Object recommendation method and device, electronic equipment and readable storage medium

Abstract

The invention relates to the technical field of object recommendation, and provides an object recommendation method and device, electronic equipment and a readable storage medium. The method comprises the steps that user side data and attribute data corresponding to a to-be-recommended object are processed through a large language model, and user preference data and extended data corresponding to the to-be-recommended object are obtained; respectively carrying out text coding processing on the user preference vector and the to-be-recommended object to obtain a user preference vector and an extension vector corresponding to the to-be-recommended object; respectively processing the user preference vector and the extension vector based on a gating mechanism to obtain a gating vector corresponding to the user preference vector and a gating vector corresponding to the extension vector; performing spatial alignment processing on the two gating vectors to obtain a target user preference vector and a target extension vector; determining a recommendation score corresponding to the object to be recommended according to the data and the vector; and based on the preset recommendation score threshold and the recommendation score corresponding to the to-be-recommended object, determining and sending the target object data, thereby improving the accuracy of the recommendation result.

Description

Technical Field

[0001] The present disclosure relates to the technical field of object recommendation, and particularly to an object recommendation method, apparatus, electronic device, and readable storage medium. Background Art

[0002] Object recommendation methods are widely used in multiple fields such as e-commerce platforms, social media, and content distribution networks. Traditional object recommendation methods mainly rely on users' historical behavior data, the attribute characteristics of objects to be recommended, and simple matching algorithms to predict the degree of preference of users for specific objects and make recommendations accordingly. However, when processing user-side data and data of objects to be recommended, traditional recommendation methods only perform shallow feature extraction, resulting in insufficient accuracy and personalization of recommendation results. When matching user preferences and the characteristics of objects to be recommended, the conventional method is to perform fixed weight allocation or simple similarity calculation, lacking dynamic adjustment and adaptive mechanisms. To address the above problems, in the prior art, it is common to introduce more complex feature engineering to extract deep features in user-side data and data of objects to be recommended; use machine learning algorithms to improve the accuracy and generalization ability of the recommendation model. However, complex feature engineering and machine learning algorithms will increase the computational complexity and model training cost, and lack effective dynamic adjustment and adaptive mechanisms when processing user preferences and the characteristics of objects to be recommended.

[0003] Therefore, there is a problem in the prior art that due to the insufficient depth of processing of user-side data and data of objects to be recommended and the lack of dynamic adjustment and adaptive mechanisms, the accuracy and personalization of recommendation results are low. Summary of the Invention

[0004] In view of this, embodiments of the present disclosure provide an object recommendation method, apparatus, electronic device, and readable storage medium to solve the problem in the prior art that due to the insufficient depth of processing of user-side data and data of objects to be recommended and the lack of dynamic adjustment and adaptive mechanisms, the accuracy and personalization of recommendation results are low.

[0005] In the first aspect of the embodiments of the present disclosure, an object recommendation method is provided, including: processing user-side data and attribute data corresponding to the object to be recommended respectively through a large language model to obtain user preference data and extended data corresponding to the object to be recommended; performing text encoding processing on the user preference data and the extended data corresponding to the object to be recommended respectively to obtain a user preference vector and an extended vector corresponding to the object to be recommended; processing the user preference vector and the extended vector corresponding to the object to be recommended respectively based on a gating mechanism to obtain a gating vector corresponding to the user preference vector and a gating vector corresponding to the extended vector; performing spatial alignment processing on the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector respectively to obtain a target user preference vector and a target extended vector corresponding to the object to be recommended; determining a recommendation score corresponding to the object to be recommended according to the user-side data, the attribute data corresponding to the object to be recommended, the historical interaction context data corresponding to the object to be recommended, the target user preference vector, and the target extended vector corresponding to the object to be recommended; determining target object data based on a preset recommendation score threshold and the recommendation score corresponding to the object to be recommended, and sending the target object data to a target terminal device so that the target object data is displayed on the current graphical user interface of the target terminal device.

[0006] In the second aspect of the embodiments of the present disclosure, an object recommendation device is provided, including: a first processing module for processing user-side data and attribute data corresponding to the object to be recommended respectively through a large language model to obtain user preference data and extended data corresponding to the object to be recommended; a second processing module for performing text encoding processing on the user preference data and the extended data corresponding to the object to be recommended respectively to obtain a user preference vector and an extended vector corresponding to the object to be recommended; a third processing module for processing the user preference vector and the extended vector corresponding to the object to be recommended respectively based on a gating mechanism to obtain a gating vector corresponding to the user preference vector and a gating vector corresponding to the extended vector; a fourth processing module for performing spatial alignment processing on the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector respectively to obtain a target user preference vector and a target extended vector corresponding to the object to be recommended; a determination module for determining a recommendation score corresponding to the object to be recommended according to the user-side data, the attribute data corresponding to the object to be recommended, the historical interaction context data corresponding to the object to be recommended, the target user preference vector, and the target extended vector corresponding to the object to be recommended; a fifth processing module for determining target object data based on a preset recommendation score threshold and the recommendation score corresponding to the object to be recommended, and sending the target object data to a target terminal device so that the target object data is displayed on the current graphical user interface of the target terminal device.

[0007] In a third aspect of the embodiments of the present disclosure, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above method are implemented.

[0008] In a fourth aspect of the embodiments of the present disclosure, a readable storage medium is provided. The readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0009] The beneficial effects of the embodiments of the present disclosure compared with the prior art are as follows: By performing text generation processing on the user-side data and the attribute data corresponding to the object to be recommended respectively through a large language model, user preference data containing user preference information can be obtained by processing the user-side data, and extended data corresponding to the object to be recommended containing information related to the object to be recommended can be obtained by processing the attribute data corresponding to the object to be recommended. Furthermore, text encoding processing can be performed on the user preference data and the extended data corresponding to the object to be recommended to obtain a user preference vector and an extended vector corresponding to the object to be recommended. Gating processing can be performed on the user preference vector and the extended vector corresponding to the object to be recommended respectively to obtain a gating vector corresponding to the user preference vector and a gating vector corresponding to the extended vector. Spatial alignment processing can be performed on the two gating vectors respectively to map the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector to the object recommendation space for data alignment, obtaining a target user preference vector and a target extended vector corresponding to the object to be recommended. According to the user-side data, the attribute data corresponding to the object to be recommended, the historical interaction context data corresponding to the object to be recommended, the target user preference vector, and the target extended vector corresponding to the object to be recommended, a recommendation score corresponding to the object to be recommended is determined. Furthermore, based on a preset recommendation score threshold and the recommendation score corresponding to the object to be recommended, target object data can be determined and sent to the current graphical user interface of the target terminal device for display, thereby improving the matching accuracy between the user preference information and the attribute information of the object to be recommended, enhancing the flexibility and adaptability of data processing, improving the relevance and accuracy of vector representation through the gating mechanism and spatial alignment processing, and enhancing the personalization and accuracy of the recommendation result. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the following drawings are only some embodiments of the present disclosure. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0011] Figure 1 It is a schematic diagram of the application scenario of the embodiments of the present disclosure;

[0012] Figure 2 is a schematic flowchart of an object recommendation method provided by an embodiment of the present disclosure;

[0013] Figure 3 is a schematic structural diagram of an object recommendation apparatus provided by an embodiment of the present disclosure;

[0014] Figure 4 is a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure. Detailed implementation manners

[0015] In the following description, specific details such as specific system architectures and technologies are presented for the purpose of illustration rather than limitation, so as to thoroughly understand the embodiments of the present disclosure. However, those skilled in the art should clearly understand that the present disclosure can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid unnecessary details from interfering with the description of the present disclosure.

[0016] It should be noted that the user information (including but not limited to terminal device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in the present disclosure are all information and data that have been authorized by the user or fully authorized by all parties, and the technical solutions involved in the present disclosure are automatically processed by an object recommendation model completed through training.

[0017] Next, an object recommendation method and apparatus according to an embodiment of the present disclosure will be described in detail with reference to the accompanying drawings.

[0018] Figure 1 is a schematic diagram of an application scenario of an embodiment of the present disclosure. This application scenario may include terminal devices 1, 2, and 3, a server 4, and a network 5.

[0019] The terminal devices 1, 2, and 3 can be hardware or software. When the terminal devices 1, 2, and 3 are hardware, they can be various electronic devices with a display screen and supporting communication with the server 4, including but not limited to smart phones, tablet computers, laptop portable computers, and desktop computers, etc.; when the terminal devices 1, 2, and 3 are software, they can be installed in the above-mentioned electronic devices. The terminal devices 1, 2, and 3 can be implemented as multiple software or software modules, or can be implemented as a single software or software module, and the embodiments of the present disclosure do not limit this. Further, various applications can be installed on the terminal devices 1, 2, and 3, such as data processing applications, instant messaging tools, social platform software, search applications, shopping applications, etc.

[0020] The server 4 may be a server that provides various services. For example, it may be a background server that receives requests sent by terminal devices with which it establishes a communication connection. This background server can receive and analyze requests sent by terminal devices and generate processing results. The server 4 may be a single server, a server cluster composed of several servers, or a cloud computing service center. The embodiments of the present disclosure do not limit this.

[0021] It should be noted that the server 4 may be hardware or software. When the server 4 is hardware, it may be various electronic devices that provide various services for the terminal devices 1, 2, and 3. When the server 4 is software, it may be multiple software or software modules that provide various services for the terminal devices 1, 2, and 3, or a single software or software module that provides various services for the terminal devices 1, 2, and 3. The embodiments of the present disclosure do not limit this.

[0022] The network 5 may be a wired network connected by coaxial cables, twisted pairs, and optical fibers, or a wireless network that can interconnect various communication devices without wiring. For example, Bluetooth, Near Field Communication (NFC), Infrared, etc. The embodiments of the present disclosure do not limit this.

[0023] Users can establish a communication connection with the server 4 via the terminal devices 1, 2, and 3 through the network 5 to receive or send information, etc. Specifically, the server 4 can obtain user-side data, attribute data corresponding to the object to be recommended, and historical interaction context data corresponding to the object to be recommended via the terminal devices 1, 2, and 3. Furthermore, the server 4 can perform text generation processing on the user-side data and the attribute data corresponding to the object to be recommended respectively through a large language model. It can process the user-side data to obtain user preference data containing user preference information, and process the attribute data corresponding to the object to be recommended to obtain extended data corresponding to the object to be recommended containing information related to the object to be recommended. Then, it can perform text encoding processing on the user preference data and the extended data corresponding to the object to be recommended to obtain a user preference vector and an extended vector corresponding to the object to be recommended. It can perform gating processing on the user preference vector and the extended vector corresponding to the object to be recommended respectively to obtain a gating vector corresponding to the user preference vector and a gating vector corresponding to the extended vector. It can perform spatial alignment processing on the two gating vectors respectively, map the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector to the object recommendation space for data alignment, and obtain a target user preference vector and a target extended vector corresponding to the object to be recommended. According to the user-side data, the attribute data corresponding to the object to be recommended, the historical interaction context data corresponding to the object to be recommended, the target user preference vector, and the target extended vector corresponding to the object to be recommended, the recommendation score corresponding to the object to be recommended is determined. Then, based on a preset recommendation score threshold and the recommendation score corresponding to the object to be recommended, target object data is determined and sent to the current graphical user interface of the target terminal device for display.

[0024] It should be noted that the specific types, quantities, and combinations of the terminal devices 1, 2, and 3, the server 4, and the network 5 can be adjusted according to the actual requirements of the application scenario, and the embodiments of the present disclosure do not limit this.

[0025] Figure 2 It is a schematic flowchart of an object recommendation method provided by an embodiment of the present disclosure. Figure 2 The object recommendation method can be executed by Figure 1 the server of Figure 2 As shown in

[0026] Step 201, process the user-side data and the attribute data corresponding to the object to be recommended respectively through a large language model to obtain user preference data and extended data corresponding to the object to be recommended.

[0027] Specifically, a pre-trained large language model can be used to perform text generation processing on the user-side data and the attribute data corresponding to the object to be recommended, obtaining user preference data and the extended data corresponding to the object to be recommended. The user-side data can be the user's attribute data, including but not limited to the information entered by the user's terminal device and / or historical behavior information, etc. The attribute data corresponding to the object to be recommended can be the description, price, category, and / or user evaluation, etc. of the object to be recommended, which is not limited here. The user preference data can contain information related to the user's interests and needs, while the extended data corresponding to the object to be recommended can be the enrichment and extension of the attribute information of the object to be recommended, including but not limited to the historical interaction information, award information, and / or release information, etc. of the object to be recommended, thereby enhancing the personalization degree of the object recommendation model, improving the accuracy of the recommendation result, and enhancing the accuracy of the matching between the user's needs and the object's attributes.

[0028] Among them, the large language model can be a deep learning model trained based on a large corpus, which can process and understand complex text information. The user-side data can be a data set related to the user's behavior and attributes, and the user-side data can be used to represent the information entered by the user's terminal device and / or historical behavior information, etc., which is not limited here. The object to be recommended can be an object prepared to be recommended to this user, including but not limited to goods, texts, videos, or images, etc. The attribute data corresponding to the object to be recommended can refer to the data used to represent the characteristics and attributes of the object to be recommended, such as the specifications, prices, and / or evaluations of goods, etc. The user preference data can be obtained by processing the user-side data through the large language model, and the user preference data can be used to represent the personalized needs and interests of this user. The extended data corresponding to the object to be recommended can be the feature information obtained by processing the attribute data corresponding to the object to be recommended through the large language model, including but not limited to the historical interaction information, award information, and / or release information, etc. of the object to be recommended.

[0029] For example, in an e-commerce platform, the object to be recommended can be sports shoes. Then, the large language model can process the user-side data such as the information entered by the user's terminal device, search history, browsing records, and purchase behavior, etc., to obtain the user's preference data for sports shoes, such as brand preference, price range, style preference, etc. At the same time, the large language model can process the attribute data of the sports shoes, such as material, weight, breathability, user evaluation, etc., to generate the extended data corresponding to the sports shoes, including potential features such as applicable scenarios and wearing comfort.

[0030] Step 202: Perform text encoding processing on the user preference data and the extended data corresponding to the object to be recommended respectively to obtain a user preference vector and an extended vector corresponding to the object to be recommended.

[0031] Specifically, through text encoding models in Natural Language Processing (NLP), including but not limited to Word2Vec or Bidirectional Encoder Representations from Transformers (BERT), etc., the user preference data and the extended data corresponding to the object to be recommended can be processed to convert the text information into numerical vectors in a high-dimensional vector space, thereby obtaining the user preference vector and the extended vector corresponding to the object to be recommended. The user preference vector can be a numerical vector obtained through text encoding processing, and the user preference vector can be used to represent the user's interests, preferences, or demand information; the extended vector corresponding to the object to be recommended can be a numerical vector extracted and encoded according to the extended data corresponding to the object to be recommended, thereby enhancing the matching accuracy between the user preference and the object to be recommended, improving the personalization degree of the object recommendation model. Through text encoding processing, the unstructured text data is converted into structured numerical vectors, enabling the recommendation algorithm to analyze the data more effectively and improving the accuracy of the recommendation results.

[0032] For example, in an e-commerce platform, the user preference data may include the user's terminal device input information, browsing history, purchase records, and search keywords, etc., and the extended data corresponding to the object to be recommended may include the product title, description, and / or user reviews, etc. Through text encoding processing, the aforementioned text data can be converted into corresponding numerical vectors, namely the user's preference vector and the extended vector corresponding to the object to be recommended.

[0033] Step 203: Based on the gating mechanism, process the user preference vector and the extended vector corresponding to the object to be recommended respectively to obtain the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector.

[0034] Specifically, through the gating mechanism, non-linear transformation can be performed on the user preference vector and the extended vector corresponding to the object to be recommended to obtain the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector respectively, thereby enhancing the dynamic adjustment ability of the object recommendation model, enabling the object recommendation model to adjust the focus of attention according to different users and objects to be recommended; improving the accuracy and personalization degree of the object recommendation model, and enhancing the robustness and adaptability of the object recommendation model.

[0035] Among them, the gating mechanism can be a network structure for controlling information flow, and the gating mechanism can determine the information to be retained, discarded, or modified through the learned weights.

[0036] The gating vector corresponding to the user preference vector can be a vector obtained by processing the user preference vector through a gating mechanism. The gating vector corresponding to the user preference vector can be used to represent the recognition and emphasis of the information in the user preference by the object recommendation model.

[0037] The gating vector corresponding to the extended vector can be a vector obtained by processing the extended vector of the object to be recommended through a gating mechanism. The gating vector corresponding to the extended vector can be used to represent the recognition and emphasis of the information in the object features by the object recommendation model.

[0038] For example, in the product recommendation of an e-commerce platform, the gating vectors corresponding to the user preference vector or the product feature extended vector that are important for the recommendation result can be determined through a gating mechanism, and the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector are obtained.

[0039] Step 204: Perform spatial alignment processing on the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector respectively to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended.

[0040] Specifically, through methods such as the attention mechanism or matrix multiplication (not limited here), perform spatial dimension alignment processing on the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended, thereby enhancing the correlation between the user preference information and the object feature information to be recommended, improving the personalization degree of the object recommendation model, and enhancing the accuracy of the object recommendation result.

[0041] Among them, the target user preference vector can be a vector representation obtained by performing spatial alignment processing on the gating vector corresponding to the user preference vector, and the target user preference vector can be used to represent the personal interests and preference information of the user.

[0042] The target extended vector corresponding to the object to be recommended can be a vector representation obtained by processing the gating vector corresponding to the extended vector through spatial alignment, so that the target extended vector corresponding to the object to be recommended and the target user preference vector are aligned in the same spatial dimension.

[0043] For example, in the recommendation system of an e-commerce platform, the user preference vector can include the user's preference information for aspects such as product type, price, and brand, and the extended vector corresponding to the product can include attribute information such as the description of the product, user evaluation, and sales volume. Through spatial alignment processing, the vector representation corresponding to the overall preference information of the user and the vector representation corresponding to the extended information of the product are obtained, that is, the target user preference vector and the target extended vector corresponding to the object to be recommended.

[0044] Step 205: Determine the recommendation score corresponding to the object to be recommended based on the user-side data, the attribute data corresponding to the object to be recommended, the historical interaction context data corresponding to the object to be recommended, the target user preference vector, and the target extension vector corresponding to the object to be recommended.

[0045] Specifically, a Multilayer Perceptron (MLP) can be used to calculate the user-side data, the attribute data corresponding to the object to be recommended, the historical interaction context data corresponding to the object to be recommended, the target user preference vector, and the target extension vector corresponding to the object to be recommended, and determine the recommendation score of the object to be recommended for the target user. The historical interaction context data corresponding to the object to be recommended includes, but is not limited to, information such as the time, location, and environment of the historical interaction between the user and the object to be recommended. This enhances the personalization of the object recommendation model, improves the accuracy and relevance of the recommendation results, and makes the recommendation results more in line with the actual needs of the user by comprehensively considering multiple data sources and features.

[0046] Among them, the recommendation score corresponding to the object to be recommended can be a quantitative indicator, and the recommendation score corresponding to the object to be recommended can be used to represent the attractiveness and recommendation value of the object to be recommended for the user. The recommendation score can be obtained through the processing of the MLP based on features such as user-side data, attribute data of the object to be recommended, historical interaction context data, target user preference vector, and target extension vector of the object to be recommended. The recommendation score can be used for aspects such as sorting, filtering recommended objects, and evaluating the recommendation effect, and is not limited here.

[0047] For example, in an e-commerce platform, when a user browses products, the recommendation score of the product for the user can be calculated based on user-side data such as the information entered by the user's terminal device, browsing history, purchase records, and search keywords, attribute data such as the category, price, brand, and sales volume of the product, and historical interaction context data such as the time, location, and environment of the user's purchase of the product, in combination with the user preference vector and the product extension vector.

[0048] Step 206: Based on a preset recommendation score threshold and the recommendation score corresponding to the object to be recommended, determine the target object data, and send the target object data to the target terminal device so that the target object data is displayed on the current graphical user interface of the target terminal device.

[0049] Specifically, according to a pre-set recommended score threshold (which can be a critical value obtained by comprehensively considering factors such as business logic, user preferences, and data quality, and can be used to distinguish whether a to-be-recommended object is worthy of being recommended), the recommended score of the to-be-recommended object can be evaluated. When the recommended score corresponding to the to-be-recommended object is greater than or equal to the pre-set recommended score threshold, the data of the to-be-recommended object is determined as target object data. When the recommended score corresponding to the to-be-recommended object is less than the pre-set recommended score threshold, the data of the to-be-recommended object is not target object data. Furthermore, the target object data can be sent to a target terminal device, which can be used to display the target object data on the currently displayed graphical user interface. The target terminal device can be the user's terminal device, including but not limited to a smart phone, a tablet computer, or a computer, etc. The current graphical user interface can be an interface visible and interactive to the user through the screen. In this way, the accuracy of the recommendation result is enhanced, the efficiency of the user to obtain useful information is improved, the user experience is enhanced, and by setting the recommended score threshold to screen the target object data, the time cost for the user to screen information is reduced, and the user's satisfaction with the recommended content is increased.

[0050] Among them, the pre-set recommended score threshold can be a score boundary for determining whether a to-be-recommended object is a target object. When the recommended score is greater than or equal to the pre-set recommended score threshold, it can be determined that the to-be-recommended object is target object data.

[0051] The target object data can be object data obtained through screening, and the recommended score of the target object data is greater than or equal to the pre-set recommended score threshold. The target object data can be product information, text content, audio content, or video content, etc., which is not limited here.

[0052] The target terminal device can be a device for receiving and displaying the target object data. The target terminal device can be a smart device that the user is using or will use, including a mobile phone, a tablet computer, a computer, etc., which is not limited here.

[0053] The current graphical user interface can be an interface displayed on the screen and available for the user to interact with when the user is using the target terminal device. The current graphical user interface can include various information elements, including but not limited to text, pictures, videos, and / or operation buttons, etc. The current graphical user interface can be used to achieve information exchange and operation control between the user and the target terminal device.

[0054] For example, in an e-commerce platform, the recommendation score of the product to be recommended is 90. According to a preset recommendation score threshold, such as 80 points, since 90 is greater than 80, it can be determined that the product to be recommended is a target product. Then, when the user opens the application program of the e-commerce platform, the data of the target product can be sent to the user's mobile phone and displayed on the current graphical user interface of the mobile phone.

[0055] According to the technical solution provided by the embodiments of the present disclosure, by performing text generation processing on the user-side data and the attribute data corresponding to the object to be recommended through a large language model, user preference data containing user preference information can be obtained by processing the user-side data, and extended data corresponding to the object to be recommended containing information related to the object to be recommended can be obtained by processing the attribute data corresponding to the object to be recommended. Furthermore, text encoding processing can be performed on the user preference data and the extended data corresponding to the object to be recommended to obtain a user preference vector and an extended vector corresponding to the object to be recommended. Gating processing can be performed on the user preference vector and the extended vector corresponding to the object to be recommended respectively to obtain a gating vector corresponding to the user preference vector and a gating vector corresponding to the extended vector. Spatial alignment processing can be performed on the two gating vectors respectively to map the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector to the object recommendation space for data alignment, obtaining a target user preference vector and a target extended vector corresponding to the object to be recommended. According to the user-side data, the attribute data corresponding to the object to be recommended, the historical interaction context data corresponding to the object to be recommended, the target user preference vector, and the target extended vector corresponding to the object to be recommended, the recommendation score corresponding to the object to be recommended is determined. Furthermore, based on the preset recommendation score threshold and the recommendation score corresponding to the object to be recommended, target object data can be determined and sent to the current graphical user interface of the target terminal device for display, thereby improving the matching accuracy between the user preference information and the attribute information of the object to be recommended, enhancing the flexibility and adaptability of data processing, improving the relevance and accuracy of vector representation through the gating mechanism and spatial alignment processing, and enhancing the personalization and accuracy of the recommendation result.

[0056] In some embodiments, processing the user-side data and the attribute data corresponding to the object to be recommended through a large language model to obtain user preference data and extended data corresponding to the object to be recommended includes: processing the user-side data based on a preset preference generation template to obtain user-side templated data; processing the attribute data corresponding to the object to be recommended based on a preset extended generation template to obtain templated attribute data corresponding to the object to be recommended; and processing the user-side templated data and the templated attribute data corresponding to the object to be recommended through a large language model to obtain user preference data and extended data corresponding to the object to be recommended.

[0057] Specifically, the user-side data can be structured through a preset preference generation template, and the user-side data can be configured to the corresponding positions of the preset preference generation template to obtain the templatized user-side data; the attribute data of the object to be recommended can be structured through a preset extension generation template to obtain the templatized attribute data of the object to be recommended. Furthermore, the templatized user-side data and the templatized attribute data corresponding to the object to be recommended can be input into the large language model for text generation processing to obtain the user preference data and the extended data corresponding to the object to be recommended.

[0058] Among them, the preset preference generation template can be a prompting engineering framework for converting user-side data into structured template data. The preset preference generation template can be used to define the rules for extracting target information from the original user data and organize the extracted target information into a format suitable for processing by the large language model.

[0059] The templatized user-side data can be structured data obtained by processing the user-side data through the application of the preset preference generation template.

[0060] The preset extension generation template can be a prompting engineering framework for converting the attribute data corresponding to the object to be recommended into structured template data. The preset extension generation template can be used to define the rules for extracting target information from the original attribute data and organize the extracted target information into a format suitable for processing by the large language model.

[0061] The templatized attribute data corresponding to the object to be recommended can be structured data obtained by processing the attribute data corresponding to the object to be recommended through the application of the preset extension generation template.

[0062] For example, in an e-commerce platform, the user-side data includes the user's terminal device entry information, browsing records, purchase records, and search keywords, etc. The object to be recommended can be the product to be recommended on the platform. The user-side data can be converted into templatized data through the preset preference generation template. For example, the user's browsing record can be converted into the form of "the user browsed {product category}", and the attribute data corresponding to the product to be recommended can be converted into templatized data through the preset extension generation template. For example, the description information of the product to be recommended can be converted into the form of "{product name} is {product category}, with {features}". Furthermore, the above templatized data can be input into the large language model for text generation processing to obtain the user preference data and the extended data corresponding to the product to be recommended.

[0063] According to the technical solution provided by the embodiments of the present disclosure, the user-side data is structurally processed through a preset preference generation template, and the user-side data is configured at the corresponding positions of the preset preference generation template to obtain the user-side templated data; the attribute data of the object to be recommended can be structurally processed through a preset extension generation template to obtain the templated attribute data of the object to be recommended. Furthermore, the user-side templated data and the templated attribute data corresponding to the object to be recommended can be input into the large language model for text generation processing to obtain the user preference data and the extended data corresponding to the object to be recommended, thereby enhancing the structural degree of the data, improving the efficiency and accuracy of data processing, and enhancing the personalized recommendation ability of the object recommendation model.

[0064] In some embodiments, determining the recommendation score corresponding to the object to be recommended according to the user-side data, the attribute data corresponding to the object to be recommended, the historical interaction context data corresponding to the object to be recommended, the target user preference vector, and the target extension vector corresponding to the object to be recommended includes: concatenating the target user preference vector and the target extension vector corresponding to the object to be recommended to obtain a recommendation enhancement vector; respectively performing feature extraction processing on the user-side data, the attribute data corresponding to the object to be recommended, and the historical interaction context data corresponding to the object to be recommended to obtain a user-side vector, an attribute vector corresponding to the object to be recommended, and a historical interaction context vector corresponding to the object to be recommended; performing vector merging processing on the recommendation enhancement vector, the user-side vector, the attribute vector corresponding to the object to be recommended, and the historical interaction context vector corresponding to the object to be recommended to obtain a target vector corresponding to the object to be recommended; and determining the recommendation score corresponding to the object to be recommended according to the target vector corresponding to the object to be recommended.

[0065] Specifically, the target user preference vector and the target extension vector corresponding to the object to be recommended can be concatenated to obtain a recommendation enhancement vector. The concatenation can be achieved through direct concatenation, dimension matching concatenation, or linear transformation concatenation, etc., which is not limited herein. Feature extraction can be respectively performed on the user-side data, the attribute data corresponding to the object to be recommended, and the historical interaction context data corresponding to the object to be recommended to obtain a user-side vector, an attribute vector corresponding to the object to be recommended, and a historical interaction context vector. The feature extraction can be achieved through a convolutional neural network, word embedding, or a recurrent neural network, etc., which is not limited herein. Furthermore, the above vectors can be processed to obtain a target vector corresponding to the object to be recommended through weighted merging or non-linear merging, etc., which is not limited herein. The recommendation score corresponding to the object to be recommended can be calculated according to the target vector corresponding to the object to be recommended, and the recommendation score corresponding to the object to be recommended can be calculated through dot product, cosine similarity, or Euclidean distance, etc.

[0066] Among them, the recommended enhanced vector can be a vector obtained by concatenating the target user preference vector and the target extended vector corresponding to the object to be recommended. The recommended enhanced vector can be used to represent the preference information of the user and the extended information of the object to be recommended. The concatenation can be achieved by direct concatenation, dimension-matching concatenation, or linear transformation concatenation, etc., which is not limited here.

[0067] The historical interaction context data corresponding to the object to be recommended can be the historical context information when the user interacts with the object to be recommended, including but not limited to time, location, current climate, and / or user behavior, etc. The historical interaction context data corresponding to the object to be recommended can be obtained by recording and analyzing the historical behavior information of the user. The historical interaction context data can be used to evaluate the user's potential interest in the object to be recommended.

[0068] The user-side vector can be a vector obtained by performing feature extraction processing on user-side data. The user-side vector can be used to represent the feature information of the user, including but not limited to the user's age, gender, and / or interest preferences, etc. The user-side vector can be used to understand the user's needs and preferences.

[0069] The attribute vector corresponding to the object to be recommended can be a vector obtained by performing feature extraction processing on the attribute data corresponding to the object to be recommended. The attribute vector corresponding to the object to be recommended can be used to represent the inherent attribute information of the object to be recommended, including but not limited to the price, brand, and / or category of the object to be recommended, etc.

[0070] The historical interaction context vector corresponding to the object to be recommended can be a vector obtained by performing feature extraction processing on the historical interaction context data corresponding to the object to be recommended. The historical interaction context vector corresponding to the object to be recommended can be used to represent the context features when the user interacts with the object to be recommended historically, including but not limited to the user's behavior patterns at different times and locations, etc. The historical interaction context vector can be used to analyze the user's preference changes for the object to be recommended in different contexts.

[0071] For example, in an e-commerce platform, when a user browses products, based on user-side data such as the user's terminal device input information, historical purchase records, browsing records, and search records, the attribute data of the product, such as price, brand, and category, the historical interaction context data between the product and the user, such as the time, location, and weather when the user clicks or purchases the product, the user's preference vector, and the extended vector of the product, such as the sales volume and evaluation of the product, a recommended enhanced vector can be obtained. Then, the above vectors can be merged and processed to obtain the target vector of the product, and the recommendation score of the product to be recommended can be calculated based on the target vector of the product.

[0072] According to the technical solution provided by the embodiments of the present disclosure, by concatenating the target user preference vector and the target extended vector corresponding to the object to be recommended, a recommendation enhancement vector can be obtained. Feature extraction can be respectively performed on the user-side data, the attribute data corresponding to the object to be recommended, and the historical interaction context data corresponding to the object to be recommended to obtain the user-side vector, the attribute vector corresponding to the object to be recommended, and the historical interaction context vector. Furthermore, the above vectors can be processed through weighted merging or non-linear merging and other methods to obtain the target vector corresponding to the object to be recommended. The recommendation score corresponding to the object to be recommended can be calculated based on the target vector corresponding to the object to be recommended, thereby enhancing the personalization degree of the object recommendation model, improving the accuracy of the recommendation result, and enhancing the user experience.

[0073] In some embodiments, determining the recommendation score corresponding to the object to be recommended according to the target vector corresponding to the object to be recommended includes: flattening the target vector corresponding to the object to be recommended to obtain the target low-dimensional vector corresponding to the object to be recommended; performing a mapping process on the target low-dimensional vector corresponding to the object to be recommended through a multi-layer perceptron to obtain the space mapping vector corresponding to the object to be recommended; performing a recommendation score calculation process on the space mapping vector corresponding to the object to be recommended to obtain the recommendation score corresponding to the object to be recommended.

[0074] Specifically, the target vector corresponding to the object to be recommended can be flattened to reduce the dimension of the target vector corresponding to the object to be recommended to obtain the target low-dimensional vector corresponding to the object to be recommended. A non-linear mapping process can be performed on the target low-dimensional vector through a multi-layer perceptron to obtain the space mapping vector corresponding to the object to be recommended. The recommendation score of the object to be recommended can be obtained by performing a recommendation score calculation process on the space mapping vector.

[0075] Among them, the target low-dimensional vector corresponding to the object to be recommended can be a low-dimensional representation obtained by flattening the high-dimensional target vector corresponding to the object to be recommended. The flattening process can be implemented through dimensionality reduction methods such as principal component analysis (PCA) or singular value decomposition (SVD), which is not limited here.

[0076] The multi-layer perceptron can be a type of feedforward artificial neural network model. The multi-layer perceptron can be composed of an input layer, several hidden layers, and an output layer. The neurons of each layer are fully connected to the neurons of the next layer, and a non-linear mapping from input to output is realized through a non-linear activation function. The non-linear activation function includes but is not limited to ReLU or sigmoid, etc.

[0077] The spatial mapping vector corresponding to the object to be recommended can be a vector representation obtained by performing a non-linear transformation on the target low-dimensional vector through a multi-layer perceptron. This spatial mapping vector representation can be processed in the hidden layer or output layer of the multi-layer perceptron.

[0078] For example, in an e-commerce platform, the target low-dimensional vector corresponding to the product to be recommended can be obtained through flattening processing. The target low-dimensional vector can be mapped through a multi-layer perceptron to obtain the spatial mapping vector corresponding to the product. Furthermore, the recommendation score corresponding to the product to be recommended can be calculated based on the spatial mapping vector.

[0079] According to the technical solution provided by the embodiments of the present disclosure, by flattening the target vector corresponding to the object to be recommended, the dimension of the target vector corresponding to the object to be recommended is reduced to obtain the target low-dimensional vector corresponding to the object to be recommended. The target low-dimensional vector can be non-linearly mapped through a multi-layer perceptron to obtain the spatial mapping vector corresponding to the object to be recommended. The recommendation score corresponding to the object to be recommended can be obtained by performing recommendation score calculation processing on the spatial mapping vector, thereby enhancing the data processing ability of the object recommendation model, improving the accuracy and precision of the recommendation result, and enhancing the user experience and satisfaction.

[0080] In some embodiments, based on a gating mechanism, the user preference vector and the extended vector corresponding to the object to be recommended are respectively processed to obtain the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector, including: performing a non-linear transformation on the user preference vector to obtain the retained vector corresponding to the user preference vector; performing a reset transformation on the retained vector corresponding to the user preference vector to obtain the gating vector corresponding to the user preference vector; performing a non-linear transformation on the extended vector corresponding to the object to be recommended to obtain the retained vector corresponding to the extended vector; performing a reset transformation on the retained vector corresponding to the extended vector to obtain the gating vector corresponding to the extended vector.

[0081] Specifically, the user preference vector can be processed through a non-linear transformation. The non-linear transformation can be implemented through a Sigmoid function or a ReLU function, which is not limited here, to obtain the retained vector corresponding to the user preference vector. The retained vector retains the target information of the user preference. The retained vector can be further processed through a reset transformation. The reset transformation can be implemented through the reset gate of the gating mechanism to obtain the gating vector corresponding to the user preference vector. The extended vector corresponding to the object to be recommended can be non-linearly transformed to obtain the retained vector corresponding to the extended vector. The retained vector can be used to characterize the important features of the object to be recommended. Furthermore, the retained vector corresponding to the extended vector can be processed through the reset gate to obtain the gating vector corresponding to the extended vector.

[0082] Among them, the retention vector corresponding to the user preference vector can be a vector obtained by processing the user preference vector through a non-linear transformation. The retention vector corresponding to the user preference vector can be used to represent the target information of the user preference. This target information can be extracted through a non-linear activation function in a neural network. The non-linear activation function includes but is not limited to Sigmoid or ReLU, etc.

[0083] The retention vector corresponding to the extended vector can be a vector obtained by processing the extended vector corresponding to the object to be recommended through a non-linear transformation. The retention vector corresponding to the extended vector can be used to represent the target feature of the object to be recommended. This target feature can be extracted through a non-linear activation function in a neural network, and this target feature can be used to evaluate the matching degree between the object to be recommended and the user preference.

[0084] For example, in an e-commerce platform, the user preference vector can include the encoding of information such as the user's preference for product categories, price ranges, and brands. The extended vector of the object to be recommended can include the encoding of information such as the detailed description of the product, user reviews, and sales volume. Through non-linear transformation, the retention vector corresponding to the user preference vector and the retention vector corresponding to the product feature can be extracted respectively. Through the reset gate, a reset transformation is performed to generate the gated vector corresponding to the user preference vector and the gated vector corresponding to the product feature.

[0085] According to the technical solution provided by the embodiments of the present disclosure, the user preference vector is processed through a non-linear transformation to obtain the retention vector corresponding to the user preference vector. This retention vector retains the target information of the user preference. Further processing can be performed on the retention vector through a reset transformation to obtain the gated vector corresponding to the user preference vector. The extended vector corresponding to the object to be recommended can be subjected to a non-linear transformation to obtain the retention vector corresponding to the extended vector. Furthermore, the retention vector corresponding to the extended vector can be processed through the reset gate to obtain the gated vector corresponding to the extended vector, thereby enhancing the expression ability of the object recommendation model for user preferences and the characteristics of the object to be recommended, and improving the accuracy and flexibility of the object recommendation model.

[0086] In some embodiments, spatial alignment processing is respectively performed on the gated vector corresponding to the user preference vector and the gated vector corresponding to the extended vector to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended, including: performing dimension alignment processing on the gated vector corresponding to the user preference vector and the gated vector corresponding to the extended vector respectively to obtain the co-dimensional vector corresponding to the user preference vector and the co-dimensional vector corresponding to the extended vector; performing spatial transformation processing on the co-dimensional vector corresponding to the user preference vector and the co-dimensional vector corresponding to the extended vector respectively to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended.

[0087] Specifically, dimensional alignment processing can be performed on the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector, so that the dimensions of the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector are the same, obtaining the co-dimensional vector corresponding to the user preference vector and the co-dimensional vector corresponding to the extended vector. Spatial transformation processing can be performed on the co-dimensional vector corresponding to the user preference vector and the co-dimensional vector corresponding to the extended vector respectively. The spatial transformation can be achieved through matrix multiplication or non-linear transformation, etc., which is not limited here. The co-dimensional vector corresponding to the user preference vector and the co-dimensional vector corresponding to the extended vector are mapped into a specified feature space to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended.

[0088] Among them, the co-dimensional vector corresponding to the user preference vector can be a vector obtained through dimensional alignment processing. The co-dimensional vector corresponding to the user preference vector and the co-dimensional vector corresponding to the extended vector have the same dimension. The co-dimensional vector corresponding to the user preference vector can be obtained by adjusting the dimension of the gating vector corresponding to the user preference vector. The adjustment process can be achieved through filling, truncation, or mapping, etc., which is not limited here.

[0089] The co-dimensional vector corresponding to the extended vector can be a vector obtained through dimensional alignment processing. The co-dimensional vector corresponding to the extended vector and the co-dimensional vector corresponding to the user preference vector have the same dimension. The co-dimensional vector corresponding to the extended vector can be obtained by adjusting the dimension of the gating vector corresponding to the extended vector.

[0090] For example, in an online shopping platform, the user preference vector can represent the user's preference characteristics for products, such as price, brand, and / or color, etc., while the extended vector can represent the detailed attributes of the product to be recommended, such as size, material, and / or evaluation, etc. By performing dimensional alignment and spatial transformation processing on the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector, more accurate target user preference vectors and target extended vectors corresponding to the objects to be recommended can be obtained.

[0091] According to the technical solution provided by the embodiments of the present disclosure, by performing dimension alignment processing on the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector, so that the dimensions of the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector are consistent, the co-dimensional vectors corresponding to the user preference vector and the co-dimensional vectors corresponding to the extended vector are obtained. The co-dimensional vectors corresponding to the user preference vector and the co-dimensional vectors corresponding to the extended vector can be respectively subjected to space transformation processing to map the co-dimensional vectors corresponding to the user preference vector and the co-dimensional vectors corresponding to the extended vector into a specified feature space, so as to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended, thereby enhancing the comparability between the user preference vector and the extended vector corresponding to the object to be recommended, improving the accuracy of the object recommendation model, and enhancing the user experience.

[0092] In some embodiments, respectively performing space transformation processing on the co-dimensional vector corresponding to the user preference vector and the co-dimensional vector corresponding to the extended vector to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended includes: performing space rotation processing on the co-dimensional vector corresponding to the user preference vector to obtain the space matching vector corresponding to the user preference vector; performing weight matching processing on the co-dimensional vector corresponding to the extended vector to obtain the weighted vector corresponding to the extended vector; and performing alignment processing on the space matching vector corresponding to the user preference vector and the weighted vector corresponding to the extended vector to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended.

[0093] Specifically, space rotation processing can be performed on the co-dimensional vector corresponding to the user preference vector to obtain a space matching vector that is more matched with the multi-layer perceptron. Weight matching processing can be performed on the co-dimensional vector corresponding to the extended vector, that is, different weights are assigned according to the importance of each dimension to obtain the weighted vector corresponding to the extended vector. Furthermore, the space matching vector corresponding to the user preference vector and the weighted vector corresponding to the extended vector can be aligned to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended.

[0094] Among them, the space matching vector corresponding to the user preference vector can be a vector obtained by performing space rotation processing on the co-dimensional vector corresponding to the user preference vector.

[0095] The weighted vector corresponding to the extended vector can be a vector obtained by performing weight matching processing on the co-dimensional vector corresponding to the extended vector, and the weights of each dimension can be used to represent the importance of that dimension in the recommendation.

[0096] For example, in an online shopping platform, the user preference vector can represent the preference information of the user for dimensions such as the price, brand, color, and / or material of a product, while the extended vector can represent additional information such as the sales volume, evaluation, and / or new product identification of the product to be recommended. Through spatial rotation processing, the same-dimensional vector corresponding to the user preference vector can be converted into a spatially matching vector, and through weight matching processing, the same-dimensional vector corresponding to the extended vector can be weighted. Furthermore, the spatially matching vector corresponding to the user preference vector and the weighted vector corresponding to the extended vector can be aligned to obtain the target user preference vector and the target extended vector corresponding to the product to be recommended.

[0097] According to the technical solution provided by the embodiments of the present disclosure, by performing spatial rotation processing on the same-dimensional vector corresponding to the user preference vector to obtain a spatially matching vector that is more matched with the multi-layer perceptron, weight matching processing can be performed on the same-dimensional vector corresponding to the extended vector to obtain a weighted vector corresponding to the extended vector. Furthermore, the spatially matching vector corresponding to the user preference vector and the weighted vector corresponding to the extended vector can be aligned to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended, thereby enhancing the correlation between the user preference and the object to be recommended, improving the accuracy of the object recommendation model, and enhancing the user experience.

[0098] All of the above optional technical solutions can be combined arbitrarily to form optional embodiments of the present disclosure, which will not be elaborated herein one by one.

[0099] The following is an embodiment of the apparatus of the present disclosure, which can be used to execute the method embodiments of the present disclosure. For details not disclosed in the embodiments of the apparatus of the present disclosure, please refer to the method embodiments of the present disclosure.

[0100] Figure 3 is a schematic diagram of an object recommendation apparatus provided by an embodiment of the present disclosure. As Figure 3 shown, the object recommendation apparatus includes:

[0101] A first processing module 301, configured to process the user-side data and the attribute data corresponding to the object to be recommended through a large language model respectively to obtain user preference data and extended data corresponding to the object to be recommended;

[0102] A second processing module 302, configured to perform text encoding processing on the user preference data and the extended data corresponding to the object to be recommended respectively to obtain a user preference vector and an extended vector corresponding to the object to be recommended;

[0103] A third processing module 303, configured to process the user preference vector and the extended vector corresponding to the object to be recommended respectively based on a gating mechanism to obtain a gating vector corresponding to the user preference vector and a gating vector corresponding to the extended vector;

[0104] The fourth processing module 304 is configured to perform spatial alignment processing on the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector respectively, so as to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended;

[0105] The determination module 305 is configured to determine the recommendation score corresponding to the object to be recommended according to the user-side data, the attribute data corresponding to the object to be recommended, the historical interaction context data corresponding to the object to be recommended, the target user preference vector, and the target extended vector corresponding to the object to be recommended;

[0106] The fifth processing module 306 is configured to determine the target object data based on a preset recommendation score threshold and the recommendation score corresponding to the object to be recommended, and send the target object data to the target terminal device, so that the target object data is displayed on the current graphical user interface of the target terminal device.

[0107] According to the technical solution provided by the embodiments of the present disclosure, by performing text generation processing on the user-side data and the attribute data corresponding to the object to be recommended through a large language model, user preference data including user preference information can be obtained by processing the user-side data, and extended data corresponding to the object to be recommended including information related to the object to be recommended can be obtained by processing the attribute data corresponding to the object to be recommended. Furthermore, text encoding processing can be performed on the user preference data and the extended data corresponding to the object to be recommended to obtain the user preference vector and the extended vector corresponding to the object to be recommended. Gating processing can be performed on the user preference vector and the extended vector corresponding to the object to be recommended respectively to obtain the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector. Spatial alignment processing can be performed on the two gating vectors respectively to map the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector to the object recommendation space for data alignment, so as to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended. The recommendation score corresponding to the object to be recommended is determined according to the user-side data, the attribute data corresponding to the object to be recommended, the historical interaction context data corresponding to the object to be recommended, the target user preference vector, and the target extended vector corresponding to the object to be recommended. Furthermore, the target object data can be determined based on a preset recommendation score threshold and the recommendation score corresponding to the object to be recommended, and the target object data is sent to the current graphical user interface of the target terminal device for display, thereby improving the matching accuracy between the user preference information and the attribute information of the object to be recommended, enhancing the flexibility and adaptability of data processing, improving the relevance and accuracy of vector representation through the gating mechanism and spatial alignment processing, and enhancing the personalization and accuracy of the recommendation result.

[0108] In some embodiments, the first processing module 301 is specifically configured to process the user-side data based on a preset preference generation template to obtain user-side templatized data; process the attribute data corresponding to the object to be recommended based on a preset extension generation template to obtain templatized attribute data corresponding to the object to be recommended; and process the user-side templatized data and the templatized attribute data corresponding to the object to be recommended respectively through a large language model to obtain user preference data and extension data corresponding to the object to be recommended.

[0109] In some embodiments, the determination module 305 is specifically configured to perform a splicing process on the target user preference vector and the target extension vector corresponding to the object to be recommended to obtain a recommendation enhancement vector; perform feature extraction processes on the user-side data, the attribute data corresponding to the object to be recommended, and the historical interaction context data corresponding to the object to be recommended respectively to obtain a user-side vector, an attribute vector corresponding to the object to be recommended, and a historical interaction context vector corresponding to the object to be recommended; perform a vector merging process on the recommendation enhancement vector, the user-side vector, the attribute vector corresponding to the object to be recommended, and the historical interaction context vector corresponding to the object to be recommended to obtain a target vector corresponding to the object to be recommended; and determine a recommendation score corresponding to the object to be recommended according to the target vector corresponding to the object to be recommended.

[0110] In some embodiments, determining a recommendation score corresponding to the object to be recommended according to the target vector corresponding to the object to be recommended is specifically configured to perform a flattening process on the target vector corresponding to the object to be recommended to obtain a target low-dimensional vector corresponding to the object to be recommended; perform a mapping process on the target low-dimensional vector corresponding to the object to be recommended through a multi-layer perceptron to obtain a space mapping vector corresponding to the object to be recommended; and perform a recommendation score calculation process on the space mapping vector corresponding to the object to be recommended to obtain a recommendation score corresponding to the object to be recommended.

[0111] In some embodiments, the third processing module 303 is specifically configured to perform a non-linear transformation process on the user preference vector to obtain a reserved vector corresponding to the user preference vector; perform a reset transformation process on the reserved vector corresponding to the user preference vector to obtain a gating vector corresponding to the user preference vector; perform a non-linear transformation process on the extension vector corresponding to the object to be recommended to obtain a reserved vector corresponding to the extension vector; and perform a reset transformation process on the reserved vector corresponding to the extension vector to obtain a gating vector corresponding to the extension vector.

[0112] In some embodiments, the fourth processing module 304 is specifically configured to perform a dimension alignment process on the gating vector corresponding to the user preference vector and the gating vector corresponding to the extension vector respectively to obtain a co-dimensional vector corresponding to the user preference vector and a co-dimensional vector corresponding to the extension vector; and perform a space transformation process on the co-dimensional vector corresponding to the user preference vector and the co-dimensional vector corresponding to the extension vector respectively to obtain a target user preference vector and a target extension vector corresponding to the object to be recommended.

[0113] In some embodiments, spatial transformation processing is respectively performed on the same-dimensional vectors corresponding to the user preference vector and the same-dimensional vectors corresponding to the extended vectors to obtain a target user preference vector and a target extended vector corresponding to the object to be recommended. Specifically, spatial rotation processing is performed on the same-dimensional vectors corresponding to the user preference vector to obtain a spatial matching vector corresponding to the user preference vector; weight matching processing is performed on the same-dimensional vectors corresponding to the extended vectors to obtain a weighted vector corresponding to the extended vectors; alignment processing is performed on the spatial matching vector corresponding to the user preference vector and the weighted vector corresponding to the extended vectors to obtain a target user preference vector and a target extended vector corresponding to the object to be recommended.

[0114] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The order of execution of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present disclosure.

[0115] Figure 4 is a schematic diagram of the electronic device 4 provided by the embodiments of the present disclosure. As Figure 4 shown, the electronic device 4 of this embodiment includes: a processor 401, a memory 402, and a computer program 403 stored in the memory 402 and executable on the processor 401. When the processor 401 executes the computer program 403, the steps in the above-mentioned method embodiments are implemented. Alternatively, when the processor 401 executes the computer program 403, the functions of each module / unit in the above-mentioned device embodiments are implemented.

[0116] The electronic device 4 may be a desktop computer, a notebook, a palm computer, a cloud server, and other electronic devices. The electronic device 4 may include, but is not limited to, the processor 401 and the memory 402. Those skilled in the art can understand that Figure 4 is only an example of the electronic device 4, and does not constitute a limitation to the electronic device 4. It may include more or fewer components than shown in the figure, or different components.

[0117] The processor 401 may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0118] The memory 402 can be an internal storage unit of the electronic device 4, for example, the hard disk or memory of the electronic device 4. The memory 402 can also be an external storage device of the electronic device 4, for example, a plug-in hard disk equipped on the electronic device 4, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. The memory 402 can also include both the internal storage unit of the electronic device 4 and the external storage device. The memory 402 is used to store computer programs and other programs and data required by the electronic device.

[0119] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the above division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0120] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium (such as a computer-readable storage medium). Based on this understanding, to implement all or part of the processes in the above embodiment methods of the present disclosure, it can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above various method embodiments can be implemented. The computer program can include computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc. The computer-readable storage medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disc, a computer memory, a Read-Only Memory (ROM), a Random Access Memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0121] The above embodiments are only used to illustrate the technical solutions of the present disclosure, rather than to limit it; although the present disclosure has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present disclosure, and should all be included within the protection scope of the present disclosure.

Claims

1. An object recommendation method, characterized in that, Including: Processing the user-side data and the attribute data corresponding to the object to be recommended respectively through a large language model to obtain user preference data and the extended data corresponding to the object to be recommended; Performing text encoding processing on the user preference data and the extended data corresponding to the object to be recommended respectively to obtain a user preference vector and the extended vector corresponding to the object to be recommended; Processing the user preference vector and the extended vector corresponding to the object to be recommended respectively based on a gating mechanism to obtain the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector; Performing spatial alignment processing on the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector respectively to obtain a target user preference vector and the target extended vector corresponding to the object to be recommended; Determining the recommendation score corresponding to the object to be recommended according to the user-side data, the attribute data corresponding to the object to be recommended, the historical interaction context data corresponding to the object to be recommended, the target user preference vector, and the target extended vector corresponding to the object to be recommended; Determining target object data based on a preset recommendation score threshold and the recommendation score corresponding to the object to be recommended, and sending the target object data to a target terminal device so that the target object data is displayed on the current graphical user interface of the target terminal device.

2. The object recommendation method according to claim 1, wherein The processing the user-side data and the attribute data corresponding to the object to be recommended respectively through a large language model to obtain user preference data and the extended data corresponding to the object to be recommended includes: Processing the user-side data based on a preset preference generation template to obtain user-side templated data; Processing the attribute data corresponding to the object to be recommended based on a preset extension generation template to obtain the templated attribute data corresponding to the object to be recommended; Processing the user-side templated data and the templated attribute data corresponding to the object to be recommended respectively through the large language model to obtain the user preference data and the extended data corresponding to the object to be recommended.

3. The object recommendation method according to claim 1, characterized in that The determining the recommendation score corresponding to the object to be recommended according to the user-side data, the attribute data corresponding to the object to be recommended, the historical interaction context data corresponding to the object to be recommended, the target user preference vector, and the target extended vector corresponding to the object to be recommended includes: Performing splicing processing on the target user preference vector and the target extended vector corresponding to the object to be recommended to obtain a recommendation enhancement vector; Performing feature extraction processing on the user-side data, the attribute data corresponding to the object to be recommended, and the historical interaction context data corresponding to the object to be recommended respectively to obtain a user-side vector, the attribute vector corresponding to the object to be recommended, and the historical interaction context vector corresponding to the object to be recommended; Performing vector merging processing on the recommendation enhancement vector, the user-side vector, the attribute vector corresponding to the object to be recommended, and the historical interaction context vector corresponding to the object to be recommended to obtain the target vector corresponding to the object to be recommended; Determine the recommendation score corresponding to the object to be recommended according to the target vector corresponding to the object to be recommended.

4. The object recommendation method according to claim 3, wherein The determining the recommendation score corresponding to the object to be recommended according to the target vector corresponding to the object to be recommended includes: Flatten the target vector corresponding to the object to be recommended to obtain the target low-dimensional vector corresponding to the object to be recommended; Perform a mapping process on the target low-dimensional vector corresponding to the object to be recommended through a multi-layer perceptron to obtain the space mapping vector corresponding to the object to be recommended; Perform a recommendation score calculation process on the space mapping vector corresponding to the object to be recommended to obtain the recommendation score corresponding to the object to be recommended.

5. The object recommendation method according to claim 1, wherein The processing the user preference vector and the extended vector corresponding to the object to be recommended respectively based on a gating mechanism to obtain the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector includes: Perform a non-linear transformation process on the user preference vector to obtain the retained vector corresponding to the user preference vector; Perform a reset transformation process on the retained vector corresponding to the user preference vector to obtain the gating vector corresponding to the user preference vector; Perform a non-linear transformation process on the extended vector corresponding to the object to be recommended to obtain the retained vector corresponding to the extended vector; Perform a reset transformation process on the retained vector corresponding to the extended vector to obtain the gating vector corresponding to the extended vector.

6. The object recommendation method according to claim 1, wherein The respectively performing a space alignment process on the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended includes: Perform a dimension alignment process on the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector respectively to obtain the co-dimensional vector corresponding to the user preference vector and the co-dimensional vector corresponding to the extended vector; Perform a space transformation process on the co-dimensional vector corresponding to the user preference vector and the co-dimensional vector corresponding to the extended vector respectively to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended.

7. The object recommendation method according to claim 6, wherein The performing a space transformation process on the co-dimensional vector corresponding to the user preference vector and the co-dimensional vector corresponding to the extended vector respectively to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended includes: Perform a space rotation process on the co-dimensional vector corresponding to the user preference vector to obtain the space matching vector corresponding to the user preference vector; Perform a weight matching process on the co-dimensional vector corresponding to the extended vector to obtain the weighted vector corresponding to the extended vector; Perform an alignment process on the space matching vector corresponding to the user preference vector and the weighted vector corresponding to the extended vector to obtain the target user preference vector and the target extended vector corresponding to the object to be recommended.

8. An object recommendation device, characterized in that, Includes: A first processing module, configured to process the user-side data and the attribute data corresponding to the object to be recommended respectively through a large language model to obtain user preference data and the extended data corresponding to the object to be recommended; A second processing module, configured to perform text encoding processing on the user preference data and the extended data corresponding to the to-be-recommended object respectively, so as to obtain a user preference vector and an extended vector corresponding to the to-be-recommended object; A third processing module, configured to process the user preference vector and the extended vector corresponding to the to-be-recommended object respectively based on a gating mechanism, so as to obtain a gating vector corresponding to the user preference vector and a gating vector corresponding to the extended vector; A fourth processing module, configured to perform spatial alignment processing on the gating vector corresponding to the user preference vector and the gating vector corresponding to the extended vector respectively, so as to obtain a target user preference vector and a target extended vector corresponding to the to-be-recommended object; A determination module, configured to determine a recommendation score corresponding to the to-be-recommended object according to the user-side data, the attribute data corresponding to the to-be-recommended object, the historical interaction context data corresponding to the to-be-recommended object, the target user preference vector, and the target extended vector corresponding to the to-be-recommended object; A fifth processing module, configured to determine target object data based on a preset recommendation score threshold and the recommendation score corresponding to the to-be-recommended object, and send the target object data to a target terminal device, so that the target object data is displayed on a current graphical user interface of the target terminal device.

9. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images