Interaction method, interaction device, equipment, medium and program product

Abstract

The invention provides an interaction method which can be applied to the technical fields of Internet technology, electronic technology, artificial intelligence and the like. The interaction method comprises the following steps: in response to a search operation of a user on a search interface, displaying an interaction control corresponding to a target category to which a search keyword belongs; in response to a trigger operation of the user on the interaction control, initiating interaction information for the target category to the user in a dialogue window of the interaction control; under the condition that the user intention is determined to be an article recommendation intention according to the reply of the user, performing user preference interaction with the user, and determining preference information of the user for a target category; and according to the preference information of the user, recommending a target article to the user, the target article being determined from all articles of the target category. The invention further provides an interaction device, equipment, a medium and a program product.

Description

Technical Field

[0001] The present disclosure relates to technical fields such as Internet technology, electronic technology and artificial intelligence, and more specifically to an interaction method, interaction device, equipment, medium and program product. Background Art

[0002] With the rapid development of Internet technology and the popularization of e-commerce, e-commerce platforms have gradually become one of the main channels for consumers to shop. However, e-commerce platforms are also facing the challenge of matching massive information with diverse user needs. The recommendation system of current e-commerce platforms can understand the user's inquiry semantics through preset scripts and keyword matching technology, and provide users with relevant recommendation lists. Users can obtain detailed information about products by browsing the information in the recommendation list or interacting with the customer service system.

[0003] In the process of realizing the concept of the present disclosure, the inventors found that there are at least the following problems in the related technology: although the existing customer service systems can handle some standardized consultations, they often lack the ability to make accurate personalized recommendations, and the customer service systems are often limited to a single merchant and it is difficult to provide services across merchants. Users cannot obtain comprehensive market information, which affects the efficiency and quality of shopping decisions. Summary of the invention

[0004] In view of this, the present disclosure provides an interaction method, an interaction device, a device, a medium, and a program product.

[0005] One aspect of the present disclosure provides an interaction method, including: in response to a user's search operation in a search interface, displaying an interaction control corresponding to a target category to which a search keyword belongs; in response to a user's triggering operation on the interaction control, initiating interaction information for the target category to the user in a dialog window of the interaction control; in a case where it is determined based on the user's reply that the user's intention is an item recommendation intention, performing user preference interaction with the user to determine the user's preference information for the target category; and recommending a target item to the user based on the user's preference information, wherein the target item is determined from all items in the target category.

[0006] According to an embodiment of the present disclosure, user preference interaction is performed with a user to determine the user's preference information for a target category, including: performing user preference interaction with the user based on preference questions under different preference dimensions to determine the user's preference information for the target category under different preference dimensions; wherein the multiple preference dimensions are obtained by processing historical interaction information about items in the target category using a semantic understanding large language model.

[0007] According to an embodiment of the present disclosure, user preference interaction is performed with the user based on preference questions under different preference dimensions to determine the user's preference information for the target category under different preference dimensions, including: utilizing a semantic understanding large language model to process the user's reply to the preference questions of the current round to obtain the user's preference information corresponding to the preference dimensions of the current round; filling the user's preference information corresponding to the preference dimensions of the current round into an information template as a preference slot value, wherein the information template includes information slots to be filled corresponding to a plurality of preference dimensions one by one; determining the preference questions for the next round of interaction with the user according to the filling status of the information template; and obtaining the preference information based on the preference slot values ​​in the filled information template.

[0008] According to an embodiment of the present disclosure, in response to a user's search operation in a search interface, search results corresponding to the search keywords are also displayed in the search interface; target items are recommended to the user based on the user's preference information, including: based on the user's preference information, using a large language model for item recommendation to filter the search results to determine recommended target items; and displaying the recommended target items in a dialog window.

[0009] According to an embodiment of the present disclosure, after recommending a target item to a user based on the user's preference information, the method further includes, in response to receiving supplementary interaction information from the user in a dialogue window, processing the supplementary interaction information using a semantic understanding large language model to obtain a semantic recognition result; and when it is determined that the semantic recognition result is the user's supplementary preference information for the target category, updating the recommended target item based on the preference information and the supplementary preference information.

[0010] According to an embodiment of the present disclosure, the above method also includes: in a case where it is determined that the user's intention is to consult a question based on the user's reply or supplementary interaction information, determining the type of question the user is consulting based on the context information of the user's reply or supplementary interaction information; based on the type of question the user is consulting, using the large language model to answer the question to call the knowledge base corresponding to the question type, and obtaining an answer corresponding to the user's question.

[0011] According to an embodiment of the present disclosure, based on the type of question asked by the user, a large answering language model is used to call the knowledge base corresponding to the question type, and an answer corresponding to the user's question is obtained, including: when it is determined based on the user's reply or supplementary interaction information that the user is asking for general knowledge of the target category, the general knowledge base of the target category is called by the large answering language model to obtain an answer corresponding to the user's question; and when it is determined based on the user's reply or supplementary interaction information that the user is asking for a recommended target item, the dedicated knowledge base of the recommended target item is called by the large answering language model to obtain an answer corresponding to the user's question.

[0012] According to an embodiment of the present disclosure, the above method also includes: when it is determined that the user's intention is to chat based on the user's reply or supplementary interaction information, using the response large language model to process the user's interaction information and output a reply to the interaction information.

[0013] Another aspect of the present disclosure provides an interactive device, including: a search module, for displaying an interactive control corresponding to a target category to which a search keyword belongs in response to a user's search operation in a search interface; an interactive module, for initiating interactive information for the target category to the user in a dialog window of the interactive control in response to a user's triggering operation on the interactive control; a determination module, for performing user preference interaction with the user to determine the user's preference information for the target category when it is determined based on the user's reply that the user's intention is an item recommendation intention; and a recommendation module, for recommending a target item to the user based on the user's preference information, wherein the target item is determined from all items in the target category.

[0014] Another aspect of the present disclosure provides an electronic device, including: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described above.

[0015] Another aspect of the present disclosure provides a computer-readable storage medium storing computer-executable instructions, which are used to implement the above-mentioned method when executed.

[0016] Another aspect of the present disclosure provides a computer program product, which includes computer executable instructions. When the instructions are executed, they are used to implement the method described above.

[0017] According to the embodiments of the present disclosure, because an interactive control corresponding to the target category to which the search keyword belongs is displayed in response to the user's search operation in the search interface, and interactive information for the target category is initiated to the user in the dialog window of the interactive control, since the interactive control interacts with the user based on the target category, the interactive control can provide services across merchants, breaking down the information barriers of different merchants, allowing users to obtain comprehensive market information. Furthermore, in the case where the user's intention is determined to be an item recommendation intention based on the user's reply, user preference interaction is performed with the user to determine the user's preference information for the target category, and the target item is recommended to the user based on the user's preference information, thereby enabling accurate personalized recommendations for the user. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0019] Figure 1 An exemplary system architecture to which the interactive method and apparatus of the present disclosure can be applied is schematically shown;

[0020] Figure 2 A flowchart of an interaction method according to an embodiment of the present disclosure is schematically shown;

[0021] Figure 3 A schematic diagram of displaying interactive controls according to an embodiment of the present disclosure is schematically shown;

[0022] Figure 4 A schematic diagram schematically shows a user interaction process according to an embodiment of the present disclosure;

[0023] Figure 5 A schematic diagram schematically illustrates a user preference interaction according to an embodiment of the present disclosure;

[0024] Figure 6 A schematic diagram schematically shows a multi-round interaction according to another embodiment of the present disclosure;

[0025] Figure 7 The overall flow chart of the interaction method according to another embodiment of the present disclosure is schematically shown;

[0026] Figure 8 The schematic diagram shows an architecture diagram of an integrated multifunctional large language model according to an embodiment of the present disclosure;

[0027] Fig. 9 A block diagram schematically shows an interaction device according to an embodiment of the present disclosure; and

[0028] Fig.10 A block diagram of an electronic device suitable for implementing the above-described interaction method according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0029] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0030] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0031] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0032] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0033] In the embodiments of the present disclosure, the collection, updating, analysis, processing, use, transmission, provision, disclosure, storage, etc. of the data involved (for example, including but not limited to user personal information) are in compliance with the provisions of relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. In particular, necessary measures are taken for user personal information to prevent illegal access to user personal information data and maintain the security of user personal information, network security, and national security.

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

[0035] The inventors found that the current intelligent customer service systems of e-commerce platforms are all set up separately based on a certain store or brand. When users need to consult more detailed information about items on the e-commerce platform, they need to enter the store or brand, trigger the intelligent customer system of the store or brand, and consult the item information related to the store or brand. The interactive information of this intelligent interactive system is limited to a single merchant or brand, and it is difficult to provide services across multiple merchants or brands. Users cannot obtain comprehensive market information, which leads to the problem of information islands and affects the efficiency and quality of user shopping decisions. Furthermore, although the existing intelligent customer service systems can handle some standardized consultations, they often lack the ability to make accurate personalized recommendations and cannot provide customized item recommendations based on the unique needs and preferences of users. And there are still limitations in understanding complex user intentions and contexts, which limits their ability to handle complex consultations and provide in-depth interactions. Although large language model technology has brought new possibilities for intelligent customer service, how to effectively integrate these models into existing e-commerce platforms and achieve good utilization is still a technical challenge.

[0036] An embodiment of the present disclosure provides an interaction method, including: in response to a user's search operation in a search interface, displaying an interaction control corresponding to a target category to which a search keyword belongs; in response to a user's triggering operation on the interaction control, initiating interaction information for the target category to the user in a dialog window of the interaction control; in a case where it is determined based on the user's reply that the user's intention is an item recommendation intention, performing user preference interaction with the user to determine the user's preference information for the target category; and recommending a target item to the user based on the user's preference information, wherein the target item is determined from all items in the target category.

[0037] Figure 1 The exemplary system architecture 100 of the interactive method and device according to the embodiment of the present disclosure is schematically shown. It should be noted that: Figure 1 What is shown is merely an example of a system architecture to which the embodiments of the present disclosure can be applied, in order 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.

[0038] like Figure 1 As shown, the system architecture 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 is used to provide a medium for a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0039] The user may use the first terminal device 101, the second terminal device 102, and the third terminal device 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 first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, and / or social platform software, etc. (only for example).

[0040] The first terminal device 101, the second terminal device 102, and the third terminal device 103 may be various electronic devices having display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

[0041] The server 105 may be a server that provides various services, such as a background management server (only as an example) that provides support for websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0042] It should be noted that the interactive method provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the interactive device provided in the embodiment of the present disclosure can generally be set in the server 105. The interactive method provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Correspondingly, the interactive device provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Alternatively, the interactive method provided in the embodiment of the present disclosure can also be executed by the first terminal device 101, the second terminal device 102 or the third terminal device 103, or can also be executed by other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103. Accordingly, the interaction device provided in the embodiment of the present disclosure may also be arranged in the first terminal device 101, the second terminal device 102 or the third terminal device 103, or in other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103.

[0043] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0044] Figure 2 The flowchart of the interaction method according to the embodiment of the present disclosure is schematically shown.

[0045] like Figure 2 As shown, the method includes operations S201 to S204.

[0046] In operation S201, in response to a search operation of a user on a search interface, an interactive control corresponding to a target category to which the search keyword belongs is displayed.

[0047] In operation S202, in response to a triggering operation of the interactive control by the user, interactive information for a target category is initiated to the user in a dialog window of the interactive control.

[0048] In operation S203, when it is determined that the user's intention is to recommend an item based on the user's reply, user preference interaction is performed with the user to determine the user's preference information for the target category.

[0049] In operation S204, a target item is recommended to the user according to the user's preference information, wherein the target item is determined from all items in the target category.

[0050] According to an embodiment of the present disclosure, in response to a user's search operation on a search interface, an interactive control corresponding to a target category to which the search keyword belongs is displayed. A mapping relationship between the search keyword and the target category may be established first. When a user searches for the target category on a network platform or client using the search keyword, various item information of the target category will be displayed on the search interface based on the mapping relationship between the keyword and the corresponding target category. At the same time, the interactive control corresponding to the target category will also be displayed on the search interface.

[0051] According to an embodiment of the present disclosure, an interactive control represents a control for interacting with a user, and can be suspended on a search interface in the form of an icon. The interactive control can call the item information of the target category corresponding to the search keyword to interact with the user, and the user can obtain the item information of each merchant or brand of the target category through the interactive control, thereby realizing cross-merchant or cross-brand interaction.

[0052] According to an embodiment of the present disclosure, the interactive control can be triggered by clicking on the icon of the interactive control or by triggering operations such as voice wake-up. The triggering operation will switch to the dialogue window of the interactive control. After entering the dialogue window, the interactive control will actively ask the user questions about the target category, while ensuring that the interaction with the user can be carried out smoothly and asking about the user's needs.

[0053] According to an embodiment of the present disclosure, based on the category of the item, an initial question corresponding to the category can be preset, and the initial question can be a question to inquire about the user's intention, for example, to inquire about the user's purchase requirements for the target category. While inquiring, multiple candidate answers can also be provided, and the user can reply by customizing the reply or clicking on multiple candidate answers, and the user's intention is determined based on the user's reply.

[0054] According to the embodiments of the present disclosure, when the user's intention is determined to be an item recommendation intention based on the user's reply, a series of targeted questions can be designed to interact with the user's user preferences based on the target category, and the user's preference information for the target category can be understood based on the user's answers and feedback. Based on the collected preference information, all items in the target category are screened, items that do not meet the user's preference information are excluded, and one or more recommended target items that meet the user's preference information are screened and recommended to the user.

[0055] According to the embodiments of the present disclosure, because an interactive control corresponding to the target category to which the search keyword belongs is displayed in response to the user's search operation in the search interface, and interactive information for the target category is initiated to the user in the dialog window of the interactive control, since the interactive control interacts with the user based on the target category, the interactive control can provide services across merchants, breaking down the information barriers of different merchants, allowing users to obtain comprehensive market information. Furthermore, in the case where the user's intention is determined to be an item recommendation intention based on the user's reply, user preference interaction is performed with the user to determine the user's preference information for the target category, and the target item is recommended to the user based on the user's preference information, thereby enabling accurate personalized recommendations for the user.

[0056] Reference below Figure 3~Figure 8 , combined with specific embodiments Figure 2 The method shown is further explained.

[0057] Figure 3 A schematic diagram of displaying interactive controls according to an embodiment of the present disclosure is schematically shown.

[0058] like Figure 3 As shown, when the user searches for a refrigerator in the search interface 301, multiple search results 302 such as refrigerator A, refrigerator B, refrigerator C, refrigerator D, etc. may be displayed in the search interface. At the same time, an interactive control icon 303 may also be displayed in the search interface. By clicking the interactive control icon 303, the interactive control 303 may be triggered to initiate interactive information with the user regarding the refrigerator.

[0059] According to an embodiment of the present disclosure, in response to a user triggering an interactive control, a question is asked to the user on the dialogue page of the interactive control. A question template may be preset first, and the preset question template may be filled with corresponding content to obtain an initial question according to the category of items corresponding to the search keyword. The user is asked a question in the dialogue window, and the user's intention is preliminarily determined based on the user's answer to the initial question. Different dialogue operations are performed according to different user intentions.

[0060] According to an embodiment of the present disclosure, when it is determined that the user intention is to recommend an item based on the user's reply, user preference interaction is performed with the user to determine the user's preference information for the target category, including: performing user preference interaction with the user based on preference questions under different preference dimensions to determine the user's preference information for the target category under different preference dimensions; wherein the multiple preference dimensions are obtained by processing historical interaction information about items in the target category using a semantic understanding large language model.

[0061] According to an embodiment of the present disclosure, the preference dimension represents the dimension that the user focuses on for the target category, for example, the brand, function, price, size and other dimensions of the item. Based on the different categories of items, the dimensions that the user focuses on will also be different. For example, in the category of washing machines, the user may focus on several dimensions such as "brand, price, impeller / drum, and washing capacity". For the milk category, the user may focus on several dimensions such as "brand, price, taste, and nutritional ingredients". According to the category of each item, determine the preference dimensions that the user focuses on for each category.

[0062] According to an embodiment of the present disclosure, a semantic understanding big language model is deployed in the interactive control, and the semantic understanding big language model is used to process the historical interaction information about items in the target category. The advanced semantic understanding ability of the big language model can be used to extract several dimensions that the user focuses on for the target category in the historical interaction information, thereby determining multiple preference dimensions.

[0063] According to an embodiment of the present disclosure, user preference interaction is performed with the user based on preference questions under different preference dimensions to determine the user's preference information for the target category under different preference dimensions, including: asking the user questions in a dialogue window regarding the preference questions under multiple preference dimensions; and using a semantic understanding large language model to process the user's responses to the preference questions to obtain the user's respective preference information for the target category in the multiple preference dimensions.

[0064] According to the embodiments of the present disclosure, by using the semantic understanding large language model to process the historical interaction information about the items in the target category, it is possible to more deeply understand the user's behavior and preferences, thereby refining multiple specific preference dimensions, which helps to more accurately capture the user's diverse needs under the target category. Designing preference questions based on different preference dimensions and interacting with users can provide users with a more personalized experience.

[0065] Figure 4 A schematic diagram schematically shows a user interaction process according to an embodiment of the present disclosure.

[0066] like Figure 4 As shown, the method includes: operations S401~S404.

[0067] In operation S401, a semantic understanding large language model is used to process the user's reply to the current round preference question to obtain the user's preference information corresponding to the current round preference dimension.

[0068] In operation S402, the user's preference information corresponding to the current round preference dimension is filled into the information template as a preference slot value, wherein the information template includes information slots to be filled corresponding to multiple preference dimensions one by one.

[0069] In operation S403, the preference questions for the next round of interaction with the user are determined according to the filling status of the information template.

[0070] In operation S404, preference information is obtained based on the preference slot value in the filled information template.

[0071] According to the embodiments of the present disclosure, the first round of preference questions can be randomly set based on one or more dimensions, and the first round of preference questions can be asked to the user in the dialogue window of the interactive control. The user's response to the first round of preference questions is processed using the semantic understanding large language model to obtain the preference information corresponding to the preference dimension in the first round of preference questions.

[0072] According to an embodiment of the present disclosure, the preference information of the first round is filled into the information template as the filling slot value, and the filling status of the information template is detected. When there is a vacancy in the information slot of the information template, it means that the preference information of the preference dimension corresponding to the vacant information slot has not been collected. At this time, based on the preference questions under the dimension corresponding to the vacant information slot, the user is asked a second round of preference questions in the dialogue window of the interactive control, and the semantic understanding large language model is used to process the user's response to the second round of preference questions and perform the corresponding information template filling operation, and the above operation is repeated until the information template is completely filled. Collect the user's preference information under multiple preference dimensions through multiple rounds of preference questions.

[0073] According to the embodiments of the present disclosure, by gradually asking the user questions about different preference dimensions and filling the user's answers into the information template in real time, the system can gradually understand the user's complete preferences in a structured manner. By gradually filling in the information template, the system can ensure that the user's complete preference information about all preference dimensions is collected. This makes the user interaction process more orderly and efficient, and avoids information confusion and omissions.

[0074] According to the embodiments of the present disclosure, the historical interaction information about the items in the target category is processed by using the semantic understanding large language model, and the importance ranking of different preference dimensions can also be obtained. Questions are asked to the user in sequence according to the importance ranking of the preference dimensions, and the user's preference information under different preference dimensions is identified and collected according to the user's reply.

[0075] Figure 5 A schematic diagram of user preference interaction according to an embodiment of the present disclosure is schematically shown.

[0076] like Figure 5 As shown in the figure, when the user triggers the interactive control of the mobile phone category, the interactive control initiates two rounds of questions to the user on the dialogue page, obtains the user's preference information on the three preference dimensions of price, brand and model of the mobile phone, and recommends the target item to the user based on the preference information of these three preference dimensions.

[0077] According to an embodiment of the present disclosure, in response to a user's search operation on a search interface, search results corresponding to the search keyword are also displayed on the search interface.

[0078] According to an embodiment of the present disclosure, a large language model for item recommendation is also deployed in the interactive control.

[0079] According to an embodiment of the present disclosure, a target item is recommended to a user based on the user's preference information, including: based on the user's preference information, using a large language model for item recommendation to filter search results and determine recommended target items; and displaying the recommended target items in a dialog window.

[0080] According to an embodiment of the present disclosure, the user's preference information and the search results corresponding to the search keywords are input into the item recommendation large language model. The item recommendation large language model can filter the search results based on the user's preference information and score or sort the filtered results, and select one or more items with high rankings as recommended target items based on the sorting or scoring results.

[0081] According to an embodiment of the present disclosure, displaying the recommended target item in the dialog window may include displaying the item description information and the item's network link. When there are multiple recommended target items, they may be displayed in the dialog window in the order of ranking so that the user can directly view and select.

[0082] According to the embodiments of the present disclosure, the semantic understanding big language model, combined with natural language processing capabilities, can understand the nuances and personalized needs of users. Then, the item recommendation big language model can accurately filter the search results according to the user's preference information, so as to provide personalized recommendations for users.

[0083] According to an embodiment of the present disclosure, after recommending a target item to a user according to the user's preference information, the method further includes, in response to receiving the user's supplementary interaction information in the dialogue window, processing the supplementary interaction information using a semantic understanding large language model to obtain a semantic recognition result. If it is determined that the semantic recognition result is the user's supplementary preference information for the target category, the recommended target item is updated based on the preference information and the supplementary preference information.

[0084] According to an embodiment of the present disclosure, preference information, supplementary preference information, and search results corresponding to search keywords are input into a large language model for item recommendation. The large language model for item recommendation performs secondary screening of the search results based on the user's preference information and supplementary preference information, and scores or sorts the secondary screening results. According to the sorting or scoring results, the recommended target items are updated.

[0085] Figure 6A schematic diagram of multiple rounds of interaction according to another embodiment of the present disclosure is schematically shown.

[0086] like Figure 6 As shown, in Figure 5 On the basis shown in the figure, after receiving the recommendation information, the user continued to raise the supplementary preference information of "Is there any suitable for the elderly?" Based on the preference information of the preference dimension of suitable for the elderly and the preference information of the user on the three preference dimensions of price, brand and model of mobile phones obtained in the first two rounds, the target items were re-recommended to the user.

[0087] According to the embodiments of the present disclosure, since different users may have deviations in the dimensions of interest of items, in order to meet the more refined needs of users, after recommending target items to users, by further receiving the user's supplementary preference information, the advanced semantic understanding and dialogue management capabilities of the semantic understanding big semantic model can better understand the complex user intentions and contexts, so that the interactive controls can conduct multiple rounds of dialogues, thereby improving the interactive capabilities, and continuously updating the recommended target items according to the user's refined needs, so that the final recommendation results are more in line with the user's needs.

[0088] According to an embodiment of the present disclosure, a large response language model is also deployed in the interactive control.

[0089] According to an embodiment of the present disclosure, when the user's intention is determined to be a question consultation intention based on the user's reply or supplementary interaction information, the type of question the user consults is determined based on the context information of the user's reply or supplementary interaction information. According to the type of question the user consults, the knowledge base corresponding to the question type is called using the answering large language model, and the answer corresponding to the user's question is obtained.

[0090] According to the embodiments of the present disclosure, the recognition of user intent can occur throughout the entire process of interaction with the user. The semantic understanding large language model is used to identify the user's intent based on the user input content. When the user's intent is identified as a question consultation, the knowledge response process is triggered. For example, asking for the attribute information of a recommended target item, asking for the policy information of the merchant corresponding to a certain item, etc. are all within the scope of knowledge response. At this time, the response large language model is used to conduct a corresponding response dialogue with the user.

[0091] According to an embodiment of the present disclosure, question types may include general knowledge types and special knowledge types. General knowledge types represent general knowledge of a certain category of items, such as what is a drum washing machine, what does the upper drainage of a washing machine mean, etc. Special knowledge types represent knowledge that a certain item has specifically, such as configuration information and function information of a target recommended item.

[0092] According to an embodiment of the present disclosure, by identifying the user's intention and question type and using a large language model to call the corresponding knowledge base, the interactive control can more accurately understand and answer the user's questions, thereby improving the accuracy of question and answer.

[0093] According to an embodiment of the present disclosure, according to the type of question consulted by the user, the knowledge base corresponding to the question type is called by using the answering large language model, and the answer corresponding to the user's question is obtained, including: when it is determined according to the user's reply or supplementary interaction information that the user is consulting the general knowledge of the target category, the general knowledge base of the target category is called by using the answering large language model to obtain the answer corresponding to the user's question. When it is determined according to the user's reply or supplementary interaction information that the user is consulting the recommended target item, the dedicated knowledge base of the recommended target item is called by using the answering large language model to obtain the answer corresponding to the user's question.

[0094] According to the embodiments of the present disclosure, a semantic understanding large language model is used to identify the context information of the user's reply or supplementary interactive information. If the SKU (Stock Keeping Unit) of the item is identified from the context information, it can be further determined that the user is consulting for specialized knowledge of a specific item. For example, when a recommended target item is sent to a user in a dialog window, and the user asks about the size, function, and other information of the recommended target item, since the description information and link information of the recommended target item are sent to the user, the SKU of the recommended target item can be identified through the context information, and the specialized knowledge base of the recommended target item can be called based on the SKU of the item to obtain the configuration information of the recommended target item, thereby using the large language model for answering the user's question.

[0095] If the SKU of the item is not identified from the context information, it can be further determined that the user is asking for general knowledge about a certain category. The general knowledge base of the category can be called and the large language model for answering the user's question can be used.

[0096] According to the embodiments of the present disclosure, the interactive control can accurately call the corresponding knowledge base according to the type of question the user asks. Whether the user asks for general knowledge or professional knowledge of a specific item, the user can get accurate answers provided by the system, thereby improving user experience and satisfaction. The user does not need to search for answers in multiple channels or platforms, which improves the efficiency of consultation.

[0097] According to an embodiment of the present disclosure, when it is determined that the user's intention is to chat based on the user's reply or supplementary interaction information, the user's interaction information is processed using a large response language model, and a reply to the interaction information is output.

[0098] According to the embodiments of the present disclosure, the responses to users' small talk can be generated directly using a large language model, or by formulating a small talk knowledge base and letting the large language model select small talk answers based on the knowledge base.

[0099] According to an embodiment of the present disclosure, when the model recognizes the user's conversation as a small talk intention, the large response language model is used to directly perform the small talk response, which can improve the user experience while ensuring the integrity of the conversation.

[0100] Figure 7 The overall flow chart of the interaction method according to another embodiment of the present disclosure is schematically shown.

[0101] like Figure 7 As shown, the interaction method includes operations S701 to S715.

[0102] In operation S701, a target category is determined according to the search keyword.

[0103] In operation S702, the interactive control is triggered to perform information interaction with the user for the target category.

[0104] In operation S703, the user intention is recognized.

[0105] In operation S704, it is determined that the intention is to recommend an item, and operation S705 is executed. In operation S705, the user's preference information is identified and the preference information is filled into the information slot. In operation S706, the user is asked a question to obtain new preference information, and the obtained preference information is refilled into the information slot. In operation S707, the item recommendation large language model is used to make a recommendation. In operation S708, the recommended target item is obtained.

[0106] In operation S709, it is determined that the intention is to chat, and operation S710 is performed. In operation S710, a chat response is performed with the user.

[0107] In operation S711, it is determined that the intention is to ask a question, and operation S712 is executed. In operation S712, the SKU of the item is identified. When the SKU of the item can be identified, operation S713 is executed. In operation S713, a dedicated knowledge base is called to respond to the user. When the SKU of the item is not identified, operation S714 is executed. In operation S714, a general knowledge base is called to respond to the user. In operation S715, the answer to the question is obtained.

[0108] Figure 8 The schematic diagram shows an architecture diagram of an integrated multifunctional large language model according to an embodiment of the present disclosure.

[0109] like Figure 8As shown, the interactive control of the present invention adopts an integrated large language model deployment method, that is, combined with the multi-task capability of the large language model, the tasks with similar functions in the interactive control are merged and supported by the same large language model service. Specifically, the interactive control integrates a semantic understanding large language model, a response large language model, and an item recommendation large language model. User intent classification, user preference information identification, information slot filling, and item SKU identification are all functions that require semantic understanding of user questions, and are uniformly performed using a semantic understanding large language model. For functions in the interactive control such as user chatting, general knowledge responses, and special knowledge responses that require the use of natural language generation technology to generate answers, a unified response large language model is used to execute. The function of item recommendation is executed using the item recommendation large language model.

[0110] According to the embodiments of the present disclosure, the deployment and operation of large language models are optimized through integrated design, the demand for computing resources is reduced, the deployment resource consumption of large language models is reduced, and the application efficiency of large language model services is improved, thereby reducing operating costs.

[0111] Fig. 9 The block diagram schematically shows an interaction device according to an embodiment of the present disclosure.

[0112] like Fig. 9 As shown, the interaction device includes a search module 910 , an interaction module 920 , a determination module 930 and a recommendation module 940 .

[0113] The search module 910 is used to display an interactive control corresponding to the target category to which the search keyword belongs in response to a user's search operation in the search interface.

[0114] The interaction module 920 is used to initiate interaction information for a target category to the user in the dialog window of the interaction control in response to the user's triggering operation on the interaction control.

[0115] The determination module 930 is used to interact with the user on user preferences and determine the user's preference information for the target category when it is determined that the user's intention is to recommend items based on the user's reply.

[0116] The recommendation module 940 is used to recommend a target item to the user according to the user's preference information, wherein the target item is determined from all items in the target category.

[0117] According to an embodiment of the present disclosure, the determination module 930 includes a preference information determination submodule.

[0118] The preference information determination submodule is used to interact with users based on preference questions under different preference dimensions, and determine the user's preference information for the target category under different preference dimensions; wherein, multiple preference dimensions are obtained by using a semantic understanding large language model to process historical interaction information about items in the target category.

[0119] According to an embodiment of the present disclosure, the preference information determination submodule includes a first processing unit, a filling unit, a preference question determination unit and a preference information determination unit.

[0120] The first processing unit is used to process the user's reply to the current round preference question by using the semantic understanding large language model to obtain the user's preference information corresponding to the current round preference dimension.

[0121] A filling unit is used to fill the user's preference information corresponding to the current round preference dimension as a preference slot value into the information template, wherein the information template includes information slots to be filled corresponding to multiple preference dimensions one by one.

[0122] The preference question determination unit is used to determine the preference questions for the next round of interaction with the user according to the filling status of the information template.

[0123] The preference information determining unit is used to obtain the preference information based on the preference slot value in the filled information template.

[0124] According to an embodiment of the present disclosure, the recommendation module 940 includes a screening submodule and a sending submodule.

[0125] The filtering submodule is used to filter the search results based on the user's preference information and use the item recommendation large language model to determine the recommended target items.

[0126] The sending submodule is used to display the recommended target items in the dialogue window.

[0127] According to an embodiment of the present disclosure, the interaction device further includes a second processing unit and an updating unit.

[0128] The second processing unit, in response to receiving the user's supplementary interaction information in the dialogue window, processes the supplementary interaction information using the semantic understanding large language model to obtain a semantic recognition result.

[0129] The updating unit is used to update the recommended target item based on the preference information and the supplementary preference information when determining that the semantic recognition result is the user's supplementary preference information for the target category.

[0130] According to an embodiment of the present disclosure, the interaction device further includes an identification module and a response module.

[0131] The identification module is used to determine the type of question the user is asking based on the context information of the user's reply or supplementary interaction information when the user's intention is determined to be question-asking intention based on the user's reply or supplementary interaction information.

[0132] The answering module is used to call the knowledge base corresponding to the question type using the answering language model according to the question type asked by the user, and obtain the answer corresponding to the user's question.

[0133] According to an embodiment of the present disclosure, the response module includes a first response submodule and a second response submodule.

[0134] The first response submodule is used to use the response large language model to call the general knowledge base of the target category to obtain an answer corresponding to the user's question when it is determined based on the user's reply or supplementary interaction information that the user is consulting the general knowledge of the target category.

[0135] The second response submodule is used to use the response large language model to call the dedicated knowledge base of the recommended target item to obtain an answer corresponding to the user's question when it is determined based on the user's reply or supplementary interaction information that the user is asking for advice on the recommended target item.

[0136] According to an embodiment of the present disclosure, the response module further includes a third response submodule.

[0137] The third response submodule is used to use the response large language model to process the user's interaction information and output a response to the interaction information when it is determined that the user's intention is to chat based on the user's reply or supplementary interaction information.

[0138] According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits, or at least part of the functions of any one of them can be implemented in one module. According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits can be split into multiple modules for implementation. According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits can be at least partially implemented as hardware circuits, such as field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), systems on chips, systems on substrates, systems on packages, application specific integrated circuits (ASICs), or can be implemented by hardware or firmware in any other reasonable way of integrating or packaging circuits, or can be implemented in any one of the three implementation methods of software, hardware, and firmware, or in any appropriate combination of any of them. Alternatively, according to the embodiments of the present invention, one or more of the modules, submodules, units, and subunits can be at least partially implemented as computer program modules, and when the computer program modules are run, the corresponding functions can be performed.

[0139] For example, any of the search module 910, the interaction module 920, the determination module 930, and the recommendation module 940 can be combined into one module / unit / sub-unit for implementation, or any of the modules / units / sub-units can be split into multiple modules / units / sub-units. Alternatively, at least part of the functions of one or more of the modules / units / sub-units can be combined with at least part of the functions of other modules / units / sub-units and implemented in one module / unit / sub-unit.

[0140] According to an embodiment of the present disclosure, at least one of the search module 910, the interaction module 920, the determination module 930, and the recommendation module 940 may be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application-specific integrated circuit (ASIC), or may be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware, and firmware, or in any appropriate combination of any of them. Alternatively, at least one of the search module 910, the interaction module 920, the determination module 930, and the recommendation module 940 may be at least partially implemented as a computer program module, and when the computer program module is executed, the corresponding function may be performed.

[0141] It should be noted that the interactive device part in the embodiments of the present disclosure corresponds to the interactive method part in the embodiments of the present disclosure. The description of the interactive device part specifically refers to the interactive method part, which will not be repeated here.

[0142] Fig.10 A block diagram of an electronic device suitable for implementing the above-described interaction method according to an embodiment of the present disclosure is schematically shown.

[0143] Fig.10 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0144] like Fig.10As shown, the electronic device 1000 according to an embodiment of the present disclosure includes a processor 1001, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage part 1008 to a random access memory (RAM) 1003. The processor 1001 may include, for example, a general-purpose microprocessor (such as a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (for example, an application-specific integrated circuit (ASIC)), etc. The processor 1001 may also include an onboard memory for caching purposes. The processor 1001 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0145] In RAM 1003, various programs and data required for the operation of electronic device 1000 are stored. Processor 1001, ROM 1002 and RAM 1003 are connected to each other via bus 1004. Processor 1001 performs various operations of the method flow according to the embodiment of the present disclosure by executing the program in ROM 1002 and / or RAM 1003. It should be noted that the program can also be stored in one or more memories other than ROM 1002 and RAM 1003. Processor 1001 can also perform various operations of the method flow according to the embodiment of the present disclosure by executing the program stored in one or more memories.

[0146] According to an embodiment of the present disclosure, the electronic device 1000 may further include an input / output (I / O) interface 1005, which is also connected to the bus 1004. The electronic device 1000 may further include one or more of the following components connected to the input / output (I / O) interface 1005: an input portion 1006 including a keyboard, a mouse, etc.; an output portion 1007 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 1008 including a hard disk, etc.; and a communication portion 1009 including a network interface card such as a LAN card, a modem, etc. The communication portion 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the input / output (I / O) interface 1005 as needed. A removable medium 1011, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1010 as needed, so that a computer program read therefrom is installed into the storage portion 1008 as needed.

[0147] According to an embodiment of the present disclosure, the method flow according to an embodiment of the present disclosure can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program contains a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 1009, and / or installed from the removable medium 1011. When the computer program is executed by the processor 1001, the above-mentioned functions defined in the system of the embodiment of the present disclosure are executed. According to an embodiment of the present disclosure, the system, equipment, device, module, unit, etc. described above can be implemented by a computer program module.

[0148] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present disclosure is implemented.

[0149] According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium. For example, it may include, but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0150] For example, according to an embodiment of the present disclosure, the computer-readable storage medium may include the ROM 1002 and / or the RAM 1003 described above and / or one or more memories other than the ROM 1002 and the RAM 1003 .

[0151] An embodiment of the present disclosure also includes a computer program product, which includes a computer program, and the computer program contains a program code for executing the method provided by the embodiment of the present disclosure. When the computer program product runs on an electronic device, the program code is used to enable the electronic device to implement the interactive method provided by the embodiment of the present disclosure.

[0152] When the computer program is executed by the processor 1001, the above functions defined in the system / device of the embodiment of the present disclosure are executed. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0153] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 1009, and / or installed from the removable medium 1011. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0154] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).

[0155] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions. It can be understood by those skilled in the art that the features recorded in the various embodiments of the present disclosure can be combined and / or combined in a variety of ways, even if such a combination or combination is not explicitly recorded in the present disclosure. In particular, without departing from the spirit and teaching of the present disclosure, the features described in the various embodiments of the present disclosure may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present disclosure.

[0156] The embodiments of the present disclosure are described above. However, these embodiments are only for illustrative purposes and are not intended to limit the scope of the present disclosure. Although the embodiments are described above, this does not mean that the measures in the various embodiments cannot be used in combination to advantage. Without departing from the scope of the present disclosure, those skilled in the art may make a variety of substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. An interactive method, comprising: In response to a user's search operation on the search interface, an interactive control corresponding to the target category to which the search keyword belongs is displayed; In response to the user's triggering operation on the interactive control, initiating interactive information for the target category to the user in the dialog window of the interactive control; When it is determined that the user's intention is to recommend an item according to the user's reply, performing user preference interaction with the user to determine the user's preference information for the target category; According to the preference information of the user, a target item is recommended to the user, wherein the target item is determined from all items in the target category.

2. The method according to claim 1, wherein: The performing user preference interaction with the user to determine the user's preference information for the target category includes: Performing user preference interaction with the user based on preference questions under different preference dimensions to determine the user's preference information for the target category under different preference dimensions; Among them, the multiple preference dimensions are obtained by processing the historical interaction information about the items in the target category using a semantic understanding large language model.

3. The method according to claim 2, wherein: The performing user preference interaction with the user based on preference questions under different preference dimensions to determine the user's preference information for the target category under the different preference dimensions includes: Using the semantic understanding large language model to process the user's response to the current round preference question, obtaining the user's preference information corresponding to the current round preference dimension; Filling the user's preference information corresponding to the current round preference dimension into the information template as a preference slot value, wherein the information template includes information slots to be filled corresponding to a plurality of the preference dimensions one by one; Determining the preference questions for the next round of interaction with the user according to the filling status of the information template; and The preference information is obtained based on the preference slot value in the filled information template.

4. The method according to claim 1, wherein: In response to a search operation by a user on a search interface, search results corresponding to the search keyword are also displayed on the search interface; The recommending a target item to the user according to the preference information of the user includes: Based on the preference information of the user, the search results are screened using a large language model for item recommendation to determine the recommended target item; as well as The recommended target item is displayed in the dialog window.

5. The method according to claim 1, wherein: After recommending the target item to the user according to the preference information of the user, the method further includes: In response to receiving the supplementary interaction information of the user in the dialog window, processing the supplementary interaction information using a semantic understanding large language model to obtain a semantic recognition result; In a case where it is determined that the semantic recognition result is the user's supplementary preference information for the target category, the recommended target item is updated based on the preference information and the supplementary preference information.

6. The method according to claim 5, wherein: The method further comprises: In the case where the user's intention is determined to be a question consultation intention according to the user's reply or the supplementary interaction information, determining the type of question consulted by the user based on the context information of the user's reply or the supplementary interaction information; According to the type of question asked by the user, the knowledge base corresponding to the question type is called using the large answering language model, and the answer corresponding to the user's question is obtained.

7. The method according to claim 6, wherein: According to the type of question asked by the user, the knowledge base corresponding to the question type is called by using the answering language model, and the answer corresponding to the user's question is obtained, including: When it is determined according to the user's reply or supplementary interaction information that the user is consulting general knowledge of the target category, the general knowledge base of the target category is called by using the large answer language model to obtain an answer corresponding to the user's question; and When it is determined based on the user's reply or supplementary interaction information that the user is asking a question about the recommended target item, a dedicated knowledge base of the recommended target item is called using the response large language model to obtain an answer corresponding to the user's question.

8. The method according to claim 6, wherein: The method further comprises: When it is determined that the user's intention is to chat based on the user's reply or supplementary interaction information, the user's interaction information is processed using the response large language model, and a reply to the interaction information is output.

9. An interactive device, comprising: A search module, for displaying an interactive control corresponding to a target category to which the search keyword belongs in response to a user's search operation on the search interface; An interaction module, configured to initiate interaction information for the target category to the user in a dialog window of the interaction control in response to a trigger operation of the user on the interaction control; A determination module, configured to, when it is determined according to the user's reply that the user's intention is to recommend an item, interact with the user about user preferences to determine the user's preference information for the target category; as well as A recommendation module is used to recommend a target item to the user according to the preference information of the user, wherein the target item is determined from all items in the target category.

10. An electronic device, comprising: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 8.

11. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enables the processor to implement the method according to any one of claims 1 to 8.

12. A computer program product, comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 8.

Images