Method and device for pushing information

Abstract

The embodiments of the present disclosure disclose a method and device for pushing information. A specific implementation of the method includes: obtaining a user's behavior feature vector and a feature vector corresponding to the information; updating the user's behavior feature vector using the feature vector of the information and a pre-trained feature extraction model, wherein the feature extraction model includes an encoding network and a decoding network, the decoding network uses the feature vector of the information as a query and the output result of the encoding network as a key and a value; determining the user's attention to the information based on the updated behavior feature vector; in response to determining that the attention meets a preset condition, pushing the information to the terminal device used by the user. This implementation helps to improve the accuracy of the information pushed to the user.

Description

Technical Field

[0001] Embodiments of the present disclosure relate to the field of computer technology, and more particularly to a method and device for pushing information. Background Art

[0002] Personalized push notifications are currently widely used in various information push scenarios to push information that users are most interested in from the massive amount of information on the Internet, achieving personalized and accurate push notifications, such as video push notifications, item push notifications, article push notifications, friend recommendations, and so on.

[0003] Information push usually includes multiple steps such as recall, rough sorting, fine sorting, and re-sorting. Among them, recall generally refers to the quick selection of a part of information from all specified information, rough sorting generally refers to further screening out some information from the recalled information based on personalization, fine sorting generally refers to a more detailed personalization information screening based on the rough sorting results, and re-sorting generally re-sorts the fine sorting results in a combined form, and then filters them through a series of preset rules to obtain the final information push results and present them to the user.

[0004] Existing personalized information push methods mainly convert the user's historical behavior information into a feature vector of fixed length, and then use this feature vector to perform similarity calculations to determine the information that the user is more interested in. Summary of the invention

[0005] The embodiments of the present disclosure provide a method and device for pushing information.

[0006] In a first aspect, an embodiment of the present disclosure provides a method for pushing information, the method comprising: obtaining a user's behavior feature vector and a feature vector corresponding to the information; updating the user's behavior feature vector using the feature vector of the information and a pre-trained feature extraction model, wherein the feature extraction model comprises an encoding network and a decoding network, the decoding network using the feature vector of the information as a query and the output result of the encoding network as a key and a value; determining the user's attention to the information based on the updated behavior feature vector; and in response to determining that the attention meets a preset condition, pushing the information to a terminal device used by the user.

[0007] In a second aspect, an embodiment of the present disclosure provides a device for pushing information, the device comprising: an acquisition unit, configured to acquire a user's behavior feature vector and a feature vector corresponding to the information; an update unit, configured to update the user's behavior feature vector using the feature vector of the information and a pre-trained feature extraction model, wherein the feature extraction model comprises an encoding network and a decoding network, the decoding network uses the feature vector of the information as a query and the output result of the encoding network as a key and a value; a determination unit, configured to determine the user's attention to the information based on the updated behavior feature vector; and a push unit, configured to push the information to a terminal device used by the user in response to determining that the attention meets a preset condition.

[0008] In a third aspect, an embodiment of the present disclosure provides an electronic device, comprising: one or more processors; a storage device for storing one or more programs; when the one or more programs are executed by one or more processors, the one or more processors implement the method described in any implementation manner in the first aspect.

[0009] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any implementation manner in the first aspect.

[0010] The method and device for pushing information provided by the embodiments of the present disclosure use a decoding network in a feature extraction model to update a user's behavior feature vector using the feature vector of the information to be pushed as a query and the output result of the encoding network as a key and a value, thereby strengthening the features in the user's behavior feature vector that are highly relevant to the information to be pushed, while weakening the features in the user's behavior feature vector that are less relevant or irrelevant to the information to be pushed. Further, the user's attention to the information to be pushed can be more accurately determined based on the updated behavior feature vector, and then determining whether to push the information to the user based on the attention helps to improve the accuracy of the information pushed to the user. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Other features, objects and advantages of the present disclosure will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0012] Figure 1 is an exemplary system architecture diagram in which an embodiment of the present disclosure may be applied;

[0013] Figure 2 is a flow chart of an embodiment of a method for pushing information according to the present disclosure;

[0014] Figure 3It is a schematic structural diagram of the feature extraction model;

[0015] Figure 4 is a flow chart of another embodiment of a method for pushing information according to the present disclosure;

[0016] Figure 5 is a flow chart of another embodiment of a method for pushing information according to the present disclosure;

[0017] Figure 6 is a schematic diagram of an application scenario of a method for pushing information according to an embodiment of the present disclosure;

[0018] Figure 7 is a structural schematic diagram of an embodiment of a device for pushing information according to the present disclosure;

[0019] Figure 8 It is a schematic diagram of the structure of an electronic device suitable for implementing the embodiments of the present disclosure. DETAILED DESCRIPTION

[0020] The present disclosure is further described in detail below in conjunction with the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the relevant parts are shown in the accompanying drawings.

[0021] It should be noted that the data collection involved in the embodiments of the present disclosure (such as user behavior data, behavior characteristics, attribute characteristics of information, etc.) is carried out on the basis of having obtained the authorization of the relevant subjects and complies with the provisions of relevant laws and regulations.

[0022] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other. The present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0023] Figure 1 An exemplary architecture 100 is shown to which an embodiment of a method for pushing information or an apparatus for pushing information of the present disclosure can be applied.

[0024] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0025] The terminal devices 101, 102, 103 interact with the server 105 through the network 104 to receive or send messages, etc. Various client applications may be installed on the terminal devices 101, 102, 103. For example, browser applications, search applications, instant messaging tools, social platforms, information flow applications, etc.

[0026] Terminal devices 101, 102, 103 can be hardware or software. When terminal devices 101, 102, 103 are hardware, they can be various electronic devices, including but not limited to smart phones, tablet computers, e-book readers, laptop portable computers and desktop computers, etc. When terminal devices 101, 102, 103 are software, they can be installed in the electronic devices listed above. It can be implemented as multiple software or software modules (for example, multiple software or software modules for providing distributed services), or it can be implemented as a single software or software module. No specific limitation is made here.

[0027] The server 105 may be a server that provides various services, such as a backend server that provides service support for client applications installed on the terminal devices 101, 102, and 103. The server 105 may obtain the behavior feature vectors of the users corresponding to the terminal devices 101, 102, and 103 and the feature vectors corresponding to the information to be pushed, and then update the user's behavior feature vectors using a pre-trained feature extraction model, and determine the user's attention to the information based on the updated user's behavior feature vectors. If the attention meets the preset conditions, the information may be further pushed to the terminal devices 101, 102, and 103.

[0028] It should be noted that the method for pushing information provided in the embodiments of the present disclosure is generally executed by the server 105 , and accordingly, the device for pushing information is generally disposed in the server 105 .

[0029] It should also be pointed out that information processing applications may also be installed in the terminal devices 101, 102, and 103. The terminal devices 101, 102, and 103 may also obtain the user's behavior feature vector and the feature vector corresponding to the information to be pushed based on the information processing application, and then use the pre-trained feature extraction model to update the user's behavior feature vector, and determine the user's attention to the information based on the updated user's behavior feature vector. If the attention meets the preset conditions, the information is displayed to the corresponding user.

[0030] It should be noted that the server 105 can be hardware or software. When the server 105 is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or it can be implemented as a single server. When the server 105 is software, it can be implemented as multiple software or software modules (for example, multiple software or software modules for providing distributed services), or it can be implemented as a single software or software module. No specific limitation is made here.

[0031] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0032] Continue to refer Figure 2 , which shows a process 200 of an embodiment of a method for pushing information according to the present disclosure. The method for pushing information includes the following steps:

[0033] Step 201, obtaining a user's behavior feature vector and a feature vector corresponding to the information.

[0034] In this embodiment, the user may be a user of any terminal device. The user's behavior feature vector may be used to characterize the user's historical behavior features. The information may be various types of information to be pushed. For example, the information may be an image, text, audio, video, item information, service information, etc. The feature vector of the information may be used to characterize the features of the information.

[0035] The execution subject of the method used to push information (such as Figure 1 The server 105 shown in the figure can obtain the user's behavior feature vector and the feature vector corresponding to the information from a local or other storage device. It should be noted that the data sources corresponding to the user's behavior feature vector and the feature vector corresponding to the information can be the same or different.

[0036] Step 202: Update the user's behavior feature vector using the feature vector of the information and a pre-trained feature extraction model.

[0037] In this embodiment, the feature extraction model may adopt the framework of an encoding network (Encoder) and a decoding network (Decoder), and the specific model structure may be various neural networks (such as convolutional neural networks, recurrent neural networks, etc.). For example, the feature extraction model may be a Transform model.

[0038] Among them, the decoding network can be implemented based on the attention mechanism. Specifically, the decoding network can use the feature vector of the information as the Query and the output results of the encoding network as the Key and Value to implement the attention mechanism.

[0039] Generally, the implementation process of the attention mechanism includes: scoring all elements in the behavior feature vector according to each element in the user's behavior feature vector (usually the dot product of the query vector and the key vector) to indicate the relevance of each element to each element in the behavior feature vector, and normalizing the score, and then multiplying the value vector by the normalized score as the final score. Then, the final score is used as the weight to calculate the updated value generated by each element after the attention mechanism is processed in a weighted sum manner.

[0040] It should be noted that the above-mentioned attention mechanism is one of the main technologies currently studied and applied in the field of natural language processing based on deep learning, and will not be elaborated here.

[0041] Step 203: Determine the user's attention to the information based on the updated behavior feature vector.

[0042] In this embodiment, the user's attention to information can represent the user's interest in the information, and various representation methods can be flexibly adopted according to actual application scenarios or application requirements. For example, the attention can be represented by the user's click rate or browsing volume of the information.

[0043] After obtaining the updated behavior feature vector of the user, various methods can be used to determine the user's attention to the information. For example, the distance between the updated behavior feature vector of the user and the feature vector of the information can be calculated, and the calculation result is used as the user's attention to the information.

[0044] Step 204: in response to determining that the degree of attention meets a preset condition, push information to a terminal device used by the user.

[0045] In this embodiment, the preset condition can be pre-set by a technician according to actual application requirements or application scenarios. For example, the preset condition can be that the attention level is greater than a preset threshold. At this time, if the attention level of the determined information is greater than the preset threshold, the information can be pushed to the terminal device used by the user to display the information to the user.

[0046] In some optional implementations of this embodiment, the feature extraction model may include a first decoding network and a second decoding network. The first decoding network may generate a first behavior feature vector of the user based on the feature vector of the information and the output result of the encoding network. The second decoding network may generate a second behavior feature vector of the user based on the feature vector of the information and the first behavior feature vector output by the first decoding network. At this time, the first behavior feature vector and the second behavior feature vector respectively output by the first decoding network and the second decoding network may be determined as the updated behavior feature vector of the user.

[0047] like Figure 3 As shown, it shows a schematic structural diagram 300 of a feature extraction model. The feature extraction model 300 may include an encoding network 301, a first decoding network 302, and a second decoding network 303. Among them, the first decoding network 302 includes an attention module 3021 and a processing module 3022. The second decoding network 303 includes an attention module 3031 and a processing module 3032. Among them, the processing module 3022 and the processing module 3032 can be various feedforward neural networks, so as to generate a first behavior feature vector and a second behavior feature vector respectively by performing fusion, dimensional transformation and other processing according to the output results of the attention module.

[0048] Since the relatively high-dimensional behavior feature vectors and relatively low-dimensional behavior feature vectors output by the two decoding networks can respectively reflect the user's behavior characteristics from different granularities, the use of the user's multi-dimensional behavior feature vectors can more comprehensively and accurately express the user's behavior characteristics, thereby helping to improve the accuracy of the push information determined based on the user's behavior characteristics.

[0049] In some optional implementations of this embodiment, the characteristic behavior vector of the user can be used to characterize the interaction information set corresponding to the user in at least one historical period. The interaction information can represent the information that the user has interacted with historically. Interaction can refer to various types of interactions, such as browsing, clicking, sharing, commenting, purchasing, etc. Generally, the time corresponding to each historical period is different and non-repeating. The duration corresponding to each historical period can be the same or different.

[0050] As an example, the user's behavior feature vector can be determined by the following steps:

[0051] Step 1: Obtain the information that the user has interacted with during the target historical period, and form an information sequence according to the corresponding interaction time.

[0052] The target historical period can be flexibly set by the technicians according to the actual application scenario. For example, the item sequence formed by the items clicked by the user in the past month can be obtained.

[0053] Step 2: Divide the target historical period into at least one historical period.

[0054] Step 3: For each historical period obtained by division, select the information subsequence corresponding to the historical period from the information sequence, and map the selected information subsequence into a feature vector.

[0055] Step 4: Fuse the feature vectors corresponding to each historical period to obtain the user's behavior feature vector.

[0056] The fusion method may adopt various vector fusion methods. For example, the feature vectors corresponding to each historical period may be sequentially spliced.

[0057] By segmenting the user's interaction information within a certain historical period by time and mapping each segment to a corresponding feature vector, we can more reasonably statistically analyze the user's historical interaction information and thus more accurately express the user's behavioral characteristics.

[0058] In the prior art, user historical behavior features are usually compressed into a feature vector of fixed length. This method limits the ability to express user behavior features and cannot accurately represent user interests. The extraction of user behavior feature vectors is usually irrelevant to the information sent instead.

[0059] The method provided by the above-mentioned embodiment of the present disclosure uses the feature vector of the information to be pushed as the query through the decoding network in the feature extraction model and the output result of the encoding network as the key and value to update the user's behavior feature vector, thereby strengthening the features in the user's behavior feature vector that are highly relevant to the information to be pushed, while weakening the features in the user's behavior feature vector that are less relevant or irrelevant to the information to be pushed, and then according to the updated behavior feature vector, the user's attention to the information to be pushed can be more accurately determined, and then determining whether to push the information to the user based on the attention helps to improve the accuracy of the information pushed to the user. In this way, for each piece of information to be pushed, the features related to the information in the user's behavior features can be extracted in combination with the specific features of the information, and each piece of information to be pushed corresponds to the behavior features of different users, which can improve the overall expression ability of the user's behavior features to more accurately reflect the user's various interests.

[0060] Further references Figure 4 , which shows a process 400 of another embodiment of a method for pushing information. The process 400 of the method for pushing information includes the following steps:

[0061] Step 401, obtaining a user's behavior feature vector and a feature vector corresponding to the information.

[0062] Step 402: Update the user's behavior feature vector using the feature vector of the information and a pre-trained feature extraction model.

[0063] Step 403, obtaining the associated feature vector.

[0064] In this embodiment, the execution subject may obtain the associated feature vector from a local or other storage device. The associated feature vector may be used to represent at least one of the following: attribute features of information, attribute features of users, and interaction features of users with respect to information.

[0065] Among them, the attribute characteristics of information can be used to represent various attributes of information. For example, for an item, the attributes of the item include category, price, color, sales volume, etc. The attribute characteristics of users can be used to represent various attributes of users. For example, basic attributes and behavioral attributes of users. The interactive characteristics of users for information can be used to represent the characteristics of the interaction between users for the information. For example, the interactive characteristics include the interactive attributes of users for various information under the category to which the information belongs (such as click-through rate, sales volume, page views, etc.).

[0066] When the associated feature vector represents at least two of the attribute characteristics of information, the attribute characteristics of users, and the interaction characteristics of users with respect to information, the feature vectors corresponding to the at least two items can be concatenated, and then dimensionality transformation and other processing (such as dimensionality reduction) can be performed to obtain the associated feature vector.

[0067] Step 404: Determine the user's attention to the information based on the associated feature vector, the updated behavior feature vector, and the feature vector of the information.

[0068] In this embodiment, various methods can be used to determine the user's attention to the information based on the associated feature vector, the updated behavior feature vector, and the feature vector of the information. For example, the distance between any two feature vectors among the associated feature vector, the updated behavior feature vector, and the feature vector of the information can be calculated, and then the average value of each distance can be calculated as the user's attention to the information.

[0069] Step 405: In response to determining that the degree of attention meets a preset condition, information is pushed to a terminal device used by the user.

[0070] In some optional implementations of this embodiment, the associated feature vector may be obtained based on low-frequency filtering. The low-frequency filtering may refer to filtering out features with lower frequencies. Specifically, after the associated feature vector is obtained, the associated feature vector may be subjected to low-frequency filtering to update the associated feature vector.

[0071] Through low-frequency filtering, the associated feature vector can reduce the impact of features with lower frequency of occurrence on subsequent calculations, which helps to more accurately represent the user, the information to be pushed, and the interactive relationship between the two.

[0072] Optionally, the associated feature vector may also be normalized to facilitate subsequent calculations.

[0073] In some optional implementations of this embodiment, the associated feature vector, the updated behavior feature vector, and the feature vector of the information may be input into a pre-trained attention generation model to obtain the user's attention to the information.

[0074] The attention generation model can be used to predict the user's attention to the information based on the associated feature vector, the user's updated behavior feature vector and the feature vector of the information. The attention generation model can be various existing neural network models, such as a convolutional neural network based on deep learning.

[0075] The contents not specifically described in this embodiment can be referred to Figure 2 The relevant descriptions in the corresponding embodiments will not be repeated here.

[0076] After obtaining the updated user behavior feature vector corresponding to the information, the method provided by the above-mentioned embodiment of the present disclosure can combine the attribute characteristics of the information, the attribute characteristics of the user and the interaction characteristics of the user with respect to the information to comprehensively determine the user's attention to the information, thereby further improving the accuracy of the determined attention, and further helping to improve the accuracy of the information pushed to the user based on the attention.

[0077] Further references Figure 5 , which shows a process 500 of another embodiment of a method for pushing information. The process 500 of the method for pushing information includes the following steps:

[0078] Step 501: Obtain a user's behavior feature vector and a feature vector corresponding to information in a target information set.

[0079] In this embodiment, the target information set may be composed of a number of information, which may be pre-specified by a technician. For example, the target information set may be an information set obtained in any step of the above information push process, such as a rough sorting result or a fine sorting result, etc. The above execution subject may obtain the feature vectors corresponding to each piece of information in the target information set.

[0080] Step 502: for the information in the target information, using the feature vector and feature extraction model of the information, update the user's behavior feature vector, and determine the user's attention to the information according to the updated behavior feature vector of the user.

[0081] In this embodiment, for each information in the target information set, the feature vector and feature extraction model of the information can be used to update the user's behavior feature vector to obtain an updated behavior feature vector corresponding to the information. Then, the user's attention to the information is determined based on the updated behavior feature vector corresponding to the information.

[0082] Step 503: Select information from the target information set for push according to the attention levels corresponding to the information in the target information set.

[0083] In this embodiment, the information may be sorted according to the degree of attention, and then a preset number of information may be selected in descending order for push.

[0084] In some optional implementations of this embodiment, the above-mentioned various models, such as feature extraction models, attention generation models, etc., can be uniformly trained in an end-to-end manner to reduce the complexity and computational complexity of model training.

[0085] The contents not specifically described in this embodiment can be referred to Figure 2 The relevant descriptions in the corresponding embodiments will not be repeated here.

[0086] Continue to see Figure 6 , Figure 6 FIG. 6 is an exemplary application scenario 600 of the method for pushing information according to this embodiment. Figure 6 In the application scenario, the items clicked by the user in the past month can be obtained 601, and the sequence of items clicked by the user in chronological order can be mapped to the user's behavior feature sequence 602. At the same time, the rough sorting result 603 (i.e., the set of items to be pushed) obtained in the rough sorting step in the information push process for the user is obtained, and then the feature vector 604 of each item in the set of items to be pushed is obtained.

[0087] Afterwards, for the feature vector of each item, the item feature vector 604 can be used as a query and the output result of the encoding network can be used as the key and value, and the feature extraction model 605 can be used to update the behavior feature vector 602 to obtain the first behavior feature vector 606 generated by the first decoding network 6051 and the second behavior feature vector 607 generated by the second decoding network 6052.

[0088] For each item in the set of items to be pushed, the attribute characteristics of the item, the attribute characteristics of the user, and the user's interaction characteristics 608 with the item are input into the click rate prediction model 609 to obtain the user's click rate 610 for the item. Afterwards, some items can be filtered from the rough sorting results in descending order of click rates to obtain item screening results 611, and the items in the item screening results 611 can be pushed to the user terminal for display.

[0089] The method provided by the above-mentioned embodiments of the present disclosure associates the information to be pushed during the process of extracting the user's behavioral features, so as to filter out the user's behavioral features corresponding to each generation of pushed information from the user's behavioral feature vector, and then use the updated user's behavioral features corresponding to the information to determine the user's attention to the information, and then select the push information according to the attention corresponding to each information. The user's attention to different information can be determined from different feature angles, so as to more accurately filter the push information.

[0090] Further references Figure 7 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a device for pushing information. Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0091] like Figure 7 As shown, the apparatus 700 for pushing information provided in this embodiment includes an acquisition unit 701, an update unit 702, a determination unit 703 and a push unit 704. The acquisition unit 701 is configured to acquire a user's behavior feature vector and a feature vector corresponding to the information; the update unit 702 is configured to update the behavior feature vector using the feature vector of the information and a pre-trained feature extraction model, wherein the feature extraction model includes an encoding network and a decoding network, the decoding network uses the feature vector of the information as a query and the output result of the encoding network as a key and a value; the determination unit 703 is configured to determine the user's attention to the information according to the updated behavior feature vector; the push unit 704 is configured to push the information to the terminal device used by the user in response to determining that the attention meets the preset conditions.

[0092] In the device for pushing information 700, the specific processing of the acquisition unit 701, the updating unit 702, the determination unit 703 and the pushing unit 704 and the technical effects thereof can be referred to in the respective Figure 2 The relevant descriptions of step 201, step 202, step 203 and step 204 in the corresponding embodiment are not repeated here.

[0093] In some optional implementations of this embodiment, the feature extraction model includes a first decoding network and a second decoding network, wherein the first decoding network generates a first behavior feature vector based on the feature vector of the information and the output result of the encoding network, and the second decoding network generates a second behavior feature vector based on the feature vector of the information and the first behavior feature vector; and the above-mentioned updating unit 702 is further configured to: determine the first behavior feature vector and the second behavior feature vector as the updated behavior feature vector.

[0094] In some optional implementations of this embodiment, the above-mentioned behavior feature vector is used to characterize the interaction information set corresponding to the user in at least one historical period.

[0095] In some optional implementations of this embodiment, the above-mentioned acquisition unit 701 is further configured to: acquire an associated feature vector, wherein the associated feature vector is used to characterize at least one of the following: attribute characteristics of the information, attribute characteristics of the user, and interaction characteristics of the user with respect to the information; and the above-mentioned determination unit 703 is further configured to: determine the user's attention to the information based on the associated feature vector, the updated behavior feature vector and the feature vector of the information.

[0096] In some optional implementations of this embodiment, the above-mentioned associated feature vector is obtained based on low-frequency filtering processing.

[0097] In some optional implementations of this embodiment, the above-mentioned determination unit 703 is further configured to: input the above-mentioned associated feature vector, the updated behavior feature vector and the feature vector of the information into a pre-trained attention generation model to obtain the user's attention to the information.

[0098] In some optional implementations of this embodiment, the information belongs to a target information set; and the above-mentioned updating unit 702 is further configured to: for the information in the target information, use the feature vector and feature extraction model of the information to update the user's behavior feature vector; the above-mentioned determination unit 703 is further configured to: for the information in the target information, determine the user's attention to the information according to the updated behavior feature vector corresponding to the information; and the above-mentioned push unit 704 is further configured to select information from the target information set for push according to the attention corresponding to the information in the target information set.

[0099] The device provided by the above-mentioned embodiment of the present disclosure obtains the user's behavior feature vector and the feature vector corresponding to the information through the acquisition unit; the updating unit updates the behavior feature vector using the feature vector of the information and the pre-trained feature extraction model, wherein the feature extraction model includes an encoding network and a decoding network, the decoding network uses the feature vector of the information as a query and the output result of the encoding network as a key and a value; the determination unit determines the user's attention to the information according to the updated behavior feature vector; the push unit pushes the information to the terminal device used by the user in response to determining that the attention meets the preset conditions. Thus, by updating the user's behavior feature vector through the feature extraction model, the features in the user's behavior feature vector that are highly relevant to the information to be pushed can be strengthened, while the features in the user's behavior feature vector that are less relevant or irrelevant to the information to be pushed can be weakened, and then the user's attention to the information to be pushed can be more accurately determined according to the updated behavior feature vector, and then determining whether to push the information to the user based on the attention helps to improve the accuracy of the information pushed to the user.

[0100] Reference below Figure 8 , which shows an electronic device (eg, Figure 1 A schematic diagram of the structure of the server in FIG. 800. Figure 8 The server shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0101] like Figure 8 As shown, the electronic device 800 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage device 808 into a random access memory (RAM) 803. In the RAM 803, various programs and data required for the operation of the electronic device 800 are also stored. The processing device 801, the ROM 802, and the RAM 803 are connected to each other via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0102] Typically, the following devices may be connected to the I / O interface 805: input devices 806 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 807 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 808 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 809. The communication device 809 may allow the electronic device 800 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 8The electronic device 800 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead. Figure 8 Each block shown in the figure may represent one device, or may represent multiple devices as required.

[0103] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device 809, or installed from a storage device 808, or installed from a ROM 802. When the computer program is executed by the processing device 801, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0104] It should be noted that the computer-readable medium described in the embodiments of the present disclosure may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the embodiments of the present disclosure, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, device or device. In the embodiments of the present disclosure, the computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0105] The computer-readable medium may be included in the electronic device; or it may exist independently without being installed in the electronic device. The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device: obtains the user's behavior feature vector and the feature vector corresponding to the information; updates the user's behavior feature vector using the feature vector of the information and a pre-trained feature extraction model, wherein the feature extraction model includes an encoding network and a decoding network, and the decoding network uses the feature vector of the information as a query and the output result of the encoding network as a key and a value; determines the user's attention to the information based on the updated behavior feature vector; and pushes the information to the terminal device used by the user in response to determining that the attention meets the preset conditions.

[0106] Computer program code for performing the operations of embodiments of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on a user's computer, partially on a user's computer, as a separate software package, partially on a user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0107] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each 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 implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0108] The units involved in the embodiments described in the present disclosure may be implemented by software or by hardware. The described units may also be provided in a processor, for example, may be described as: a processor including an acquisition unit, an update unit, a determination unit, and a push unit. The names of these units do not, in some cases, constitute limitations on the units themselves, for example, the acquisition unit may also be described as a "unit for acquiring a user's behavior feature vector and a feature vector corresponding to the information".

[0109] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the embodiments of the present disclosure to form a technical solution.

Claims

1. A method for pushing information, comprising: Obtaining the user's behavior feature vector and the feature vector corresponding to the information; Using the feature vector of the information and a pre-trained feature extraction model, the behavior feature vector is updated, wherein the feature extraction model includes an encoding network and a decoding network, the decoding network uses the feature vector of the information as a query and the output result of the encoding network as a key and a value; Determining the user's attention to the information according to the updated behavior feature vector; In response to determining that the attention degree meets a preset condition, pushing the information to a terminal device used by the user; The feature extraction model includes a first decoding network and a second decoding network, wherein the first decoding network generates a first behavior feature vector according to a feature vector of the information and an output result of the encoding network, and the second decoding network generates a second behavior feature vector according to the feature vector of the information and the first behavior feature vector; and The updating of the behavior feature vector by using the feature vector of the information and a pre-trained feature extraction model comprises: The first behavior feature vector and the second behavior feature vector are determined as updated behavior feature vectors.

2. The method according to claim 1, wherein: The behavior feature vector is used to characterize an interaction information set corresponding to the user in at least one historical period.

3. The method according to claim 1, wherein: The method further comprises: Acquire an associated feature vector, wherein the associated feature vector is used to represent at least one of the following: an attribute feature of the information, an attribute feature of the user, and an interaction feature of the user with respect to the information; and The determining, according to the updated behavior feature vector, the user's attention to the information includes: The user's attention to the information is determined according to the associated feature vector, the updated behavior feature vector, and the feature vector of the information.

4. The method according to claim 3, wherein: The associated feature vector is obtained based on low-frequency filtering.

5. The method according to claim 3, wherein: The determining, according to the associated feature vector, the updated behavior feature vector, and the feature vector of the information, the user's attention to the information includes: The associated feature vector, the updated behavior feature vector, and the feature vector of the information are input into a pre-trained attention generation model to obtain the user's attention to the information.

6. The method according to any one of claims 1 to 5, wherein: The information belongs to a target information set; and The updating of the behavior feature vector by using the feature vector of the information and a pre-trained feature extraction model, and determining the user's attention to the information according to the updated behavior feature vector comprises: For the information in the target information, using the feature vector and feature extraction model of the information, updating the behavior feature vector; determining the user's attention to the information according to the updated behavior feature vector; and In response to determining that the attention degree meets a preset condition, pushing the information to a terminal device used by the user includes: According to the attention levels respectively corresponding to the information in the target information set, information is selected from the target information set for push.

7. A device for pushing information, wherein: The device comprises: An acquisition unit, configured to acquire a user's behavior feature vector and a feature vector corresponding to the information; An updating unit is configured to update the behavior feature vector using the feature vector of the information and a pre-trained feature extraction model, wherein the feature extraction model includes an encoding network and a decoding network, the decoding network uses the feature vector of the information as a query and uses the output result of the encoding network as a key and a value; a determining unit configured to determine the user's attention to the information according to the updated behavior feature vector; a push unit, configured to push the information to a terminal device used by the user in response to determining that the attention degree meets a preset condition; The feature extraction model includes a first decoding network and a second decoding network, wherein the first decoding network generates a first behavior feature vector according to a feature vector of the information and an output result of the encoding network, and the second decoding network generates a second behavior feature vector according to the feature vector of the information and the first behavior feature vector; and The updating unit is further configured to: determine the first behavior feature vector and the second behavior feature vector as updated behavior feature vectors.

8. The device according to claim 7, wherein: The behavior feature vector is used to characterize an interaction information set corresponding to the user in at least one historical period.

9. The device according to claim 7, wherein: The acquisition unit is further configured to: acquire an associated feature vector, wherein the associated feature vector is used to characterize at least one of the following: an attribute feature of the information, an attribute feature of the user, and an interaction feature of the user with respect to the information; and The determining unit is further configured to determine the user's attention to the information based on the associated feature vector, the updated behavior feature vector, and the feature vector of the information.

10. The device according to claim 9, wherein: The associated feature vector is obtained based on low-frequency filtering.

11. The device according to claim 9, wherein: The determining unit is further configured to: The associated feature vector, the updated behavior feature vector, and the feature vector of the information are input into a pre-trained attention generation model to obtain the user's attention to the information.

12. The device according to any one of claims 7 to 11, wherein: The information belongs to a target information set; and The updating unit is further configured to: for information in the target information, update the behavior feature vector using the feature vector and feature extraction model of the information; The determining unit is further configured to: determine, for information in the target information, the user's attention to the information according to the updated behavior feature vector corresponding to the information; and The push unit is further configured to select information from the target information set for push according to the attention levels respectively corresponding to the information in the target information set.

13. An electronic device comprising: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 6.

14. A computer readable medium having a computer program stored thereon, wherein: When the program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images