is determined whether the operation is triggered by the user clicking on an agent sharing post published in the social data interaction platform. Specifically, the determination process can be completed based on an event tracking mechanism.

[0091] According to an embodiment of the present disclosure, when it is confirmed that the operation originates from an agent sharing post, the agent entry component built into the agent sharing post is activated. This component has data reading permissions and can read the historical interaction data pre-authorized by the sharer. After the data is loaded, usage instructions and authorization agreements for the historical interaction data will be dynamically generated and displayed on the interaction interface.

[0092] According to an embodiment of the present disclosure, after the user expresses agreement to the data usage terms through an explicit confirmation operation on the interaction interface, the interaction request input by the user is structurally parsed, and a multi-dimensional query is performed in the historical interaction information. Specifically, the query process can adopt semantic similarity calculation and intention recognition algorithms to verify whether there is a preset response in the historical interaction information that conforms to the current interaction request.

[0093] According to an embodiment of the present disclosure, if there is a matching item, the corresponding interaction response is directly output. If there is no matching result, the current interaction request and the information characterizing the content features of the target functional unit are jointly input into the large model, so as to output an interaction response adapted to the user requirements and context features through feature space mapping and generative reasoning.

[0094] According to an embodiment of the present disclosure, since the waiting time for generating an interaction response through the large model is relatively long, by reasonably judging the trigger conditions of operations and effectively using historical interaction data, the efficiency and experience of the user's interaction with the intelligent agent can be improved, making the service of the intelligent agent more in line with the user's needs.

[0095] According to an embodiment of the present disclosure, in the process of the user interacting with the intelligent agent, it is usually in the form of question-and-answer pairs. As Figure 5C shown, after the user enters the interaction interface, a first interaction request is first sent, such as "Interaction Request 1" in the figure. After receiving this request, the intelligent agent will generate and display a corresponding interaction response, that is, "Interaction Response 1", thus completing the first interaction and forming the first question-and-answer pair 503. If the user needs further interaction, a second interaction request can be sent, such as "Interaction Request 2" in the figure, and the intelligent agent will also generate and display the corresponding "Interaction Response 2", thus completing the second interaction and forming the second question-and-answer pair 504.

[0096] According to an embodiment of the present disclosure, through the instant response of the intelligent agent to the user's request, the user can obtain a quick feedback, and the form of forming question-and-answer pairs simulates the process of human natural conversation, making the interaction more intuitive and natural.

[0097] Figure 6 Schematically shows a flowchart of risk detection according to an embodiment of the present disclosure.

[0098] As Figure 6 shown, the method includes operations S601 to S610.

[0099] In operation S601, the system receives an interaction request input by the user on the interaction interface.

[0100] In operation S602, the system performs risk detection on the interaction request input by the user based on the preset word list matching rules and the platform filtering rules built in the social data interaction platform.

[0101] In operation S603, it is judged whether there is a risk in the interaction request.

[0102] In operation S604, the interaction request is intercepted.

[0103] In operation S605, it is judged whether to input the interaction request into the large model.

[0104] In operation S606, the output content is filtered according to the risk management and control strategy of the large model to obtain an interactive response.

[0105] In operation S607, risk detection is performed on the interaction response based on the preset vocabulary matching rule and the platform filtering rule built into the social data interaction platform.

[0106] In operation S608 , it is determined whether the interactive response passes the risk detection.

[0107] In operation S609, the interaction response is displayed in the interaction interface.

[0108] In operation S610, the interactive response is intercepted.

[0109] According to an embodiment of the present disclosure, an interaction request input by a user in an interactive interface is received, and a risk detection is performed on the interaction request through a double filtering mechanism. Specifically, risk detection involves two aspects: on the one hand, the content is screened according to preset vocabulary matching rules to identify and filter obvious illegal information. On the other hand, the compliance of the request is verified using the platform filtering rules built into the social data interaction platform. The vocabulary matching rules are mainly used to filter simpler abnormal illegal information, while the platform filtering rules contain nearly a hundred risk control strategies, which review the interaction requests through asynchronous interception, thereby effectively reducing the risk of message abnormalities.

[0110] According to an embodiment of the present disclosure, after completing multi-dimensional risk detection, a risk assessment is performed on the interaction request. If illegal content is identified, the system will immediately execute the operation of intercepting the interaction request. For the interaction request that passes the detection, it is further determined whether it needs to be submitted to the large model for processing. When an interaction request is submitted to the large model, the system will simultaneously activate the risk control strategy within the large model. The strategy begins to adjust during the content generation phase to ensure that the output results meet security standards. At the same time, the interactive response generated by the large model also needs to be double-checked by the vocabulary matching rules and the platform filtering rules again, thus forming a complete content security closed loop.

[0111] According to the embodiments of the present disclosure, the interactive responses that pass the risk detection are delivered to the interactive interface for display, while the interactive responses that fail the detection are intercepted. Specifically, the interception operation may include withdrawing the message or coding the message content. Through the full-link risk management from input to output, the potential risks are effectively controlled while ensuring the freedom of interaction.

[0112] According to an embodiment of the present disclosure, the interaction method further includes: continuously monitoring the activity index of a target functional unit by using a hotspot capture mechanism; analyzing the operation data sets of multiple functional components in the target functional unit to determine the active preference when it is determined that the dynamic gain value of the activity index reaches a preset threshold; and deploying an agent matching the active preference in the target functional unit.

[0113] According to an embodiment of the present disclosure, the activity index of the target functional unit is continuously tracked by using a hotspot capture mechanism. This mechanism adopts a sliding time window algorithm to dynamically collect and analyze data in dimensions such as user access frequency, stay duration, and operation depth. When it is monitored that the dynamic gain value of the activity index exceeds the preset threshold, the operation data sets of multiple functional components in the target functional unit are obtained.

[0114] According to an embodiment of the present disclosure, multi-dimensional analysis is performed on the operation data sets of each functional component in the target functional unit to identify the usage preference patterns of user groups, establish an activity evaluation matrix for functional components, and finally determine the dominant active preference.

[0115] According to an embodiment of the present disclosure, based on the identified active preference, a matching agent is dynamically deployed in the target functional unit. This deployment process adopts a modular architecture to ensure that the functional components of the agent are highly consistent with the user preferences. After the deployment is completed, the system will also continuously collect feedback data to continuously optimize the service strategy of the agent through a reinforcement learning mechanism.

[0116] According to an embodiment of the present disclosure, by continuously monitoring the activity index through a hotspot capture mechanism to analyze the operation data sets to determine the active preference when the dynamic gain value reaches the threshold and deploying a matching agent, the agent in the target functional unit can be closely matched with the current active demand, improving the adaptability of the agent and the operation efficiency of the functional unit, and optimizing the user experience.

[0117] According to an embodiment of the present disclosure, analyzing the operation data sets of multiple functional components in the target functional unit to determine the active preference includes: determining the activity of each functional component in the high-frequency period by using a preset quantization algorithm according to the operation data sets of multiple functional components, where the functional components include at least one of a Q&A component, a voting component, a lottery component, and a scoring component, and the high-frequency period includes the time period when the posting frequency of the target functional unit exceeds the target frequency; determining at least one target component from multiple functional components according to the activity of each functional component; and determining the active preference according to the attribute information of the target component and the topic keywords in the high-frequency period.

[0118] According to an embodiment of the present disclosure, based on a preset quantization algorithm, a periodic analysis is performed on the operation data sets of each functional component (including but not limited to a question-and-answer component, a voting component, a lottery component, and a scoring component) within a target functional unit. This algorithm comprehensively considers dimensional indicators such as component call frequency, user participation depth, and interaction duration, and focuses on calculating the activity values of each component during high-frequency periods. Among them, the determination criterion for the high-frequency period is the time interval during which the posting frequency of the target functional unit continuously exceeds a preset threshold.

[0119] According to an embodiment of the present disclosure, by establishing an activity evaluation matrix for functional components, normalization processing and weighted calculation are performed on the performance indicators of each functional component, so as to obtain the activity of each functional component. Subsequently, a sliding window statistical method is used to identify functional components whose activity is significantly higher than the average level, and one or more target components are determined from the candidate set in combination with a preset selection strategy (such as the top N screening method or the threshold filtering method).

[0120] According to an embodiment of the present disclosure, the attribute characteristics of the target component and the topic keywords during the high-frequency period are integrated to construct a multi-dimensional feature vector space. Through feature weight calculation and pattern matching algorithms, the topic keywords are associated and analyzed with the attribute information, and finally an active preference representing the behavior preferences of the user group is output.

[0121] According to an embodiment of the present disclosure, based on the operation data sets of the functional components, the activity of each functional component during the high-frequency period is determined, and then the target components are screened out. This process can accurately grasp the preference tendency of the functional unit during the active period, provide a strong basis for subsequent targeted deployment of related intelligent agents, and thus effectively improve user participation and experience.

[0122] According to an embodiment of the present disclosure, the interaction method further includes: extracting operation interaction data associated with the operation rules from the target functional unit; using text clustering technology to integrate the operation interaction data into structured guiding information; and presetting the question-and-answer pairs in the guiding information into the response template of the dialogue robot to generate an intelligent agent for guiding an object to obtain operation information in the target functional unit.

[0123] According to an embodiment of the present disclosure, extracting the operation interaction data associated with the operation rules from the target functional unit specifically includes information such as user query records, system response content, and operation logs. After removing noise information through a data cleaning process, the remaining valid data is standardized to unify the data format and coding specification, and a standardized data basis is established for subsequent analysis.

[0124] According to an embodiment of the present disclosure, a text clustering algorithm based on deep learning is adopted to perform multi-dimensional feature extraction and semantic analysis on the preprocessed operation interaction data. The core topic clusters in the data are identified through topic modeling technology, and the logical associations between question-and-answer pairs are established by using the relation extraction method. Finally, guiding information with a hierarchical structure is generated.

[0125] According to an embodiment of the present disclosure, high-quality question-and-answer pairs in the structured guiding information are pre-set into the response template of the dialogue robot, and a multi-turn dialogue process tree is constructed based on the semantic features of the question-and-answer pairs. At the same time, a context association mechanism and a speech optimization strategy are configured, so as to generate an agent that matches the guiding information. The finally generated agent has the ability to dynamically evaluate the user's query intention and can adaptively select the most matching guiding strategy according to the user's cognitive level and operation scenario.

[0126] According to an embodiment of the present disclosure, by generating an agent based on the operation interaction data in the target functional unit, the agent can automatically answer some common questions, provide basic user guidance, etc., thus effectively reducing the work burden of community managers.

[0127] According to an embodiment of the present disclosure, the creation of the agent can also be initiated by the user spontaneously. When a request for creating an agent triggered by the user is received, the creation interface is dynamically loaded. According to the feature information input by the user in the creation interface, the corresponding agent is created. As Figure 7 shown, in the creation interface, the user can input multiple attribute information in the attribute input box 701, and these attribute information include attribute a, attribute b, attribute c, and attribute d. After the user completes the input of all required attributes, by clicking the creation button 702 at the bottom of the interface, the creation process of the agent can be completed. This design allows the user to customize multiple attributes of the agent to meet personalized needs.

[0128] Figure 8 A block diagram of an interaction device according to an embodiment of the present disclosure is schematically shown.

[0129] As Figure 8 shown, the interaction device 800 includes an interface loading module 810 and a request response module 820.

[0130] A selection receiving module 810, configured to receive the selection of the agent by the object in the social data interaction platform.

[0131] An interface display module 820, configured to display an interaction interface corresponding to the agent in response to the selection, wherein the agent is deployed in a target functional unit of the social data interaction platform.

[0132] A request response module 830, configured to input an interaction request entered by an object in an interaction interface and information characterizing the content features of a target functional unit into a large model, and output an interaction response that matches the content features.

[0133] According to an embodiment of the present disclosure, the interaction device 800 further includes a record determination module, a guidance generation module, and a guidance display module.

[0134] The record determination module is configured to determine the interaction record between the object and the intelligent agent in the target functional unit.

[0135] The guidance generation module is configured to dynamically generate guidance information based on the interaction record, where the guidance information is used to guide the object to input an interaction request.

[0136] The guidance display module is configured to display the guidance information when the interaction interface is loaded.

[0137] According to an embodiment of the present disclosure, the guidance generation module includes a guidance determination sub-module and a guidance generation sub-module.

[0138] The guidance determination sub-module is configured to determine that the guidance information is the basic guidance for the intelligent agent when the interaction record represents that the object and the intelligent agent are in a target interaction.

[0139] The guidance generation sub-module is configured to generate guidance information that matches the target functional unit according to the interaction data generated by the historical interaction when the interaction record represents that there is a historical interaction between the object and the intelligent agent.

[0140] According to an embodiment of the present disclosure, the guidance generation sub-module includes a feature extraction unit and an information fusion unit.

[0141] The feature extraction unit is configured to extract the interaction features of the object from the interaction data, where the interaction features include at least one of semantic features, context features, and behavior pattern features.

[0142] The information fusion unit is configured to fuse the interaction features with the real-time dynamic information of the target functional unit to generate guidance information, where the real-time dynamic information includes the real-time behavior data of the object in the target functional unit and the heat change information in the target functional unit.

[0143] According to an embodiment of the present disclosure, the interaction device 800 further includes a history reading module, a usage display module, a response query module, and a query output module.

[0144] The history reading module is configured to call the intelligent agent entry component embedded in the intelligent agent sharing post to read the historical interaction data created by the object sharing the intelligent agent post when it detects that the object triggers an operation by clicking on the intelligent agent sharing post in the social data interaction platform.

[0145] Use a display module to display instructions for using historical interaction data on the interaction interface.

[0146] A response query module is used to perform a response query in the historical interaction information according to the interaction request input by the object when it is determined that the object agrees to use the historical interaction data based on the usage instructions.

[0147] A query output module is used to output the queried interaction response when it is determined that the historical interaction information includes an interaction response corresponding to the interaction request.

[0148] According to an embodiment of the present disclosure, the interaction device 800 further includes a request detection module, a request interception module, a content filtering module, and a response detection module.

[0149] The request detection module is used to perform risk detection on the interaction request input by the object based on a preset word list matching rule and a platform filtering rule built in the social data interaction platform.

[0150] The request interception module is used to intercept the interaction request when it is determined that the interaction request is risky.

[0151] The content filtering module is used to filter the output content according to the risk control strategy of the large model to obtain an interaction response when it is determined that the interaction request passes the risk detection and the interaction request is input to the large model.

[0152] The response detection module is used to display the interaction response on the interaction interface when it is determined that the interaction response passes the risk detection, and the risk detection is based on a preset word list matching rule and a platform filtering rule.

[0153] According to an embodiment of the present disclosure, the interaction device 800 further includes a hot spot capture module, a preference determination module, and an intelligent deployment module.

[0154] The hot spot capture module is used to continuously monitor the activity index of the target functional unit by using a hot spot capture mechanism.

[0155] The preference determination module is used to analyze the operation data sets of multiple functional components in the target functional unit to determine the active preference when it is determined that the dynamic gain value of the activity index reaches a preset threshold.

[0156] The intelligent deployment module is used to deploy an agent matching the active preference in the target functional unit.

[0157] According to an embodiment of the present disclosure, the preference determination module includes an activity determination sub-module, a target determination sub-module, and a preference determination sub-module.

[0158] An active determination sub-module, configured to determine the activity levels of each functional component during high-frequency periods according to the operation data sets of multiple functional components by using a preset quantization algorithm, where the functional components include at least one of a question-and-answer component, a voting component, a lottery component, and a scoring component, and the high-frequency periods include time periods when the posting frequency of the target functional unit exceeds the target frequency.

[0159] A target determination sub-module, configured to determine at least one target component from multiple functional components according to the activity levels of each functional component.

[0160] A preference determination sub-module, configured to determine an active preference according to the attribute information of the target component and the topic keywords during the high-frequency periods.

[0161] According to an embodiment of the present disclosure, the interaction device 800 further includes an operation extraction module, a text clustering module, and an intelligent generation module.

[0162] An operation extraction module, configured to extract operation interaction data associated with operation rules from the target functional unit.

[0163] A text clustering module, configured to integrate the operation interaction data into structured guidance information by using text clustering technology.

[0164] An intelligent generation module, configured to preset the question-and-answer pairs in the guidance information into the response template of the dialogue robot to generate an intelligent agent for guiding an object to obtain the operation information in the target functional unit.

[0165] According to an embodiment of the present disclosure, the present disclosure further provides an electronic device, a readable storage medium, and a computer program product.

[0166] According to an embodiment of the present disclosure, an electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method as described above.

[0167] According to an embodiment of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the method as described above.

[0168] According to an embodiment of the present disclosure, a computer program product includes a computer program, and the computer program, when executed by a processor, implements the method as described above.

[0169] Figure 9A block diagram of an electronic device suitable for implementing an interaction method according to an embodiment of the present disclosure is schematically shown. The electronic device is intended to represent various forms of digital computers, such as, for example, a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as, for example, a personal digital processor, a cellular phone, a smart phone, a wearable device, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0170] As Figure 9 shown, the device 900 includes a computing unit 901 that can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 902 or a computer program loaded from a storage unit 908 into a random access memory (RAM) 903. In the RAM 903, various programs and data required for the operation of the device 900 can also be stored. The computing unit 901, the ROM 902, and the RAM 903 are connected to each other via a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0171] A plurality of components in the device 900 are connected to the I / O interface 905, including: an input unit 906, such as a keyboard, a mouse, etc.; an output unit 907, such as various types of displays, speakers, etc.; a storage unit 908, such as a disk, an optical disc, etc.; and a communication unit 909, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 909 allows the device 900 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0172] The computing unit 901 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 901 executes the various methods and processes described above, such as the interaction method. For example, in some embodiments, the interaction method can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 908. In some embodiments, part or all of the computer program can be loaded and / or installed onto the device 900 via the ROM 902 and / or the communication unit 909. When the computer program is loaded into the RAM 903 and executed by the computing unit 901, one or more steps of the interaction method described above can be executed. Alternatively, in other embodiments, the computing unit 901 can be configured to execute the interaction method by any other suitable means (e.g., by means of firmware).

[0173] The various embodiments of the systems and techniques described above in this document can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a dedicated or general-purpose programmable processor, and can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0174] The program code for implementing the methods of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the program codes are executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program code can be executed entirely on the machine, partially on the machine, as an independent software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0175] In the context of this disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0176] In order to provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, speech input, or tactile input).

[0177] The systems and techniques described herein can be implemented in a computing system including backend components (e.g., as a data server), or a computing system including middleware components (e.g., an application server), or a computing system including frontend components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system including any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0178] A computer system can include a client and a server. The client and the server are generally remote from each other and typically interact through a communication network. The client-server relationship is generated by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, a server of a distributed system, or a server incorporating a blockchain.

[0179] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in the present disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved, and no limitation is imposed herein.

[0180] The above specific embodiments do not constitute a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub - combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

1. An interaction method, comprising: Receiving a selection of an agent by an object in a social data interaction platform; In response to the selection, displaying an interaction interface corresponding to the agent, wherein the agent is deployed in a target functional unit of the social data interaction platform; Inputting an interaction request entered by the object in the interaction interface and information characterizing the content features of the target functional unit into a large model, so as to present an interaction response matching the content features on the interaction interface.

2. The method according to claim 1, wherein, Further comprising: Determining an interaction record between the object and the agent in the target functional unit; Dynamically generating guiding information based on the interaction record, wherein the guiding information is used to guide the object to input an interaction request; Displaying the guiding information when the interaction interface is loaded.

3. The method according to claim 2, wherein, The dynamically generating guiding information based on the interaction record includes: When the interaction record indicates that the interaction between the object and the agent is a target interaction, determining that the guiding information is a basic guide for the agent; When the interaction record indicates that there is a historical interaction between the object and the agent, generating guiding information matching the target functional unit according to the interaction data generated by the historical interaction.

4. The method according to claim 3, wherein, The dynamically generating guiding information matching the target functional unit according to the interaction data generated by the historical interaction includes: Extracting interaction features of the object from the interaction data, wherein the interaction features include at least one of semantic features, context features, and behavior pattern features; Fusing the interaction features with real-time dynamic information of the target functional unit to generate the guiding information, wherein the real-time dynamic information includes real-time behavior data of the object in the target functional unit and heat change information in the target functional unit.

5. The method according to claim 1, wherein Further comprising: When it is detected that the object triggers the operation by clicking on an agent sharing post in the social data interaction platform, invoking an agent entry component embedded in the agent sharing post to read historical interaction data created by the object sharing the agent sharing post; Displaying an instruction for using the historical interaction data on the interaction interface; When it is determined that the object agrees to use the historical interaction data based on the instruction, performing a response query in the historical interaction information according to the interaction request input by the object; When it is determined that the historical interaction information includes an interaction response corresponding to the interaction request, outputting the queried interaction response.

6. The method according to claim 1, wherein, Further comprising: Performing risk detection on the interaction request input by the object based on a preset word list matching rule and a platform filtering rule built in the social data interaction platform; When it is determined that the interaction request has a risk, intercepting the interaction request; When it is determined that the interaction request passes the risk detection and the interaction request is input into the large model, filtering the output content according to the risk control strategy of the large model to obtain the interaction response; In the case where it is determined that the interaction response passes the risk detection, the interaction response is displayed in the interaction interface, and the risk detection is performed based on the preset vocabulary matching rule and the platform filtering rule.

7. The method according to claim 1, wherein, It further includes: Using a hotspot capture mechanism to continuously monitor the activity index of the target functional unit; In the case where it is determined that the dynamic gain value of the activity index reaches a preset threshold, analyzing the operation data sets of multiple functional components in the target functional unit to determine the active preference; Deploying an agent matching the active preference in the target functional unit.

8. The method according to claim 7, wherein, The analyzing the operation data sets of multiple functional components in the target functional unit to determine the active preference includes: According to the operation data sets of the multiple functional components, using a preset quantization algorithm to determine the activity of each functional component in the high-frequency period, where the functional component includes at least one of a question-and-answer component, a voting component, a lottery component, and a scoring component, and the high-frequency period includes a time period when the posting frequency of the target functional unit exceeds the target frequency; Determining at least one target component from the multiple functional components according to the activity of each functional component; Determining the active preference according to the attribute information of the target component and the topic keywords in the high-frequency period.

9. The method according to claim 1, wherein It further includes: Extracting operation interaction data associated with the operation rule from the target functional unit; Using text clustering technology to integrate the operation interaction data into structured guidance information; Pre-setting the question-and-answer pairs in the guidance information into the response template of the dialogue robot to generate an agent for guiding the object to obtain the operation information in the target functional unit.

10. An interaction device, comprising: A selection receiving module, configured to receive the selection of an agent by an object in a social data interaction platform; An interface display module, configured to display an interaction interface corresponding to the agent in response to the selection, where the agent is deployed in a target functional unit of the social data interaction platform; A request response module, configured to input the interaction request input by the object in the interaction interface and the information characterizing the content feature of the target functional unit into a large model, so as to present an interaction response matching the content feature on the interaction interface.

11. An electronic device, comprising: At least one processor; And A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method according to any one of claims 1-9.

12. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to execute the method according to any one of claims 1-9.

13. A computer program product, comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1-9.

Images