A method for expanding interest portrait, an interest portrait expansion device and a storage medium

Abstract

The embodiment of the present application discloses an interest portrait expansion method, an interest portrait expansion device and a storage medium, which are used to expand a second interest portrait based on the user's existing first interest portrait, so that the user's interest portrait is richer, the description of the user's interest status is more comprehensive and accurate, and products that can enhance the user experience are recommended to the user based on the enriched second interest portrait. The interest portrait expansion method provided by the embodiment of the present application includes: obtaining a first interest portrait; training the first interest portrait based on a preset training model to obtain a first similarity relationship, the first similarity relationship being the association relationship between the first interest portrait and the second interest portrait; determining the second interest portrait based on the first interest portrait and the first similarity relationship.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of computer technology, and specifically to an interest portrait expansion method, an interest portrait expansion device, and a storage medium. Background Art

[0002] At present, user interest portraits have become one of the functions that various industries are vying for. The main function of interest portraits is to help businesses or customers accurately understand user needs, hobbies, interests and other aspects.

[0003] In order to enrich the user's interest profile, there are two main methods commonly used in the industry: the introduction of externally available data sources and the use of knowledge graphs. Figure 1a , which is a flowchart of the introduction of externally available data sources, from Figure 1a As can be seen from the introduction of external data sources, the main purpose is to introduce the user's behavior data in external products such as videos, browsers, and Tiantian Kuaibao, calculate the user's external portrait based on these external data, and then recommend other content that the user is interested in based on the external portrait and make it effective online. You can also refer to Figure 1b , is a schematic diagram of the use of knowledge graphs, from Figure 1b It can be seen that the knowledge graph uses the connection between interest portraits in the graph, such as the connection between face recognition, feature expression, reinforcement learning, natural language processing, etc. and deep learning applications, etc., to analyze the content of the interest itself, expand the user's other interests, and then make recommendations and take effect online. However, the introduction of external data sources often makes it difficult to enrich the interest portraits due to the small intersection between the user's interest portraits and the poor quality of individual data, which makes it difficult for other recommended products to meet the actual needs of users; and the knowledge graph expansion method cannot reflect the unique characteristics of external products due to the slow update cycle of the relatively static graph, so it is also difficult to enrich the interest portraits, which makes it difficult for other recommended products to meet the actual needs of users.

[0004] Therefore, how to enrich and expand the user's interest portrait is an urgent problem that needs to be solved. Summary of the invention

[0005] The embodiments of the present application provide an interest portrait expansion method, an interest portrait expansion device and a storage medium, which are used to expand a second interest portrait based on the user's existing first interest portrait, so that the user's interest portrait is richer, the description of the user's interest status is more comprehensive and accurate, and based on the enriched second interest portrait, products that can enhance the user experience are recommended to the user.

[0006] In a first aspect, an embodiment of the present application provides a method for expanding an interest portrait, including:

[0007] Get the first interest portrait;

[0008] Training the first interest portrait based on a preset training model to obtain a first similarity relationship, where the first similarity relationship is an association relationship between the first interest portrait and the second interest portrait;

[0009] The second interest portrait is determined based on the first interest portrait and the first similarity relationship.

[0010] Optionally, in combination with the first aspect, in a first possible implementation manner, training the first interest portrait based on a preset training model to obtain a first similarity relationship includes:

[0011] Determine N interest words in the first interest portrait, where N is an integer greater than or equal to 0;

[0012] Training at least two of the N interest words based on the preset training model to obtain a latent vector corresponding to each of the interest words;

[0013] Determine a cosine similarity value between latent vectors corresponding to each interest word, wherein the cosine similarity value is used to represent the first similarity relationship.

[0014] Optionally, in combination with the first possible implementation of the first aspect, in a second possible implementation, the method further includes:

[0015] One-hot encoding is performed on the N interest words.

[0016] Optionally, in combination with the first aspect and the first to second possible implementations, in a third possible implementation, determining the second interest portrait based on the first interest portrait and the first similarity relationship includes:

[0017] Determining an interest score corresponding to the first interest portrait;

[0018] The second interest portrait is determined based on the interest score corresponding to the first interest portrait and the first similarity relationship.

[0019] Optionally, in combination with the first aspect and the first to third possible implementations, in a fourth possible implementation, obtaining the first interest portrait includes:

[0020] User behavior data is obtained, where the user behavior data includes a first interest portrait.

[0021] In a second aspect, an embodiment of the present application provides an interest portrait expansion device, which may include:

[0022] An acquiring unit, used for acquiring a first interest portrait;

[0023] A training unit, configured to train the first interest portrait obtained by the acquisition unit based on a preset training model to obtain a first similarity relationship, wherein the first similarity relationship is an association relationship between the first interest portrait and the second interest portrait;

[0024] A determination unit is used to determine the second interest portrait based on the first interest portrait obtained by the acquisition unit and the first similarity relationship obtained by the training unit.

[0025] Optionally, in combination with the second aspect, in a first possible implementation manner, the training unit may include:

[0026] A first determination module, configured to determine N interest words in the first interest portrait, where N is an integer greater than or equal to 0;

[0027] A training module, configured to train at least two of the N interest words determined by the first determination module based on the preset training model to obtain a latent vector corresponding to each interest word;

[0028] The first determination module is used to determine the cosine similarity value between the latent vectors corresponding to each interest word, and the cosine similarity value is used to represent the first similarity relationship.

[0029] Optionally, in combination with the first possible implementation method of the second aspect above, in a second possible implementation method, the interest portrait expansion device may further include: an encoding unit for one-hot encoding the N interest words.

[0030] Optionally, in combination with the second aspect and the first to second possible implementations of the second aspect, in a third possible implementation, the determining unit includes:

[0031] A second determination module, used to determine the interest score corresponding to the first interest portrait;

[0032] The second determination module is used to determine the second interest portrait based on the interest score corresponding to the first interest portrait and the first similarity relationship.

[0033] Optionally, in combination with the above-mentioned second aspect and the first to third possible implementation methods of the second aspect, in a fourth possible implementation method, the acquisition unit is used to obtain user behavior data, and the user behavior data includes a first interest portrait.

[0034] In the third aspect, an embodiment of the present application provides an interest portrait expansion device, comprising: a processor and a memory; the memory is used to store program instructions, and when the interest portrait expansion device is running, the processor executes the program instructions stored in the memory to enable the client to execute the interest portrait expansion method such as the first aspect mentioned above or any possible implementation method of the first aspect.

[0035] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, comprising instructions, which, when executed on a computer, enables the computer to execute a method as in the first aspect or any possible implementation of the first aspect.

[0036] In a fifth aspect, an embodiment of the present application provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute a method as in the first aspect or any possible implementation of the first aspect.

[0037] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0038] In an embodiment of the present application, the first interest portrait is trained by using a preset training model to obtain a first similarity relationship, thereby combining the first similarity relationship with the first interest portrait, and expanding the second interest portrait based on the user's existing first interest portrait, so that the user's interest portrait is richer, and the description of the user's interest status is more comprehensive and accurate, and products that can enhance the user experience are recommended to the user based on the enriched second interest portrait. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1a It is a flowchart of the introduction of externally available data sources;

[0041] Figure 1b This is a schematic diagram of the use of knowledge graph;

[0042] Figure 1c is a schematic diagram of a scenario of a service channel provided in an embodiment of the present application;

[0043] Figure 2 is a schematic diagram of an embodiment of the method for expanding interest portraits provided in this embodiment;

[0044] Figure 3It is a logic diagram for determining a latent vector provided in an embodiment of the present application;

[0045] Figure 4 is a schematic diagram of a second interest portrait expanded in an embodiment of the present application;

[0046] Figure 5 It is a schematic diagram of an embodiment of an interest portrait expansion device provided in an embodiment of the present application;

[0047] Figure 6 It is a schematic diagram of the hardware structure of the communication device in the embodiment of the present application. DETAILED DESCRIPTION

[0048] The embodiments of the present application provide an interest portrait expansion method, an interest portrait expansion device and a storage medium, which are used to expand a second interest portrait based on the user's existing first interest portrait, so that the user's interest portrait is richer, the description of the user's interest status is more comprehensive and accurate, and based on the enriched second interest portrait, products that can enhance the user experience are recommended to the user.

[0049] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0050] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices. The naming or numbering of steps in this application does not mean that the steps in the method flow must be executed in the time / logical sequence indicated by the naming or numbering. The process steps that have been named or numbered can change the execution order according to the technical purpose to be achieved, as long as the same or similar technical effects can be achieved.

[0051] Artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines so that machines have the functions of perception, reasoning and decision-making.

[0052] Artificial intelligence technology is a comprehensive discipline that covers a wide range of fields, including both hardware-level and software-level technologies. The basic technologies of artificial intelligence generally include sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics and other technologies. Artificial intelligence software technology mainly includes computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning (DL).

[0053] Among them, machine learning (ML) is a multi-disciplinary interdisciplinary subject, involving probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and other disciplines. It specializes in studying how computers simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications are spread across all areas of artificial intelligence. Machine learning and deep learning usually include artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and self-learning. With the development of machine learning technology, machine learning is applied in more and more scenarios. For example, for each point of interest in an interest portrait, the corresponding latent vector can be determined based on machine learning.

[0054] In the process of learning the corresponding latent vector, the latent vector calculation tool Interest To Vector (Inte2Vec) is often used. Inte2Vec can be efficiently trained on millions of dictionaries and hundreds of millions of data sets; the training result is the latent vector. The latent vector can well measure the similarity between the interest words in each interest portrait. After obtaining the similarity, the similarity between the interest words in each interest portrait can be determined based on calculation formulas such as cosine similarity. In this way, after obtaining the similarity relationship corresponding to the similarity, the enriched interest portrait can be determined based on the similarity relationship.

[0055] It should be understood that the method proposed in the embodiment of the present application is mainly used in the scenario of personalized recommendation, and is used to recommend products such as articles and videos to users. Figure 1c , is a schematic diagram of a scenario of a service channel provided in an embodiment of the present application. Figure 1c It can be seen that multiple business channels can be included, and each business channel includes multiple candidate recommended articles. For example, in the "Recommendation" channel, it can include but is not limited to "*****, the increase in 2019 may be less than 5%?", "***** enters the critical moment: these two requirements are the key", "Attention! ***** will have new changes, let you understand it with a picture", etc. The business channels described above include but are not limited to "News", "Entertainment", "Video", etc. In actual applications, other channels such as "Beijing" can also be included, which will not be limited in this application.

[0056] In order to enrich the user's interest profile, the commonly used methods in the industry mainly include two types: the introduction of external available data sources and the use of knowledge graphs. At present, the main method for introducing external data sources is to introduce the user's external product behavior data, calculate the user's external profile based on these external data, and then recommend other content of interest to the user based on the external profile. The knowledge graph method uses the connection relationship between interest profiles in the graph, analyzes the content of the interest itself, and expands the user's other interests for recommendation. However, the introduction of external data sources often has a large amount of data, inconsistent device numbers, few user intersections, and poor quality of individual data. There is a certain threshold for small and medium-sized companies to obtain data outside of products. The knowledge graph has a slow update cycle and is relatively static; the graph is trained from the entire network data, and sometimes cannot reflect the unique characteristics of the product; it requires a lot of manual review work. Therefore, it is difficult to enrich the interest profile using the current two methods, which makes it difficult for other recommended products to meet the actual needs of users.

[0057] Therefore, in order to solve the above problems, an embodiment of the present application provides a method for expanding an interest portrait. Based on the similarity relationship, a second interest portrait is expanded on the basis of the user's existing first interest portrait, so that the user's interest portrait is richer, and the description of the user's interest status is more comprehensive and accurate, and products that enhance the user experience are recommended to the user based on the enriched second interest portrait.

[0058] The interest portrait expansion method in this embodiment can be applied to the above Figure 1c In addition to the scenarios shown, the present invention can also be applied to other personalized scenarios, which are not specifically limited here.

[0059] In order to better understand the solution proposed in the embodiment of the present application, the specific process of this embodiment is introduced below. Figure 2, is a schematic diagram of an embodiment of a method for extending an interest portrait provided in this embodiment, and the method may include:

[0060] 201. Obtain a first interest portrait.

[0061] In this embodiment, the first interest portrait is the interest portrait that the user has currently generated based on the user behavior data, wherein the user behavior data may be behavior data generated by the user's browsing behavior or reading behavior on video products, browser products, APP installation and other products. For example, the first interest portrait may be "La Liga: Sports", "Science: Three Body", etc., which will not be specifically limited in the embodiments of this application.

[0062] Optionally, the first interest portrait may be obtained by acquiring user behavior data, thereby obtaining a corresponding first interest portrait.

[0063] 202. Train the first interest portrait based on a preset training model to obtain a first similarity relationship, where the first similarity relationship is an association relationship between the first interest portrait and the second interest portrait.

[0064] In this embodiment, in order to learn the similarity relationship between the first interest portrait and the second interest portrait that needs to be expanded, the first interest portrait can be trained by a preset training model to obtain a first similarity relationship between the first interest portrait and the second interest portrait, and the similarity value corresponding to the first similarity relationship indicates the degree of similarity between the first interest portrait and the second interest portrait. It should be understood that the higher the degree of similarity, the more content is recommended based on the first interest portrait, and similar content can also be recommended based on the second interest portrait.

[0065] Optionally, in some embodiments, the first interest portrait is trained based on a preset training model to obtain a first similarity relationship. Specifically, this can be done by determining N interest words in the first interest portrait, where N is an integer greater than or equal to 0; training at least two of the N interest words based on the preset training model to obtain a latent vector corresponding to each interest word; and determining a cosine similarity value between the latent vectors corresponding to each interest word, wherein the cosine similarity value is used to represent the first similarity relationship.

[0066] In other words, the preset training model can be an Inte2Vec model. In order to obtain the first similarity relationship between the first interest portrait and the second interest portrait, the preset training model can use a method similar to item2Vec, treat the user's first interest portrait as an independent article, and use the skip-gram model (skip-gram) method to learn the latent vector representation of the first object data, that is, predict the surrounding words through the N interest words in the first interest portrait. For details, please refer to Figure 3, which is a logic diagram for determining latent vectors provided in an embodiment of the present application.

[0067] from Figure 3 It can be seen that the preset training model includes an input layer, a hidden layer, and an output layer. At this time, we can first determine the N interest words in the first interest portrait to build a vocabulary, and then use adjacent pairs of interest words as input data for the preset training model to train the model. The hidden layer in the middle is responsible for the training calculation of the weight matrix value. The weight matrix represents the calculation of the similarity relationship between each interest word, so that each row in the weight matrix can be used to represent the latent vector of each interest word. The output layer uses a softmax regression classifier, and each of its nodes will output a value between 0 and 1. The value at this time can represent the predicted probability of each surrounding word relative to the interest word, that is, Figure 3 The probability that the interest word (N-1) described in appears in the adjacent position of the interest word (N). At this time, the cosine similarity between the latent vectors corresponding to each interest word is calculated, so as to represent the first similarity relationship based on the cosine similarity.

[0068] It should be noted that the higher the probability, the higher the similarity between the first interest profile and the second interest profile. The adjacent pairs of interest words described above can be represented by at least two interest words among the N interest words, for example, interest words that may appear simultaneously in an article of interest to a certain user.

[0069] Optionally, in other embodiments, after determining the N interest words in the first interest portrait, a vocabulary can be constructed based on the N interest words, so that each interest word is one-hot encoded, so that the N interest words that appear in text form can be converted into numbers that can be input into a preset training model. For example, when there are only two interest words in the vocabulary, "computer" and "mathematics", the one-hot codes obtained are [0,1] and [1,0] respectively.

[0070] 203. Determine a second interest portrait based on the first interest portrait and the first similarity relationship.

[0071] In this embodiment, after obtaining the first similarity relationship, the user's existing first interest portrait is combined to expand a second interest portrait with similarity, so that more abundant interest portraits can be recommended to the user.

[0072] Optionally, in some other embodiments, the interest score corresponding to the first interest portrait may be determined; and then the second interest portrait may be determined based on the interest score corresponding to the first interest portrait and the first similarity relationship.

[0073] In other words, after obtaining the first similarity relationship, the second interest portrait is determined by the weight formula. That is: ,in, represents the interest score corresponding to the first interest portrait, representing the first similarity relationship, Therefore, through the user's existing first interest portrait and the trained first similarity relationship, the user can obtain a richer second interest portrait.

[0074] For example, see Figure 4 , is a schematic diagram of the second interest portrait expanded in the embodiment of the present application. Figure 4 It can be seen that, for example, the user's existing first interest portrait includes "La Liga", and after combining the first similarity relationship, a richer second interest portrait of the user can be determined, such as "Serie A", "Champions League", etc., where 0.1447 represents the similarity between "Serie A" and "La Liga", and 0.1437 represents the similarity between "Champions League" and "La Liga". It should be understood that the 0.1447 and 0.1437 described above are only for illustrative purposes. In actual applications, they should be determined according to actual conditions and will not be specifically limited in the embodiments of this application.

[0075] In addition, from Figure 4 It can also be seen that the newly generated first similarity relationship can be captured in time, such as the similarity relationship between the recently popular "science fiction movies" and "The Wandering Earth". It should be understood that other newly generated first similarity relationships can also be captured in practical applications, which will not be limited in the embodiments of this application.

[0076] In an embodiment of the present application, the first interest portrait is trained by using a preset training model to obtain a first similarity relationship, thereby expanding the second interest portrait based on the user's existing first interest portrait, thereby enriching the user's interest portrait and providing a more comprehensive and accurate description of the user's interest status, and recommending products that can enhance the user experience to the user based on the enriched second interest portrait.

[0077] The above mainly introduces the scheme provided by the embodiment of the present application from the perspective of the method. It can be understood that in order to realize the above functions, the hardware structure and / or software module corresponding to the execution of each function are included. Those skilled in the art should easily realize that, in combination with the modules and algorithm steps of each example described in the embodiment disclosed in this application, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a certain function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0078] The embodiment of the present application can divide the functional modules of the device according to the above method example. For example, each functional module can be divided according to each function, or two or more functions can be integrated into one processing module. The above integrated module can be implemented in the form of hardware or in the form of software functional modules. It should be noted that the division of modules in the embodiment of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation.

[0079] The following is a detailed description of the interest portrait expansion device 50 in the embodiment of the present application. Figure 5 , Figure 5 This is a schematic diagram of an embodiment of an interest portrait expansion device 50 provided in an embodiment of the present application. The interest portrait expansion device 50 includes:

[0080] An acquisition unit 501 is used to acquire a first interest portrait;

[0081] A training unit 502 is used to train the first interest portrait obtained by the acquisition unit 501 based on a preset training model to obtain a first similarity relationship, where the first similarity relationship is an association relationship between the first interest portrait and the second interest portrait;

[0082] The determination unit 503 is used to determine the second interest portrait based on the first interest portrait obtained by the acquisition unit 501 and the first similarity relationship obtained by the training unit 502.

[0083] Optionally, in the above Figure 5 On the basis of the corresponding embodiment, in another embodiment of the interest portrait expansion device 50 provided in the embodiment of the present application, the training unit 502 may include:

[0084] A first determination module, configured to determine N interest words in the first interest portrait, where N is an integer greater than or equal to 0;

[0085] A training module, configured to train at least two of the N interest words determined by the first determination module based on the preset training model to obtain a latent vector corresponding to each interest word;

[0086] The first determination module is used to determine the cosine similarity value between the latent vectors corresponding to each interest word, and the cosine similarity value is used to represent the first similarity relationship.

[0087] Optionally, in the above Figure 5On the basis of the corresponding optional embodiment, in another embodiment of the interest portrait expansion device 50 provided in the embodiment of the present application, the interest portrait expansion device 50 may also include: an encoding unit for one-hot encoding the N interest words.

[0088] Optionally, in the above Figure 5 as well as Figure 5 On the basis of the corresponding optional embodiment, in another embodiment of the interest portrait expansion device 50 provided in the embodiment of the present application, the determining unit 503 includes:

[0089] A second determination module, used to determine the interest score corresponding to the first interest portrait;

[0090] The second determination module is used to determine the second interest portrait based on the interest score corresponding to the first interest portrait and the first similarity relationship.

[0091] Optionally, in the above Figure 5 as well as Figure 5 On the basis of the corresponding optional embodiment, the acquisition unit 501 is used to acquire user behavior data, and the user behavior data includes a first interest portrait.

[0092] The above describes the interest portrait expansion device 50 in the embodiment of the present application from the perspective of modular functional entities, and the following describes the interest portrait expansion device 50 in the embodiment of the present application from the perspective of hardware processing. Figure 6 FIG. 1 is a schematic diagram of the hardware structure of the communication device in the embodiment of the present application. Figure 6 As shown, the communication device may include:

[0093] The communication device includes at least one processor 601 , a communication line 607 , a memory 603 and at least one communication interface 604 .

[0094] Processor 601 can be a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (server IC), or one or more integrated circuits used to control the execution of the program of the present application.

[0095] The communication link 607 may include a pathway to transmit information between the above-mentioned components.

[0096] The communication interface 604 uses any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0097] The memory 603 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, and the memory may exist independently and be connected to the processor via a communication line 607. The memory may also be integrated with the processor.

[0098] The memory 603 is used to store computer execution instructions for executing the solution of the present application, and the execution is controlled by the processor 601. The processor 601 is used to execute the computer execution instructions stored in the memory 603, thereby realizing the interest portrait expansion method provided in the above embodiment of the present application.

[0099] Optionally, the computer-executable instructions in the embodiments of the present application may also be referred to as application code, which is not specifically limited in the embodiments of the present application.

[0100] In a specific implementation, as an embodiment, the communication device may include multiple processors, such as Figure 6 601 and processor 602 in the embodiment of the present invention. Each of these processors may be a single-CPU processor or a multi-CPU processor. The processor herein may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0101] In a specific implementation, as an embodiment, the communication device may further include an output device 605 and an input device 606. The output device 605 communicates with the processor 601 and may display information in a variety of ways. The input device 606 communicates with the processor 601 and may receive user input in a variety of ways. For example, the input device 606 may be a mouse, a touch screen device, or a sensor device.

[0102] The communication device can be a general purpose device or a dedicated device. In a specific implementation, the communication device can be a desktop computer, a portable computer, a network server, a wireless terminal device, an embedded device or a Figure 6 The embodiments of the present application do not limit the type of communication device.

[0103] In the embodiment of the present application, the processor 601 included in the computer device also has the following functions: obtaining a first interest portrait;

[0104] Training the first interest portrait based on a preset training model to obtain a first similarity relationship, where the first similarity relationship is an association relationship between the first interest portrait and the second interest portrait;

[0105] The second interest portrait is determined based on the first interest portrait and the first similarity relationship.

[0106] The above-mentioned acquisition unit 501 can be implemented through the input device 606, and the training unit 502, the determination unit 503, the first determination module, the training module, the encoding unit, and the second determination module can be implemented through the processor 601 or the processor 602.

[0107] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product.

[0108] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0109] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the unit is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0110] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0111] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0112] If the integrated 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 computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0113] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for expanding interest portraits, characterized in that: include: Get the first interest portrait; Training the first interest portrait based on a preset training model to obtain a first similarity relationship, where the first similarity relationship is an association relationship between the first interest portrait and the second interest portrait, and a similarity value corresponding to the first similarity relationship represents a degree of similarity between the first interest portrait and the second interest portrait; Determining an interest score corresponding to the first interest portrait; determining the second interest portrait based on the interest score corresponding to the first interest portrait and the first similarity relationship; The step of training the first interest portrait based on a preset training model to obtain a first similarity relationship includes: Determine N interest words in the first interest portrait, where N is an integer greater than or equal to 0; Training at least two of the N interest words based on the preset training model to obtain a latent vector corresponding to each of the interest words; Determine a cosine similarity value between latent vectors corresponding to each interest word, wherein the cosine similarity value is used to represent the first similarity relationship.

2. The method according to claim 1, characterized in that The method further comprises: One-hot encoding is performed on the N interest words.

3. The method according to any one of claims 1 to 2, characterized in that: The obtaining of the first interest portrait includes: User behavior data is obtained, where the user behavior data includes a first interest portrait.

4. An interest portrait expansion device, characterized in that: include: An acquisition unit, used for acquiring a first interest portrait; A training unit, configured to train the first interest portrait obtained by the acquisition unit based on a preset training model to obtain a first similarity relationship, wherein the first similarity relationship is an association relationship between the first interest portrait and the second interest portrait, and a similarity value corresponding to the first similarity relationship represents a degree of similarity between the first interest portrait and the second interest portrait; A determining unit, configured to determine an interest score corresponding to the first interest portrait; determining the second interest portrait based on the interest score corresponding to the first interest portrait and the first similarity relationship; The training unit comprises: A first determination module, configured to determine N interest words in the first interest portrait, where N is an integer greater than or equal to 0; A training module, configured to train at least two of the N interest words determined by the first determination module based on the preset training model to obtain a latent vector corresponding to each interest word; The first determination module is used to determine the cosine similarity value between the latent vectors corresponding to each interest word, and the cosine similarity value is used to represent the first similarity relationship.

5. The interest portrait expansion device according to claim 4, characterized in that: The interest portrait expansion device also includes: an encoding unit for one-hot encoding the N interest words.

6. The interest portrait expansion device according to any one of claims 4 to 5, characterized in that: The acquisition unit is used to acquire user behavior data, and the user behavior data includes a first interest portrait.

7. An interest portrait expansion device, characterized in that: The interest portrait expansion device comprises: Input / output (I / O) interfaces, processors, and memory, The memory stores program instructions; The processor is used to execute program instructions stored in the memory to perform the method according to any one of claims 1 to 3.

8. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a computer device, the computer device is caused to perform the method according to any one of claims 1 to 3.

9. A computer program product comprising instructions, characterized in that When the method is executed on a computer, the computer device is caused to execute the method according to any one of claims 1 to 3.

Images