Object recommendation method and device, equipment and storage medium

Abstract

The embodiment of the invention provides an object recommendation method and device, equipment and a storage medium. The method comprises the steps that multiple pieces of first text information aiming at a target user and multiple pieces of second text information associated with a group of multiple to-be-recommended objects are obtained, the multiple pieces of first text information are user requirements of the target user described in a natural language, and the multiple pieces of second text information comprise description texts aiming at the group of to-be-recommended objects; determining a plurality of candidate objects recommended to the target user from the plurality of to-be-recommended objects based on the plurality of pieces of first text information and the plurality of pieces of second text information through a trained recommendation model; and determining one or more target objects recommended to the target user from the plurality of candidate objects based on the plurality of candidate objects and a predetermined condition. In this way, the user demand and the product can be comprehensively considered, so that the appropriate product is recommended to the user.

Description

Technical Field

[0001] Example embodiments of the present disclosure generally relate to the field of computer technology, and more particularly, to methods, devices, apparatuses, and storage media for object recommendation. Background Art

[0002] By understanding the needs of users, suitable products can be recommended to target users. When recommending suitable products to target users, it is usually necessary to consider the relationship between multiple products (for example, main-increase relationship, mutually exclusive relationship, etc.) and the matching between products and target user needs. Therefore, how to recommend suitable logistics products to target users based on specific scenarios and needs is worthy of attention and urgent. Summary of the invention

[0003] In a first aspect of the present disclosure, a method for object recommendation is provided. The method comprises: obtaining a plurality of first text information for a target user and a plurality of second text information associated with a group of multiple objects to be recommended, wherein the plurality of first text information is a user demand of the target user described in natural language, and the plurality of second text information comprises a description text for a group of objects to be recommended; determining a plurality of candidate objects recommended to the target user from a plurality of objects to be recommended based on the plurality of first text information and the plurality of second text information through a trained recommendation model; and determining one or more target objects recommended to the target user from the plurality of candidate objects based on the plurality of candidate objects and a predetermined condition.

[0004] In a second aspect of the present disclosure, a device for object recommendation is provided. The device includes: an acquisition module configured to acquire a plurality of first text information for a target user and a plurality of second text information associated with a group of multiple objects to be recommended, wherein the plurality of first text information is a user demand of the target user described in natural language, and the plurality of second text information includes description text for a group of objects to be recommended; a first determination module configured to determine a plurality of candidate objects recommended to the target user from a plurality of objects to be recommended based on the plurality of first text information and the plurality of second text information through a trained recommendation model; and a second determination module configured to determine one or more target objects recommended to the target user from a plurality of candidate objects based on the plurality of candidate objects and a predetermined condition.

[0005] In a third aspect of the present disclosure, an electronic device is provided. The electronic device includes at least one processing unit; and at least one memory, the at least one memory is coupled to the at least one processing unit and stores instructions for execution by the at least one processing unit, and when the instructions are executed by the at least one processing unit, the electronic device executes the method of the first aspect of the present disclosure.

[0006] In a fourth aspect of the present disclosure, a computer-readable storage medium is provided, wherein a computer program is stored on the computer-readable storage medium, and the computer program can be executed by a processor to perform the method according to the first aspect of the present disclosure.

[0007] In a fifth aspect of the present disclosure, a computer program product is provided, which includes computer executable instructions, wherein the computer executable instructions implement the method according to the first aspect of the present disclosure when executed by a processor.

[0008] It should be understood that the contents described in the summary of the present invention are not intended to limit the key features or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In the following, in conjunction with the accompanying drawings and with reference to the following detailed description, the above and other features, advantages and aspects of various implementations of the present disclosure will become more apparent. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, wherein:

[0010] Figure 1 A schematic diagram showing an example environment in which embodiments of the present disclosure can be implemented;

[0011] Figure 2 A flowchart of a process for object recommendation according to some embodiments of the present disclosure is shown;

[0012] Figure 3 A flowchart showing a process of determining a candidate object according to some embodiments of the present disclosure;

[0013] Figure 4 A schematic diagram showing a process of supply chain product recommendation according to some embodiments of the present disclosure;

[0014] Figure 5 A block diagram showing an apparatus for object recommendation according to some embodiments of the present disclosure; and

[0015] Figure 6 A block diagram of an electronic device is shown in which one or more embodiments of the present disclosure may be implemented. DETAILED DESCRIPTION

[0016] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0017] It should be noted that the acquisition, storage and application of user personal information involved in the technical solution of the present disclosure are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0018] In the description of the embodiments of the present disclosure, the term "including" and similar terms should be understood as open inclusion, that is, "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The term "some embodiments" should be understood as "at least some embodiments". Other explicit and implicit definitions may also be included below.

[0019] As used herein, the term "model" can learn the association between the corresponding input and output from the training data, so that after the training is completed, the corresponding output can be generated for a given input. The generation of the model can be based on machine learning technology. Deep learning is a machine learning algorithm that processes inputs and provides corresponding outputs by using multiple layers of processing units. In this article, "model" may also be referred to as "machine learning model", "machine learning network" or "network", and these terms are used interchangeably in this article.

[0020] Generally, machine learning can be roughly divided into three stages, namely the training stage, the testing stage, and the application stage (also called the inference stage). In the training stage, a given model can be trained using a large amount of training data, and the parameter values ​​are continuously updated iteratively until the model can obtain consistent inferences that meet the expected goals from the training data. Through training, the model can be considered to be able to learn the association from input to output (also called the mapping of input to output) from the training data. The parameter values ​​of the trained model are determined. In the testing stage, the test input is applied to the trained model to test whether the model can provide the correct output, thereby determining the performance of the model. In the application stage, the model can be used to process the actual input based on the parameter values ​​obtained from the training to determine the corresponding output.

[0021] As briefly mentioned above, when recommending suitable products to target users, it is usually necessary to consider the relationship between multiple products and the match between the products and the needs of the target users.

[0022] Usually, suitable products are recommended to target users manually. For example, the target users are corporate users, and the products are logistics supply chain products. Logistics supply chain products can include main products, value-added products, and product solutions. At the same time, different corporate users have different needs, and as the company develops, the needs of the same corporate user at different times are also different. However, when recommending suitable products to target users manually, it is difficult to take into account the relationship between multiple products and the matching between the products and the needs of target users, resulting in the recommended products failing to meet the needs of target users.

[0023] In order to at least partially solve the above problems and other potential problems that may exist in traditional solutions, the present disclosure provides a solution for object recommendation. According to various embodiments of the present disclosure, multiple first text information for a target user and multiple second text information associated with a group of multiple objects to be recommended are obtained, the multiple first text information is a user demand of the target user described in natural language, and the multiple second text information includes a description text for a group of objects to be recommended. Through a trained recommendation model, based on the multiple first text information and the multiple second text information, multiple candidate objects recommended to the target user are determined from the multiple objects to be recommended. And based on the multiple candidate objects and predetermined conditions, one or more target objects recommended to the target user are determined from the multiple candidate objects. In this way, user needs and products can be comprehensively considered to recommend suitable products to users.

[0024] Hereinafter, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings.

[0025] Figure 1 A schematic diagram of an example environment 100 in which embodiments of the present disclosure can be implemented is shown. In the environment 100, various client applications, such as product recommendation applications, etc., may be installed on the terminal device 110. The recommendation model 130 is configured to process the information of the object to be recommended, for example, to recommend the object to be recommended to the target user based on the information of the object to be recommended.

[0026] In some embodiments of the present disclosure, the recommendation model 130 used may be deployed on the remote device 120. The terminal device 110 may communicate with the remote device 130 (e.g., via network communication) to utilize the recommendation model 130 stored thereon to perform a model reasoning task, i.e., a task of recommending an object to be recommended to a target user based on the information of the object to be recommended. The information of the object to be recommended may be directly stored on the remote device 120, or collected by the terminal device 110 and sent to the remote device 120. In some embodiments, the recommendation model 130 may also be partially or completely deployed locally on the terminal device 110, and run by the terminal device 110 to perform the task of recommending an object to be recommended to a target user. The embodiments of the present disclosure do not impose specific limitations on this.

[0027] The terminal device 110 can be any type of mobile terminal, fixed terminal or portable terminal, including a mobile phone, a desktop computer, a laptop computer, a notebook computer, a netbook computer, a tablet computer, a media computer, a multimedia tablet, a personal communication system (PCS) device, a personal navigation device, a personal digital assistant (PDA), an audio / video player, a digital camera / camcorder, a positioning device, a television receiver, a radio broadcast receiver, an e-book device, a gaming device, or any combination of the foregoing, including accessories and peripherals of these devices or any combination thereof. The remote device 120 can include, for example, a computing system / server, such as a mainframe, an edge computing node, a computing device in a cloud environment, a virtual machine, and the like. Although a single device is shown, the remote device 130 can include multiple physical devices. In addition, although only a single terminal device 110 is shown, the remote device 120 or the model 130 deployed therein can be accessed by multiple terminal devices 110 to provide the reasoning capability of the model 130.

[0028] It should be understood that the structure and functionality of environment 100 are described for exemplary purposes only and do not imply any limitation on the scope of the present disclosure.

[0029] Figure 2 FIG. 2 is a flowchart of a process 200 for object recommendation according to some embodiments of the present disclosure. Figure 2 In the embodiment, the process 200 can be implemented in the terminal device 110 or the remote device 120. For the convenience of description, the following description will be given by taking the terminal device 110 executing the process 200 as an example. It should be understood that a similar process can also be executed by the remote device 120. Figure 1 The solution for object recommendation provided by the present disclosure is introduced in detail.

[0030] In block 210 , the terminal device 110 obtains a plurality of first text information for a target user and a plurality of second text information associated with a group of a plurality of objects to be recommended.

[0031] The plurality of first text information is a user demand of a target user described in natural language. The plurality of second text information includes a description text for a group of objects to be recommended. In some embodiments, the plurality of first text information and the plurality of second text information can be obtained based on the user's input data. The input data can include text information collected by the terminal device 110, and can also include input data received by the terminal device 110, etc.

[0032] It should be understood that the above-mentioned input data is related to the application scenario, and the input data generated in different scenarios may also be different.

[0033] For example, in the recommendation scenario of logistics supply chain products, the first text information may be feedback information for the target user. Taking the target user as an enterprise user as an example, the feedback information may include but is not limited to: the current needs of the enterprise, the business conditions of the enterprise, the order status of the enterprise, etc.

[0034] In some embodiments, in order to reflect the changes in user needs of target users, multiple pieces of feedback information within a predetermined time period can be used as multiple first text messages, and the multiple pieces of feedback information have a time sequence relationship (for example, a sequential relationship). Thus, the changes in the needs of enterprise users within the predetermined time period can be reflected through multiple first text messages. For example, the predetermined time period can be one year.

[0035] For another example, in the above scenario, the second text information may include product introduction of logistics supply chain products (for example, product definition, target groups, etc.), product business attributes (for example, weight, volume, temperature control type, etc.), product transaction attributes (for example, collection method, settlement method, etc.), product claims attributes and other information.

[0036] In some embodiments, the second text information may include long text type text information and / or discrete text information.

[0037] In some embodiments, the terminal device 110 may collect one or more raw data including user needs and associated with the object to be recommended, for example, the raw data may include data of types such as images, voices, or texts, and extract text from the raw data of types such as images, voices, or texts, thereby obtaining the first text information and the second text information. For example, the first text information and the second text information may be extracted from the raw data of types such as images, voices, or texts by a trained extraction network model. It should be understood that the trained extraction network model may adopt any network model in the relevant technology, and the embodiments of the present disclosure are not limited thereto.

[0038] In block 220 , the terminal device 110 determines, through the trained recommendation model 130 , a plurality of candidate objects to be recommended to the target user from a plurality of objects to be recommended based on the plurality of first text information and the plurality of second text information.

[0039] After obtaining the plurality of first text information and the plurality of second text information, the plurality of first text information and the plurality of second text information may be used as inputs of the recommendation model 130 , thereby outputting a plurality of candidate objects recommended to the target user from the plurality of objects to be recommended.

[0040] Combine the following Figure 3 The process of determining multiple candidate objects to be recommended to a target user from multiple objects to be recommended in the embodiment of the present disclosure is introduced in detail. Figure 3FIG. 3 is a flowchart of a process 300 for determining a candidate object according to some embodiments of the present disclosure. The process 300 may be performed by the terminal device 110 .

[0041] In some embodiments, the recommendation model 130 may include two sub-models. One of the sub-models (hereinafter referred to as the "first sub-model") may be a model capable of processing sequence data. For example, a Long Short-Term Memory (LSTM) model, a Gated Recurrent Unit (GRU), etc. The other sub-model (hereinafter referred to as the "second sub-model") may be a Multilayer Perceptron (MLP).

[0042] It should be understood that by using multiple first text information and multiple second text information as inputs of the recommendation model 130, that is, using multiple first text information as inputs of the first sub-model and using multiple second text information as inputs of the second sub-model, multiple candidate objects to be recommended to the target user can be determined from multiple objects to be recommended based on the output of the first sub-model and the output of the second sub-model.

[0043] Specifically, see Figure 3 In box 310, the terminal device 110 determines, through the trained first sub-model and based on the multiple first text information, first feature representations corresponding to the multiple first text information, where the first feature representations can indicate user needs of the target user.

[0044] After obtaining the plurality of first text information, the terminal device 110 inputs the plurality of first text information into the first sub-model, thereby outputting first feature representations corresponding to the plurality of first text information.

[0045] In some embodiments, the first feature representation may include a third feature representation indicating changes in user needs and a fourth feature representation indicating a user profile of the target user. For example, the third feature representation and the fourth feature representation may be concatenated to obtain the first feature representation.

[0046] Specifically, first, the terminal device 110 can determine the third feature representation of the target user based on the multiple first text information and the temporal relationship between the first text information through the trained first sub-model. For example, the multiple first text information is sequentially input into the first sub-model based on the temporal relationship (e.g., the order relationship) between the first text information, so as to obtain the third feature representation of the target user.

[0047] In some embodiments, before the multiple first text information is input into the first sub-model, the first text information may be preprocessed. For example, the preprocessing of the first text information may include removing punctuation marks through stop words, performing word segmentation on the first text information, and the like. In some embodiments, after the multiple first text information is preprocessed, the multiple first text information may be mapped into text vectors, and then the text vectors may be used as input to the first sub-model. For example, the multiple first text information may be mapped into text vectors through a language representation (Bidirectional Encoder Representations from Transformers, BERT) model.

[0048] Furthermore, the terminal device 110 may determine the third text information corresponding to each of the plurality of first text information based on the plurality of first text information and the predetermined prompt word. For example, the plurality of first text information may be rewritten by the predetermined prompt word.

[0049] For example, the predetermined prompt words may include, but are not limited to: business conditions, current needs, current satisfaction, etc. After rewriting the multiple first text information, key information can be extracted from the multiple first text information in a question-answer pair manner through a machine learning model and a given enumeration value, thereby obtaining multiple third text information.

[0050] As an example, the third text information may be expressed as follows:

[0051] {“question”:“The customer’s visit feedback is: {feedback content}, and the feedback content keywords are generated in the following format: Business status: {business status enumeration value (good, general, abnormal)}; Current demand: {current demand enumeration value (cold chain, medicine, large items, etc.)}; Current satisfaction: {current satisfaction enumeration value (satisfied, dissatisfied)}…}.

[0052] It should be understood that the third text information mentioned above is only an example, and any other representation method may be used according to specific applications and scenarios.

[0053] In some embodiments, the above-mentioned machine learning model can be trained by historical feedback information and labels of the historical feedback information. For example, the operating conditions, current demand, current satisfaction, etc. in the historical feedback information can be labeled, and then the above-mentioned machine learning model can be trained by the labeled historical feedback information.

[0054] Furthermore, the terminal device 110 may determine a fourth feature representation of the target user based on the third text information through the trained first sub-model.

[0055] Specifically, the terminal device 110 can collect the user portrait of the target user within a predetermined period of time (for example, one month) in the past of each third text message based on the respective times of multiple third text messages as the input of the first sub-model, thereby outputting the fourth feature representation of the target user.

[0056] For example, the user portrait of the target user may include but is not limited to: number of user orders, number of products used, products used, number of customer complaints, claims settlement, number of visits, etc.

[0057] In some embodiments, before inputting the plurality of third text messages and the user portrait corresponding to each third text message into the first sub-model, the third text messages and the user portrait corresponding to each third text message may also be preprocessed, for example, mapping the third text messages and the user portrait corresponding to each third text message into a text vector.

[0058] Since the third text information and the user portrait corresponding to each third text information include long text data and discrete data, during the mapping process, a multi-hot method can be used to map the discrete data, normalize the long text data, and concatenate the text vectors obtained by the two methods. Then, the text vectors are input into the first sub-model to obtain the fourth feature representation of the target user.

[0059] Finally, the terminal device 110 may determine the first feature representation based on the third feature representation and the fourth feature representation.

[0060] For example, the terminal device 110 may concatenate the third feature representation and the fourth feature representation to obtain the first feature representation.

[0061] In box 320, the terminal device 110 determines second feature representations corresponding to multiple objects to be recommended based on the second text information through the trained second sub-model.

[0062] In some embodiments, the second text information includes long text data and discrete data. The terminal device 110 can extract a first vector representation of the long text data and a second vector representation of the discrete data, splice the first vector representation and the second vector representation, and finally determine the second feature representation of each of the multiple objects to be recommended based on the spliced ​​first vector representation and the second vector representation through the trained second sub-model.

[0063] In some embodiments, when the terminal device 110 extracts the first vector representation of long text data (e.g., product introduction), the long text data can be used as input to a language representation (Bidirectional Encoder Representations from Transformers, BERT) model to output the first vector representation of the long text data.

[0064] In some embodiments, when the terminal device 110 extracts the second vector representation of discrete data, the discrete data can be mapped to the second vector representation in a multi-hot manner.

[0065] Next, the training process of the recommendation model 130 is introduced.

[0066] In some embodiments, the recommendation model 130 is trained based on multiple training samples. Each training sample includes multiple third text information for the user, multiple fourth text information for multiple products, and a sample label, the sample label indicates the user's contract status for multiple products, the third text information is used to describe the user's needs, and the fourth text information is used to describe multiple products. It should be understood that the detailed process of training can be implemented in any manner in the relevant technology, which will not be described in detail here.

[0067] In some embodiments, the first sub-model and the second sub-model in the recommendation model 130 can be trained separately. Since the types of products do not change much in a short period of time, the second sub-model can be trained once and used directly without multiple training. In the case of changes in products, the second sub-model can be trained.

[0068] In block 330 , the terminal device 110 determines a plurality of candidate objects to be recommended to the target user from the plurality of objects to be recommended based on the first feature representation and the second feature representation.

[0069] Specifically, first, the terminal device 110 may determine the similarity between the first feature representation and the second feature representation. Further, based on the similarity, the multiple objects to be recommended are sorted. Finally, based on the sorted multiple objects to be recommended, multiple candidate objects to be recommended to the target user are determined from the multiple objects to be recommended.

[0070] In some embodiments, the terminal device 110 may use the following formula to determine the similarity between the first feature representation and the second feature representation:

[0071]

[0072] Among them, y Q represents the first feature representation, y D represents the second feature representation, and R(Q, D) represents the similarity.

[0073] In some embodiments, multiple recommended objects may be sorted in descending order of similarity values, and the recommended objects with a predetermined ranking may be selected as candidate objects. For example, the top five recommended objects may be selected as candidate objects.

[0074] Continue to see Figure 2 In block 230 , the terminal device 110 determines one or more target objects to be recommended to the target user from the multiple candidate objects based on the multiple candidate objects and the predetermined condition.

[0075] In some embodiments, the predetermined conditions may include at least one of the following: a first condition for indicating that the target user has purchased products among multiple objects to be recommended, a second condition for indicating a mutually exclusive relationship between multiple objects to be recommended, or a third condition for indicating a master-slave relationship between multiple objects to be recommended.

[0076] It should be understood that in a mutually exclusive relationship, for example, if you buy product a, you cannot use product b. In this case, product a and product b have a mutually exclusive relationship. In a main-additive relationship, value-added products cannot usually be purchased separately. Therefore, if there are value-added products in the candidate objects, it is necessary to filter out the value-added products that have not purchased the corresponding main products based on the products that the target user has purchased.

[0077] To further explain the solutions for object recommendation provided by some embodiments of the present disclosure, the following description is made in conjunction with an example of supply chain product recommendation. Of course, it can be understood that the embodiments of the present disclosure are not limited to this application scenario.

[0078] Figure 4 A schematic diagram of a process 400 for supply chain product recommendation according to some embodiments of the present disclosure is shown. The process 400 may be implemented by the terminal device 110, for example.

[0079] In block 411 , the terminal device 110 maps the user feedback information into a text vector through a Bert model.

[0080] In box 412, the terminal device 110 inputs the text vector obtained by mapping the user feedback information into the Lstm model, and outputs the vector representation of the user demand.

[0081] In block 421 , the terminal device 110 rewrites the user feedback information based on the prompt word.

[0082] In box 422, the terminal device 110 inputs the rewritten user feedback information into the language model and outputs a user portrait.

[0083] In box 423, the terminal device 110 inputs the user portrait into the Lstm model and outputs the vector representation of the user portrait.

[0084] In box 431, the terminal device 110 maps the user portrait into a text vector through the Bert model.

[0085] In box 432, the terminal device 110 inputs the text vector obtained by mapping the user portrait into the Mlp model and outputs the vector representation of the product portrait.

[0086] In box 440, the terminal device 110 concatenates the vector representation of the user demand and the vector representation of the user portrait to obtain a vector representation of the change in user demand.

[0087] In box 450, the terminal device 110 calculates the similarity between the vector representation of the product portrait and the vector representation of the user demand change.

[0088] In block 461 , the terminal device 110 filters the purchased products.

[0089] In block 462 , the terminal device 110 filters products that have a mutually exclusive relationship.

[0090] In block 463 , the terminal device 110 filters products having a primary-increase relationship.

[0091] At block 470 , the terminal device 110 determines the top five products.

[0092] At block 480 , the terminal device 110 presents the product to the user.

[0093] It should be understood that the specific implementation of process 400 can refer to the corresponding implementation in the aforementioned embodiment, which will not be repeated here.

[0094] The comparison of the success rate of recommending products to users based on the above process 400 and recommending products to users based on manual experience is shown in Table 1 below:

[0095] Table 1

[0096]

[0097] The experimental group is the users who recommend products to the users based on the above process 400, and the control group is the users who recommend products to the users based on manual experience.

[0098] Based on the characteristics of user industry and order quantity, the users are divided into experimental group and control group. The results of users' negative feedback are shown in Table 2 below:

[0099] Table 2

[0100]

[0101] In summary, the solution for object recommendation proposed in the present disclosure can comprehensively consider user needs and product conditions, thereby recommending suitable products to users.

[0102] Figure 5 A block diagram of an apparatus 500 for object recommendation according to some embodiments of the present disclosure is shown. The apparatus 500 may be implemented in the remote device 120 and / or the terminal device 110. Each module / component in the apparatus 500 may be implemented by hardware, software, firmware, or any combination thereof.

[0103] The device 500 includes an acquisition module 510, which is configured to acquire multiple first text information for a target user and multiple second text information associated with a group of multiple objects to be recommended, the multiple first text information is a user demand of the target user described in natural language, and the multiple second text information includes description text for a group of objects to be recommended. The device 500 also includes a first determination module 520, which is configured to determine multiple candidate objects to be recommended to the target user from multiple objects to be recommended based on the multiple first text information and the multiple second text information through a trained recommendation model. The device 500 also includes a second determination module 530, which is configured to determine one or more target objects to be recommended to the target user from multiple candidate objects based on the multiple candidate objects and predetermined conditions.

[0104] In some embodiments, the predetermined condition includes at least one of the following: a first condition for indicating that an object among multiple objects to be recommended has been associated with the target user, a second condition for indicating a mutually exclusive relationship between multiple objects to be recommended, or a third condition for indicating a master-slave relationship between multiple objects to be recommended.

[0105] In some embodiments, the first determination module 520 is also configured to determine, through a trained recommendation model and based on multiple first text information, first feature representations corresponding to multiple first text information, the first feature representations indicating user needs; determine, through a trained recommendation model and based on second text information, second feature representations corresponding to multiple objects to be recommended, the second feature representations indicating feature portraits corresponding to multiple objects to be recommended; and determine, from the multiple objects to be recommended, based on the first feature representations and the second feature representations, multiple candidate objects to be recommended to the target user.

[0106] In some embodiments, the first determination module 520 is further configured to determine the similarity between the first feature representation and the second feature representation; sort the multiple objects to be recommended based on the similarity; and determine multiple candidate objects to be recommended to the target user from the multiple objects to be recommended based on the sorted multiple objects to be recommended.

[0107] In some embodiments, there is a temporal relationship between multiple first text information, and the first determination module 520 is also configured to determine the third feature representation of the target user based on the multiple first text information and the temporal relationship through a trained recommendation model, the third feature representation indicating changes in user needs; determine the third text information corresponding to each of the multiple first text information based on the multiple first text information and predetermined prompt words; determine the fourth feature representation of the target user based on the third text information through a trained recommendation model, the fourth feature representation indicating a user portrait of the target user; and determine the first feature representation based on the third feature representation and the fourth feature representation.

[0108] In some embodiments, the second text information includes long text data and discrete data, and the first determination module 520 is also configured to extract a first vector representation of the long text data and a second vector representation of the discrete data; concatenate the first vector representation and the second vector representation; and determine, through a trained recommendation model, the second feature representation of each of a plurality of objects to be recommended based on the concatenated first vector representation and the second vector representation.

[0109] In some embodiments, the recommendation model is trained based on multiple training samples, each training sample includes multiple third text information for the user, multiple fourth text information for multiple objects and sample labels, the sample labels indicate the user's contract status for multiple objects, the third text information is used to describe the user's needs, and the fourth text information is used to describe multiple objects.

[0110] The units included in the device 500 can be implemented in various ways, including software, hardware, firmware, or any combination thereof. In some embodiments, one or more units can be implemented using software and / or firmware, such as machine executable instructions stored on a storage medium. In addition to or as an alternative to machine executable instructions, some or all of the units in the device 500 can be implemented at least in part by one or more hardware logic components. As an example and not limitation, exemplary types of hardware logic components that can be used include field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0111] Figure 6 1 shows a block diagram of an electronic device 600 in which one or more embodiments of the present disclosure may be implemented. It should be understood that Figure 6 The electronic device 600 shown is merely exemplary and should not constitute any limitation on the functionality and scope of the embodiments described herein. Figure 6 The electronic device 600 shown can be used to implement Figure 1 remote device 120 and / or terminal device 110.

[0112] like Figure 6 As shown, the electronic device 600 is in the form of a general electronic device. The components of the electronic device 600 may include, but are not limited to, one or more processors or processing units 610, a memory 620, a storage device 630, one or more communication units 640, one or more input devices 650, and one or more output devices 660. The processing unit 610 may be an actual or virtual processor and is capable of performing various processes according to a program stored in the memory 620. In a multi-processor system, multiple processing units execute computer executable instructions in parallel to improve the parallel processing capability of the electronic device 600.

[0113] The electronic device 600 typically includes a plurality of computer storage media. Such media may be any available media accessible to the electronic device 600, including but not limited to volatile and non-volatile media, removable and non-removable media. The memory 620 may be a volatile memory (e.g., a register, a cache, a random access memory (RAM)), a non-volatile memory (e.g., a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), flash memory), or some combination thereof. The storage device 630 may be a removable or non-removable medium, and may include a machine-readable medium, such as a flash drive, a disk, or any other medium, which may be capable of being used to store information and / or data (e.g., training data for training) and may be accessed within the electronic device 600.

[0114] The electronic device 600 may further include additional removable / non-removable, volatile / non-volatile storage media. Figure 6 As shown in , a disk drive for reading or writing from a removable, non-volatile disk (e.g., a "floppy disk") and an optical drive for reading or writing from a removable, non-volatile optical disk may be provided. In these cases, each drive may be connected to the bus (not shown) by one or more data media interfaces. The memory 620 may include a computer program product 625 having one or more program modules that are configured to perform various methods or actions of various embodiments of the present disclosure.

[0115] The communication unit 640 implements communication with other electronic devices through a communication medium. Additionally, the functions of the components of the electronic device 600 can be implemented with a single computing cluster or multiple computing machines that can communicate through a communication connection. Therefore, the electronic device 600 can operate in a networked environment using a logical connection with one or more other servers, a network personal computer (PC), or another network node.

[0116] The input device 650 may be one or more input devices, such as a mouse, a keyboard, a tracking ball, etc. The output device 660 may be one or more output devices, such as a display, a speaker, a printer, etc. The electronic device 600 may also communicate with one or more external devices (not shown) through the communication unit 640 as needed, such as a storage device, a display device, etc., communicate with one or more devices that allow a user to interact with the electronic device 600, or communicate with any device that allows the electronic device 600 to communicate with one or more other electronic devices (e.g., a network card, a modem, etc.). Such communication may be performed via an input / output (I / O) interface (not shown).

[0117] According to an exemplary implementation of the present disclosure, a computer-readable storage medium is provided, on which one or more computer instructions are stored, wherein the one or more computer instructions are executed by a processor to implement the method described above. According to an exemplary implementation of the present disclosure, a computer program product is also provided, which is tangibly stored on a non-transitory computer-readable medium and includes computer-executable instructions, and the computer-executable instructions are executed by a processor to implement the method described above.

[0118] Various aspects of the present disclosure are described herein with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products implemented according to the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer-readable program instructions.

[0119] These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processing unit of the computer or other programmable data processing device, a device that implements the functions / actions specified in one or more boxes in the flowchart and / or block diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and / or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0120] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0121] The flow chart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to multiple implementations of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of a module, program segment or instruction includes one or more executable instructions for realizing the logical function of the specification. In some implementations as replacements, the function marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be realized by a special hardware-based system that performs the function or action of the specification, or can be realized by a combination of special hardware and computer instructions.

[0122] The above descriptions of various implementations of the present disclosure are exemplary, non-exhaustive, and not limited to the disclosed implementations. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described implementations. The selection of terms used herein is intended to best explain the principles of the implementations, practical applications, or improvements to the technology in the marketplace, or to enable other persons of ordinary skill in the art to understand the implementations disclosed herein.

Claims

1. A method for object recommendation, comprising: Acquire a plurality of first text information for a target user and a plurality of second text information associated with a group of a plurality of objects to be recommended, wherein the plurality of first text information is user needs of the target user described in natural language, and the plurality of second text information includes description text for the group of objects to be recommended; Determining, by means of a trained recommendation model, a plurality of candidate objects to be recommended to the target user from the plurality of objects to be recommended based on the plurality of first text information and the plurality of second text information; as well as Based on the multiple candidate objects and a predetermined condition, one or more target objects recommended to the target user are determined from the multiple candidate objects.

2. The method according to claim 1, wherein the predetermined condition comprises at least one of the following: A first condition for indicating an object that is associated with the target user among the plurality of objects to be recommended, A second condition for indicating a mutually exclusive relationship between the plurality of objects to be recommended, or The third condition is used to represent the master-slave relationship between the multiple objects to be recommended.

3. The method according to claim 1, wherein determining the plurality of candidate objects to be recommended to the target user from the plurality of objects to be recommended comprises: Determining, by means of the trained recommendation model, first feature representations corresponding to the plurality of first text information based on the plurality of first text information, wherein the first feature representations indicate the user demand; Determining, by means of the trained recommendation model, second feature representations corresponding to the plurality of objects to be recommended based on the second text information, wherein the second feature representations indicate feature portraits corresponding to the plurality of objects to be recommended; as well as Based on the first feature representation and the second feature representation, the multiple candidate objects to be recommended to the target user are determined from the multiple objects to be recommended.

4. The method according to claim 3, wherein determining the plurality of candidate objects based on the first feature representation and the second feature representation comprises: determining a similarity between the first feature representation and the second feature representation; Based on the similarities, sorting the multiple objects to be recommended; as well as Based on the sorted plurality of objects to be recommended, the plurality of candidate objects to be recommended to the target user are determined from the plurality of objects to be recommended.

5. The method according to claim 3, wherein the plurality of first text information have a temporal relationship therebetween, and wherein determining the first feature representation comprises: Determining, by means of the trained recommendation model, a third feature representation of the target user based on the plurality of first text information and the temporal relationship, wherein the third feature representation indicates a change in the user's needs; Based on the multiple first text messages and the predetermined prompt word, determining the third text messages corresponding to each of the multiple first text messages; Determining, by means of the trained recommendation model and based on the third text information, a fourth feature representation of the target user, wherein the fourth feature representation indicates a feature portrait of the target user; and The first feature representation is determined based on the third feature representation and the fourth feature representation.

6. The method according to claim 3, wherein the second text information includes long text data and discrete data, and wherein determining the second feature representations corresponding to the plurality of objects to be recommended includes: Extracting a first vector representation of the long text data and a second vector representation of the discrete data; concatenating the first vector representation and the second vector representation; as well as The second feature representation of each of the plurality of objects to be recommended is determined by the trained recommendation model based on the concatenated first vector representation and the second vector representation.

7. The method according to claim 1, wherein the recommendation model is trained based on multiple training samples, each training sample includes multiple third text information for users, multiple fourth text information for multiple objects and sample labels, the sample labels indicate the user's contract status for the multiple objects, the third text information is used to describe the user's needs, and the fourth text information is used to describe the multiple objects.

8. A device for object recommendation, comprising: an acquisition module, configured to acquire a plurality of first text information for a target user and a plurality of second text information associated with a group of multiple objects to be recommended, wherein the plurality of first text information is user needs of the target user described in natural language, and the plurality of second text information includes description text for the group of objects to be recommended; A first determination module is configured to determine, through a trained recommendation model, a plurality of candidate objects recommended to the target user from the plurality of objects to be recommended based on the plurality of first text information and the plurality of second text information; as well as The second determination module is configured to determine one or more target objects recommended to the target user from the multiple candidate objects based on the multiple candidate objects and a predetermined condition.

9. An electronic device, comprising: at least one processing unit; as well as At least one memory, the at least one memory being coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit, the instructions causing the electronic device to perform the method according to any one of claims 1 to 7 when executed by the at least one processing unit.

10. A computer-readable storage medium having a computer program stored thereon, wherein the computer program can be executed by a processor to implement the method according to any one of claims 1 to 7.

11. A computer program product comprising computer executable instructions, wherein the computer executable instructions, when executed by a processor, implement the method according to any one of claims 1 to 7.

Images