CN115203564A

CN115203564A - Information flow recommendation method and device and computer program product

Info

Publication number: CN115203564A
Application number: CN202210857765.1A
Authority: CN
Inventors: 邓罗丹
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2022-07-20
Filing date: 2022-07-20
Publication date: 2022-10-18
Also published as: WO2024016680A1

Abstract

The disclosure provides an information flow recommendation method, an information flow recommendation device, electronic equipment, a storage medium and a program product, and relates to the technical field of artificial intelligence, in particular to the technical field of evolution strategies. The specific implementation scheme is as follows: acquiring characteristic information of a first user in an information flow recommendation scene; determining a first weight corresponding to each factor in a factor set according to the characteristic information through a multi-factor fusion parameter network, wherein the factor in the factor set represents index information required to be considered in the information flow recommendation process; determining a second weight corresponding to each factor in the factor set according to the characteristic information through a gating screening network; determining a target factor suitable for a first user under an information flow recommendation scene in the factor set according to the first weight and the second weight; and determining and pushing a recommendation result corresponding to the first user in the information flow recommendation scene to the first user according to the target factor. The present disclosure improves the accuracy of the recommendation.

Description

Information flow recommendation method and device and computer program product

Technical Field

The present disclosure relates to the field of artificial intelligence technologies, and in particular, to the field of evolutionary strategy technologies, and in particular, to a method and an apparatus for information stream recommendation, a method and an apparatus for model training, an electronic device, a storage medium, and a computer program product for information stream recommendation, which may be used in an information stream recommendation scenario.

Background

The information flow recommendation is different from the advertisement, not only focuses on the point-to-area ratio of the resources, but also integrates a series of experience indexes such as reading duration, resource diversity, user approval amount and sharing amount as a comprehensive recommendation index. Although the number of the fused target factors is increased, different fusion factors have applicable scene limitations, for example, the primary task for a new user model is to promote innovation, and the targets of duration, diversity and the like are not the problems of important system attention. How to perform adaptive factor screening on a scene faced by a user is a common problem in an information flow recommendation system.

Disclosure of Invention

The disclosure provides an information flow recommendation method and device, a model training method and device, an electronic device, a storage medium and a computer program product.

According to a first aspect, there is provided an information flow recommendation method, comprising: acquiring characteristic information of a first user in an information flow recommendation scene; determining a first weight corresponding to each factor in a factor set according to the characteristic information through a multi-factor fusion parameter network, wherein the factor in the factor set represents index information required to be considered in the information flow recommendation process; determining a second weight corresponding to each factor in the factor set according to the characteristic information through a gating screening network; determining a target factor suitable for a first user under an information flow recommendation scene in the factor set according to the first weight and the second weight; and determining and pushing a recommendation result corresponding to the first user in the information flow recommendation scene to the first user according to the target factor.

According to a second aspect, there is provided a model training method comprising: acquiring characteristic information of a second user in an information flow recommendation scene; determining a first weight corresponding to each factor in a factor set according to the characteristic information through an initial multi-factor fusion parameter network, wherein the factor in the factor set represents index information required to be considered in the information flow recommendation process; determining a second weight corresponding to each factor in the factor set according to the characteristic information through an initial gating screening network; determining a target factor suitable for a second user under the information flow recommendation scene in the factor set according to the first weight and the second weight; determining a recommendation result corresponding to a second user in an information flow recommendation scene according to the target factor; and adjusting parameters of the initial multi-factor fusion parameter network and parameters of the initial gating screening network according to feedback information of the second user on the recommendation result by adopting an evolution strategy so as to obtain the trained multi-factor fusion parameter network and the trained gating screening network.

According to a third aspect, there is provided an information flow recommendation apparatus comprising: the information flow recommendation system comprises a first acquisition unit, a second acquisition unit and a third acquisition unit, wherein the first acquisition unit is configured to acquire characteristic information of a first user in an information flow recommendation scene; the first determining unit is configured to determine a first weight corresponding to each factor in a factor set according to the characteristic information through a multi-factor fusion parameter network, wherein the factor in the factor set represents index information required to be considered in the information flow recommendation process; the second determining unit is configured to determine a second weight corresponding to each factor in the factor set according to the characteristic information through the gated screening network; the third determining unit is configured to determine a target factor, which is suitable for the first user in the information flow recommendation scene, in the factor set according to the first weight and the second weight; and the recommending unit is configured to determine and push a recommending result corresponding to the first user in the information flow recommending scene to the first user according to the target factor.

According to a fourth aspect, there is provided a model training apparatus comprising: the second acquisition unit is configured to acquire the characteristic information of the second user in the information flow recommendation scene; a fourth determining unit, configured to determine, through the initial multi-factor fusion parameter network, a first weight corresponding to each factor in the factor set according to the feature information, where the factor in the factor set represents index information to be considered in an information flow recommendation process; a fifth determining unit configured to determine, through the initial gated screening network, a second weight corresponding to each factor in the factor set according to the feature information; a sixth determining unit, configured to determine, according to the first weight and the second weight, a target factor applicable to a second user in the information flow recommendation scenario in the factor set; the seventh determining unit is configured to determine a recommendation result corresponding to a second user in the information flow recommendation scene according to the target factor; and the training unit is configured to adjust parameters of the initial multi-factor fusion parameter network and parameters of the initial gating screening network according to feedback information of the second user on the recommendation result by adopting an evolution strategy so as to obtain the trained multi-factor fusion parameter network and the trained gating screening network.

According to a fifth aspect, there is provided an electronic device comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method as described in any one of the implementations of the first and second aspects.

According to a sixth aspect, there is provided a non-transitory computer readable storage medium storing computer instructions for causing a computer to perform the method as described in any one of the implementations of the first and second aspects.

According to a seventh aspect, there is provided a computer program product comprising: a computer program which, when executed by a processor, implements a method as described in any of the implementations of the first aspect and the second aspect.

According to the technology disclosed by the invention, the information flow recommendation method is provided, under the information flow recommendation scene, the first weight of each factor corresponding to a user is determined through a multi-factor fusion parameter network, the second weight of each factor is determined through a gating screening network, so that the target factor suitable for the user under the information flow recommendation scene is accurately determined according to the first weight and the second weight, the information flow recommendation is carried out, and the accuracy of the recommendation result is improved.

It should be understood that the statements in this section are not intended to identify key or critical features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description.

Drawings

The drawings are included to provide a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:

FIG. 1 is an exemplary system architecture diagram in which an embodiment according to the present disclosure may be applied;

FIG. 2 is a flow diagram for one embodiment of an information flow recommendation method according to the present disclosure;

fig. 3 is a schematic diagram of an application scenario of the information flow recommendation method according to the present embodiment;

FIG. 4 is a flow diagram of yet another embodiment of an information flow recommendation method according to the present disclosure;

FIG. 5 is a flow diagram of one embodiment of a model training method according to the present disclosure;

FIG. 6 is a block diagram of one embodiment of an information flow recommendation device according to the present disclosure;

FIG. 7 is a block diagram of one embodiment of a model training apparatus according to the present disclosure;

FIG. 8 is a schematic block diagram of a computer system suitable for use in implementing embodiments of the present disclosure.

Detailed Description

Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, in which various details of the embodiments of the disclosure are included to assist understanding, and which are to be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

In the technical scheme of the disclosure, the collection, storage, use, processing, transmission, provision, disclosure and other processing of the personal information of the related user are all in accordance with the regulations of related laws and regulations and do not violate the good customs of the public order.

Fig. 1 illustrates an exemplary architecture 100 to which the information flow recommendation method and apparatus, the model training method and apparatus of the present disclosure may be applied.

As shown in fig. 1, the system architecture 100 may include

terminal devices

101, 102, 103, a network 104, and a server 105. The communication connections between the

terminal devices

101, 102, 103 form a topological network and the network 104 serves as a medium for providing communication links between the

terminal devices

101, 102, 103 and the server 105. Network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, to name a few.

The

terminal devices

101, 102, 103 may be hardware devices or software that support network connections for data interaction and data processing. When the

terminal devices

101, 102, 103 are hardware, they may be various electronic devices supporting network connection, information acquisition, interaction, display, processing, and the like, including but not limited to smart phones, tablet computers, e-book readers, laptop portable computers, desktop computers, and the like. When the

terminal devices

101, 102, 103 are software, they can be installed in the electronic devices listed above. It may be implemented, for example, as multiple software or software modules to provide distributed services, or as a single software or software module. And is not particularly limited herein.

The server 105 may be a server providing various services, for example, a background processing server determining, according to feature information of users corresponding to the

terminal devices

101, 102, and 103 in an information flow recommendation scenario, first weights of factors corresponding to the users through a multi-factor fusion parameter network, determining, through a gated screening network, second weights of the factors, so as to accurately determine target factors suitable for the users according to the first weights and the second weights, and perform information flow recommendation. For another example, the background processing servers of the multi-factor fusion parameter network and the gating screening network are obtained based on the evolutionary strategy training according to the feedback information provided by the

terminal devices

101, 102, 103 for the recommendation result. As an example, the server 105 may be a cloud server.

The server may be hardware or software. When the server is hardware, it may be implemented as a distributed server cluster composed of multiple servers, or may be implemented as a single server. When the server is software, it may be implemented as multiple pieces of software or software modules (e.g., software or software modules for providing distributed services) or as a single piece of software or software module. And is not particularly limited herein.

It should be further noted that the information flow recommendation method and the model training method provided by the embodiments of the present disclosure may be executed by a server, or may be executed by a terminal device, or may be executed by the server and the terminal device in cooperation with each other. Accordingly, the information flow recommendation device and the model training device may include all the parts (for example, all the units) in the server, all the parts in the terminal device, and all the parts in the server and the terminal device.

It should be understood that the number of terminal devices, networks, and servers in fig. 1 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation. When the electronic device on which the information flow recommendation method and the model training method are operated does not need to perform data transmission with other electronic devices, the system architecture may only include the electronic device (e.g., a server or a terminal device) on which the information flow recommendation method and the model training method are operated.

Referring to fig. 2, fig. 2 is a flowchart of an information flow recommendation method provided in an embodiment of the present disclosure, where the process 200 includes the following steps:

step 201, acquiring characteristic information of a first user in an information flow recommendation scene.

In this embodiment, an execution subject (for example, a terminal device or a server in fig. 1) of the information flow recommendation method may obtain the feature information of the first user in the information flow recommendation scenario from a remote location or a local location based on a wired network connection manner or a wireless network connection manner.

The first user is a user to be subjected to information flow recommendation. The information stream recommendation scenario may be a recommendation scenario corresponding to various types of information streams. For example, in a news application, the information flow recommendation scenario is to determine news information flows in which a user is interested; in video applications, the information flow recommendation scene is to determine the video information flow in which the user is interested.

The feature information of the first user in the information flow recommendation scene comprises user feature information of the first user and scene feature information of the information flow recommendation scene. By way of example, the user characteristic information includes user liveness, age, gender, daily average usage duration of the product, usage times, and the like; the scene characteristic information includes a refresh state, a refresh number, a refresh time, and the like.

Step 202, determining a first weight corresponding to each factor in the factor set according to the characteristic information through a multi-factor fusion parameter network.

In this embodiment, the execution subject may determine the first weight corresponding to each factor in the factor set according to the feature information through a multi-factor fusion parameter network.

The factors in the factor set represent index information to be considered in the information flow recommendation process. For example, the factor combination includes factors such as reading duration, diversity of displayed resources, user approval amount, and sharing amount.

In this implementation, in order to improve the pertinence of information flow recommendation, different factor sets may be determined for different information flow recommendation scenarios. The corresponding factor combination of each information flow recommendation scene comprises a plurality of factors aiming at the information flow recommendation scene.

As an example, the multi-factor converged parameter network includes a plurality of tower networks, each tower network corresponding to one factor of the set of factors. At the bottom layer of the multi-factor fusion parameter network, a plurality of tower networks share characteristic information, and the plurality of tower networks are used for outputting first weights of corresponding factors. Each tower network may be a neural network, including but not limited to convolutional neural networks, cyclic neural networks, and like network models.

And step 203, determining a second weight corresponding to each factor in the factor set according to the characteristic information through the gating screening network.

In this embodiment, the execution main body may determine the second weight corresponding to each factor in the factor set according to the characteristic information through the gated screening network.

The gated screening network may be implemented based on a gated recurrent neural network. As an example, the execution subject may determine the network structure of the gated screening network according to the number of factors included in the factor set, so that the number of outputs of the gated screening network coincides with the number of factors included in the factor set, and the factors included in the factor set correspond to the outputs of the gated screening network one to one. And the plurality of outputs of the gating screening network are the second weights corresponding to the factors in the factor set.

And 204, determining a target factor suitable for the first user in the information flow recommendation scene in the factor set according to the first weight and the second weight.

In this embodiment, the execution subject may determine, according to the first weight and the second weight, a target factor applicable to the first user in the information flow recommendation scenario in the factor set.

As an example, for each factor in the factor set, the execution subject may determine a first weight corresponding to the factor and a second weight corresponding to the factor, and determine a total weight by summing, weighted summing, or the like; and then sequencing the total weight, and determining the factors with the preset number sequenced at the top as target factors, or determining the factors with the total weight larger than a preset numerical value as the target factors.

Step 205, according to the target factor, determining and pushing a recommendation result corresponding to the first user in the information stream recommendation scene to the first user.

In this embodiment, the execution main body may determine and push a recommendation result corresponding to the first user in an information stream recommendation scenario to the first user according to the target factor.

As an example, for each determined target factor, the executing entity may first obtain a total weight of the target factor based on a first weight and a second weight corresponding to the target factor; and then, based on the corresponding target factors and the total weight, obtaining a weighted item corresponding to each target factor, and combining the weighted items of the target factors to obtain a multi-target factor fusion formula.

After the multi-target factor fusion formula is obtained, the recommendation ranking score of the contents to be ranked in the preset contents to be recommended set can be determined through the multi-target factor fusion formula; and sequencing the contents to be sequenced in the content set to be sequenced based on the recommendation sequencing score so as to take the contents to be sequenced with the preset number in the top sequence as a recommendation result corresponding to the first user and push the recommendation result to the first user.

With continued reference to fig. 3, fig. 3 is a schematic diagram 300 of an application scenario of the information flow recommendation method according to the present embodiment. In the application scenario of fig. 3, a user 301 issues a start instruction to a short video-class application through a terminal device 302. The server 303 firstly acquires the characteristic information of the user 301 in the short video information stream recommendation scene based on the opening instruction; then, determining a first weight 305 corresponding to each factor in the factor set according to the characteristic information through a multi-factor fusion parameter network 304, wherein the factor in the factor set represents index information to be considered in the information flow recommendation process; then, determining a second weight 307 corresponding to each factor in the factor set according to the characteristic information through a gating screening network 306; then, according to the first weight 305 and the second weight 307, determining a target factor 308 suitable for the first user in the information flow recommendation scene in the factor set; and determining and pushing a recommendation result 309 corresponding to the first user in the information stream recommendation scene to the user 301 according to the target factor 308.

In the embodiment, in an information flow recommendation scene, a first weight of each factor corresponding to a user is determined through a multi-factor fusion parameter network, a second weight of each factor is determined through a gating screening network, so that a target factor suitable for the user is accurately determined according to the first weight and the second weight, information flow recommendation is performed, and accuracy of a recommendation result is improved.

In some optional implementations of this embodiment, the executing main body may execute the step 203 by:

firstly, determining initial second weights corresponding to all factors in a factor set according to characteristic information through a gating screening network; and then, converting the initial second weight corresponding to each factor into 0/1 through a preset activation function to obtain the second weight corresponding to each factor.

As an example, the value range of the preset activation function at the uppermost layer of the gated screening network is 0/1, so that the continuous value problem is skillfully converted into a 0/1 problem. As an example, the preset activation function may be:

in the implementation mode, for each factor in the factor set, the second weight output by the gating screening network is 0/1, so that the continuous value problem is ingeniously converted into the 0/1 problem, and the high efficiency and convenience of the determination process of the target factor are improved.

In some optional implementations of this embodiment, the executing main body may execute the step 204 by:

firstly, multiplying a first weight and a second weight corresponding to each factor in a factor set to obtain a weight product; and then, determining a target factor suitable for the first user under the information flow recommendation scene in the factor set according to the weight product corresponding to each factor in the factor set.

When the second weight output by the gating screening network is 0/1, the weight product corresponding to each factor can be conveniently determined to be the first weight or zero corresponding to the factor; and further, removing the factor with the weight product being zero, and keeping the factor with the weight product being non-zero to obtain the target factor.

In the implementation mode, the target factor is determined based on the weight product of the first weight and the second weight corresponding to each factor, and the convenience and the accuracy of the target factor determination process are further improved.

In some optional implementations of this embodiment, the execution main body may further perform the following operations: firstly, feedback information of a first user on a recommendation result is obtained.

The feedback information may be reflection information of the information flow in the recommendation result after the first user obtains the recommendation result. By way of example, the feedback information includes whether to click, whether to check, whether to make praise, comment, and other interactive operations.

And secondly, adjusting parameters of the multi-factor fusion parameter network and parameters of the gated screening network according to the feedback information by adopting an evolution strategy so as to execute a recommendation task of a subsequent user in an information flow recommendation scene through the adjusted multi-factor fusion parameter network and the gated screening network.

The evolutionary strategy algorithm is an algorithm based on an evolutionary theory, and the algorithm can be used for exploring and searching parameter disturbance which enables overall return of a multi-factor fusion parameter network and a gating screening network to be larger. Specifically, according to the feedback information and a preset return function, parameters of the multi-factor fusion parameter network and return values of parameters of the gating screening network are determined; guiding the adjustment process of the parameters of the multi-factor fusion parameter network and the parameters of the gating screening network based on the principle of the maximization of the return value; and iterating according to preset iteration times based on a preset evolutionary strategy algorithm to generate parameters of a multi-factor fusion parameter network and parameters of a gating screening network which satisfy the new round of Gaussian distribution of the mean value and the variance of each parameter.

And executing the follow-up user recommendation task by adopting the multi-factor fusion parameter network and the gating screening network after the parameters are adjusted.

In this implementation manner, the execution body adopts an evolutionary strategy to adjust the multi-factor fusion parameter network and the gated screening network in the application process of the multi-factor fusion parameter network and the gated screening network, so that the accuracy of the multi-factor fusion parameter network and the gated screening network can be continuously improved.

With continued reference to fig. 4, an exemplary flow 400 of yet another embodiment of an information flow recommendation method according to the present disclosure is shown, comprising the steps of:

step 401, obtaining feature information of a first user in an information flow recommendation scene.

And 402, determining a first weight corresponding to each factor in the factor set according to the characteristic information through a multi-factor fusion parameter network.

The factors in the factor set represent index information to be considered in the information flow recommendation process.

And step 403, determining initial second weights corresponding to the factors in the factor set according to the characteristic information through a gated screening network.

Step 404, converting the initial second weight corresponding to each factor into 0/1 through a preset activation function, so as to obtain the second weight corresponding to each factor.

Step 405, multiplying the first weight and the second weight corresponding to each factor in the factor set to obtain a weight product.

And step 406, determining a target factor applicable to the first user in the information flow recommendation scene in the factor set according to the weight product corresponding to each factor in the factor set.

Step 407, determining and pushing a recommendation result corresponding to the first user in the information stream recommendation scene to the first user according to the target factor.

As can be seen from this embodiment, compared with the embodiment corresponding to fig. 2, the flow 400 of the information flow recommendation method in this embodiment specifically illustrates the determining process of the second weight and the determining process of the target factor, so that the efficiency and convenience of the determining process of the target factor are further improved, and the accuracy of the recommendation result is improved.

With continuing reference to FIG. 5, an exemplary flow 500 of one embodiment of a model training method according to the present disclosure is shown, comprising the steps of:

step 501, obtaining characteristic information of a second user in an information flow recommendation scene.

In this embodiment, an executing entity (for example, a terminal device or a server in fig. 1) of the model training method may obtain the feature information of the second user in the information flow recommendation scenario from a remote location or a local location based on a wired network connection manner or a wireless network connection manner.

And the second user is a user to be subjected to information flow recommendation in the training process of the initial multi-factor fusion parameter network and the initial gating screening network. In the model training process, a plurality of second users are typically involved. The training process shown in steps 501-506 may be performed for each second user.

The information stream recommendation scene may be a recommendation scene corresponding to various types of information streams. For example, in a news application, an information flow recommendation scene is to determine news information flows of users; in the video application, the information flow recommendation scene is the video information flow of a certain user.

The feature information of the second user in the information flow recommendation scene comprises the user feature information of the second user and the scene feature information of the information flow recommendation scene. By way of example, the user characteristic information includes user liveness, age, gender, daily average usage duration of the product, usage times, and the like; the scene characteristic information includes a refresh state, refresh times, refresh time, and the like.

Step 502, determining a first weight corresponding to each factor in the factor set according to the characteristic information through an initial multi-factor fusion parameter network.

In this embodiment, the execution subject may determine the first weight corresponding to each factor in the factor set according to the feature information through an initial multi-factor fusion parameter network.

In this implementation, in order to improve the pertinence of information flow recommendation, different factor sets may be determined for different information flow recommendation scenarios. The corresponding factor set for each information stream recommendation scenario includes a plurality of factors for the information stream recommendation scenario.

As an example, the initial multi-factor converged parameter network includes a plurality of tower networks, each tower network corresponding to one factor of the set of factors. At the bottom layer of the initial multi-factor fusion parameter network, a plurality of tower networks share characteristic information, and the plurality of tower networks are used for outputting first weights of corresponding factors. Each tower network may be a neural network, including but not limited to convolutional neural networks, cyclic neural networks, and like network models.

And 503, determining a second weight corresponding to each factor in the factor set according to the characteristic information through the initial gating screening network.

In this embodiment, the execution main body may determine the second weight corresponding to each factor in the factor set according to the characteristic information through the initial gating screening network.

The initial gated screening network may be implemented based on a gated recurrent neural network. As an example, the execution subject may determine the network structure of the initial gated screening network according to the number of factors included in the factor set, so that the number of outputs of the initial gated screening network coincides with the number of factors included in the factor set, and the factors included in the factor set correspond to the outputs of the initial gated screening network one to one. And a plurality of outputs of the initial gating screening network are second weights corresponding to the factors in the factor set.

And step 504, determining a target factor suitable for the second user under the information flow recommendation scene in the factor set according to the first weight and the second weight.

In this embodiment, the execution subject may determine, according to the first weight and the second weight, a target factor applicable to a second user in an information flow recommendation scenario in the factor set.

As an example, for each factor in the factor set, the execution subject may determine a first weight corresponding to the factor and a second weight corresponding to the factor, and determine the total weight by summing, weighted summing, or the like; and then sequencing the total weight, and determining the factors with the preset number sequenced at the top as target factors, or determining the factors with the total weight larger than a preset numerical value as the target factors.

And 505, determining a recommendation result corresponding to the second user in the information flow recommendation scene according to the target factor.

In this embodiment, the execution subject may determine, according to the target factor, a recommendation result corresponding to the second user in the information flow recommendation scenario.

As an example, for each determined target factor, the executing entity may first obtain a total weight of the target factor based on the first weight and the second weight corresponding to the target factor; and further, based on the corresponding target factors and the total weight, obtaining a weighted item corresponding to each target factor, and combining the weighted items of the target factors to obtain a multi-target factor fusion formula.

After the multi-target factor fusion formula is obtained, the recommendation ranking score of the contents to be ranked in the preset contents to be recommended set can be determined through the multi-target factor fusion formula; and sequencing the contents to be sequenced in the content set to be sequenced based on the recommendation sequencing score so as to take the contents to be sequenced with the preset number in the top sequence as a recommendation result corresponding to the second user and push the recommendation result to the second user.

Step 506, adjusting parameters of the initial multi-factor fusion parameter network and parameters of the initial gating screening network according to feedback information of the second user on the recommendation result by adopting an evolution strategy so as to obtain the trained multi-factor fusion parameter network and the trained gating screening network.

In this embodiment, the execution body may adjust parameters of the initial multi-factor fusion parameter network and parameters of the initial gated screening network according to feedback information of the second user on the recommendation result by using an evolution strategy, so as to obtain the trained multi-factor fusion parameter network and gated screening network.

Specifically, according to feedback information and a preset return function, parameters of an initial multi-factor fusion parameter network and return values of parameters of an initial gating screening network are determined; guiding the adjustment process of the parameters of the initial multi-factor fusion parameter network and the parameters of the initial gating screening network based on the principle of the maximization of the return value; and iterating according to preset iteration times based on a preset evolutionary strategy algorithm to generate parameters of a multi-factor fusion parameter network and parameters of a gating screening network which meet the new round of Gaussian distribution of the mean value and the variance of each parameter, and taking the adjusted multi-factor fusion parameter network and the gating screening network as an initial multi-factor fusion parameter network and an initial gating screening network for next round of training.

And performing the training operation through iteration, and responding to the preset end condition, so as to obtain the trained multi-factor fusion parameter network and the gated screening network. The preset ending condition may be, for example, that the iteration number exceeds a preset number threshold, the training time exceeds a preset time threshold, and the like.

In the embodiment, the target factors are screened based on the gated screening network, so that the evolution efficiency of the evolution strategy is effectively improved, and meanwhile, the target factors beneficial to the global situation can be automatically screened by triggering from the perspective of global optimization.

In some optional implementations of this embodiment, the executing main body may execute the step 503 by:

firstly, determining initial second weights corresponding to all factors in a factor set according to characteristic information through an initial gating screening network; and then, converting the initial second weight corresponding to each factor into 0/1 through a preset activation function to obtain the second weight corresponding to each factor.

As an example, the value range of the preset activation function at the uppermost layer of the initial gated screening network is 0/1, so that the continuous value problem is ingeniously converted into the 0/1 problem. As an example, the preset activation function may be:

in the implementation mode, for each factor in the factor set, the second weight output by the initial gating screening network is 0/1, so that the continuous value problem is ingeniously converted into a 0/1 problem, and the evolution efficiency of the evolution strategy is further improved.

In some optional implementations of this embodiment, the executing main body may execute the step 504 as follows:

firstly, multiplying a first weight and a second weight corresponding to each factor in a factor set to obtain a weight product; and then, determining a target factor suitable for a second user under the information flow recommendation scene in the factor set according to the weight product corresponding to each factor in the factor set.

When the second weight output by the initial gating screening network is 0/1, the weight product corresponding to each factor can be conveniently determined to be the first weight or zero corresponding to the factor; and further, removing the factor with the weight product being zero, and keeping the factor with the weight product being non-zero to obtain the target factor.

In the embodiment, the target factor is determined based on the weight product of the first weight and the second weight corresponding to each factor, so that the convenience and the accuracy of the target factor determination process in the model training process are further improved.

With continuing reference to fig. 6, as an implementation of the method shown in the above-mentioned figures, the present disclosure provides an embodiment of an information flow recommendation apparatus, which corresponds to the embodiment of the method shown in fig. 2, and which can be applied in various electronic devices.

As shown in fig. 6, the information flow recommendation apparatus 600 includes: a first obtaining unit 601 configured to obtain feature information of a first user in an information flow recommendation scenario; a first determining unit 602, configured to determine, through a multi-factor fusion parameter network, a first weight corresponding to each factor in a factor set according to the feature information, where the factor in the factor set represents index information to be considered in an information flow recommendation process; a second determining unit 603 configured to determine, according to the feature information, a second weight corresponding to each factor in the factor set through the gated screening network; a third determining unit 604, configured to determine, according to the first weight and the second weight, a target factor applicable to the first user in the information flow recommendation scenario in the factor set; the recommending unit 605 is configured to determine and push a recommendation result corresponding to the first user in the information stream recommendation scenario to the first user according to the target factor.

In some optional implementations of this embodiment, the second determining unit 603 is further configured to: determining initial second weights corresponding to all factors in the factor set according to the characteristic information through a gating screening network; and converting the initial second weight corresponding to each factor into 0/1 through a preset activation function to obtain the second weight corresponding to each factor.

In some optional implementations of this embodiment, the third determining unit 604 is further configured to: multiplying the first weight and the second weight corresponding to each factor in the factor set to obtain a weight product; and determining a target factor suitable for the first user under the information flow recommendation scene in the factor set according to the weight product corresponding to each factor in the factor set.

In some optional implementations of this embodiment, the apparatus further includes: a feedback unit (not shown in the figure) configured to acquire feedback information of the first user on the recommendation result; and the evolution unit (not shown in the figure) is configured to adjust the parameters of the multi-factor fusion parameter network and the parameters of the gated screening network according to the feedback information by adopting an evolution strategy so as to execute the recommendation task of the subsequent user in the information flow recommendation scene through the adjusted multi-factor fusion parameter network and the gated screening network.

In the embodiment, an information flow recommendation device is provided, in an information flow recommendation scene, a first weight of each factor corresponding to a user is determined through a multi-factor fusion parameter network, a second weight of each factor is determined through a gated screening network, so that a target factor suitable for the user is accurately determined according to the first weight and the second weight, information flow recommendation is performed, and accuracy of a recommendation result is improved.

With continuing reference to fig. 7, as an implementation of the method shown in the above figures, the present disclosure provides an embodiment of a model training apparatus, which corresponds to the embodiment of the method shown in fig. 5, and which can be applied in various electronic devices.

As shown in fig. 7, the model training apparatus 700 includes: a second obtaining unit 701 configured to obtain feature information of a second user in an information flow recommendation scenario; a fourth determining unit 702, configured to determine, according to the feature information, a first weight corresponding to each factor in a factor set through an initial multi-factor fusion parameter network, where the factor in the factor set represents index information to be considered in an information flow recommendation process; a fifth determining unit 703 configured to determine, according to the feature information, a second weight corresponding to each factor in the factor set through the initial gating screening network; a sixth determining unit 704, configured to determine, according to the first weight and the second weight, a target factor applicable to a second user in the information flow recommendation scenario in the factor set; a seventh determining unit 705, configured to determine, according to the target factor, a recommendation result corresponding to a second user in the information flow recommendation scenario; the training unit 706 is configured to adjust parameters of the initial multi-factor fusion parameter network and parameters of the initial gating screening network according to feedback information of the second user on the recommendation result by using an evolution strategy, so as to obtain a trained multi-factor fusion parameter network and a trained gating screening network.

In some optional implementations of the present embodiment, the fifth determining unit 703 is further configured to: determining initial second weights corresponding to all factors in the factor set according to the characteristic information through an initial gating screening network; and converting the initial second weight corresponding to each factor into 0/1 through a preset activation function to obtain the second weight corresponding to each factor.

In some optional implementations of this embodiment, the sixth determining unit 704 is further configured to: multiplying the first weight and the second weight corresponding to each factor in the factor set to obtain a weight product; and determining a target factor suitable for a second user under the information flow recommendation scene in the factor set according to the weight product corresponding to each factor in the factor set.

In this embodiment, the target factor is determined based on the weight product of the first weight and the second weight corresponding to each factor, so that convenience and accuracy of the target factor determination process in the model training process are further improved.

According to an embodiment of the present disclosure, the present disclosure also provides an electronic device including: at least one processor; and a memory communicatively coupled to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor, so that the at least one processor can implement the information flow recommendation method and the model training method described in any of the above embodiments.

According to an embodiment of the present disclosure, the present disclosure further provides a readable storage medium, which stores computer instructions for enabling a computer to implement the information flow recommendation method and the model training method described in any of the above embodiments when executed.

The embodiments of the present disclosure provide a computer program product, which when executed by a processor can implement the information flow recommendation method and the model training method described in any of the embodiments above.

FIG. 8 illustrates a schematic block diagram of an example electronic device 800 that can be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not intended to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 8, the apparatus 800 includes a computing unit 801 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM) 802 or a computer program loaded from a storage unit 808 into a Random Access Memory (RAM) 803. In the RAM 803, various programs and data necessary for the operation of the device 800 can also be stored. The calculation unit 801, the ROM 802, and the RAM 803 are connected to each other by a bus 804. An input/output (I/O) interface 805 is also connected to bus 804.

A number of components in the device 800 are connected to the I/O interface 805, including: an input unit 806, such as a keyboard, a mouse, or the like; an output unit 807 such as various types of displays, speakers, and the like; a storage unit 808, such as a magnetic disk, optical disk, or the like; and a communication unit 809 such as a network card, modem, wireless communication transceiver, etc. The communication unit 809 allows the device 800 to exchange information/data with other devices via a computer network such as the internet and/or various telecommunication networks.

Computing unit 801 may be a variety of general and/or special purpose processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various dedicated Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, and so forth. The calculation unit 801 executes the respective methods and processes described above, such as the information flow recommendation method. For example, in some embodiments, the information flow recommendation method may be implemented as a computer software program tangibly embodied in a machine-readable medium, such as storage unit 808. In some embodiments, part or all of the computer program can be loaded and/or installed onto device 800 via ROM 802 and/or communications unit 809. When loaded into RAM 803 and executed by computing unit 801, may perform one or more of the steps of the information flow recommendation methods described above. Alternatively, in other embodiments, the computing unit 801 may be configured to perform the information flow recommendation method in any other suitable manner (e.g., by means of firmware).

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), system on a chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the program code, when executed by the processor or controller, causes the functions/acts specified in the flowchart and/or block diagram to be performed. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, 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 compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), and the Internet.

The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The Server can be a cloud Server, also called a cloud computing Server or a cloud host, and is a host product in a cloud computing service system, so as to solve the defects of large management difficulty and weak service expansibility existing in the traditional physical host and Virtual Private Server (VPS) service; it may also be a server of a distributed system, or a server incorporating a blockchain.

According to the technical scheme of the embodiment of the disclosure, an information flow recommendation method is provided, and in an information flow recommendation scene, a first weight of each factor corresponding to a user is determined through a multi-factor fusion parameter network, a second weight of each factor is determined through a gating screening network, so that a target factor suitable for the user is accurately determined according to the first weight and the second weight, information flow recommendation is performed, and the accuracy of a recommendation result is improved.

It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in this disclosure may be performed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions provided by this disclosure can be achieved, which are not limited herein.

The above detailed description should not be construed as limiting the scope of the disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present disclosure should be included in the scope of protection of the present disclosure.

Claims

1. An information flow recommendation method, comprising:

acquiring characteristic information of a first user in an information flow recommendation scene;

determining a first weight corresponding to each factor in a factor set according to the characteristic information through a multi-factor fusion parameter network, wherein the factor in the factor set represents index information required to be considered in the information flow recommendation process;

determining a second weight corresponding to each factor in the factor set according to the characteristic information through a gating screening network;

determining a target factor in the factor set, which is suitable for the first user under the information flow recommendation scene, according to the first weight and the second weight;

and determining and pushing a recommendation result corresponding to the first user in the information flow recommendation scene to the first user according to the target factor.

2. The method of claim 1, wherein determining, by the gated screening network, the second weight corresponding to each factor in the set of factors from the feature information comprises:

determining initial second weights corresponding to the factors in the factor set according to the characteristic information through the gating screening network;

and converting the initial second weight corresponding to each factor into 0/1 through a preset activation function to obtain the second weight corresponding to each factor.

3. The method according to claim 1 or 2, wherein the determining, according to the first weight and the second weight, a target factor of the set of factors that is applicable to the first user under the information flow recommendation scenario includes:

multiplying a first weight and a second weight corresponding to each factor in the factor set to obtain a weight product;

and determining a target factor suitable for the first user under the information flow recommendation scene in the factor set according to the weight product corresponding to each factor in the factor set.

4. The method of claim 1, further comprising:

obtaining feedback information of the first user on the recommendation result;

and adjusting parameters of the multi-factor fusion parameter network and parameters of the gated screening network according to the feedback information by adopting an evolution strategy so as to execute a recommendation task of a subsequent user in the information flow recommendation scene through the adjusted multi-factor fusion parameter network and the gated screening network.

5. A model training method, comprising:

acquiring characteristic information of a second user in an information flow recommendation scene;

determining a first weight corresponding to each factor in a factor set according to the characteristic information through an initial multi-factor fusion parameter network, wherein the factor in the factor set represents index information required to be considered in the information flow recommendation process;

determining second weights corresponding to all factors in the factor set according to the characteristic information through an initial gating screening network;

determining a target factor in the factor set, which is suitable for the second user under the information flow recommendation scene, according to the first weight and the second weight;

determining a recommendation result corresponding to the second user in the information flow recommendation scene according to the target factor;

and adjusting parameters of the initial multi-factor fusion parameter network and parameters of the initial gating screening network according to feedback information of the second user on the recommendation result by adopting an evolution strategy to obtain the trained multi-factor fusion parameter network and the trained gating screening network.

6. The method of claim 5, wherein determining, by the initial gated screening network, the second weight corresponding to each factor in the set of factors from the feature information comprises:

determining initial second weights corresponding to the factors in the factor set according to the characteristic information through the initial gating screening network;

7. The method according to claim 5 or 6, wherein the determining, according to the first weight and the second weight, a target factor of the set of factors suitable for the second user in the information flow recommendation scenario comprises:

multiplying the first weight and the second weight corresponding to each factor in the factor set to obtain a weight product;

and determining a target factor, which is applicable to the second user under the information flow recommendation scene, in the factor set according to the weight product corresponding to each factor in the factor set.

8. An information flow recommendation device comprising:

the information flow recommendation system comprises a first acquisition unit, a second acquisition unit and a third acquisition unit, wherein the first acquisition unit is configured to acquire characteristic information of a first user in an information flow recommendation scene;

a first determining unit, configured to determine, through a multi-factor fusion parameter network, a first weight corresponding to each factor in a factor set according to the feature information, where the factor in the factor set represents index information to be considered in an information flow recommendation process;

a second determining unit configured to determine, through a gated screening network, a second weight corresponding to each factor in the factor set according to the feature information;

a third determining unit, configured to determine, according to the first weight and the second weight, a target factor in the factor set that is applicable to the first user in the information flow recommendation scenario;

and the recommending unit is configured to determine and push a recommending result corresponding to the first user in the information flow recommending scene to the first user according to the target factor.

9. The apparatus of claim 8, wherein the second determining unit is further configured to:

determining initial second weights corresponding to the factors in the factor set according to the characteristic information through the gating screening network; and converting the initial second weight corresponding to each factor into 0/1 through a preset activation function to obtain the second weight corresponding to each factor.

10. The apparatus of claim 8 or 9, wherein the third determining unit is further configured to:

multiplying a first weight and a second weight corresponding to each factor in the factor set to obtain a weight product; and determining the target factor, which is applicable to the first user under the information flow recommendation scene, in the factor set according to the weight product corresponding to each factor in the factor set.

11. The apparatus of claim 8, further comprising:

a feedback unit configured to acquire feedback information of the first user on the recommendation result;

and the evolution unit is configured to adjust the parameters of the multi-factor fusion parameter network and the parameters of the gated screening network according to the feedback information by adopting an evolution strategy so as to execute the recommendation task of the subsequent user in the information flow recommendation scene through the adjusted multi-factor fusion parameter network and the gated screening network.

12. A model training apparatus comprising:

the second acquisition unit is configured to acquire characteristic information of a second user in an information flow recommendation scene;

a fourth determining unit, configured to determine, according to the feature information, a first weight corresponding to each factor in a factor set through an initial multi-factor fusion parameter network, where the factor in the factor set represents index information that needs to be considered in an information stream recommendation process;

a fifth determining unit configured to determine, through an initial gated screening network, a second weight corresponding to each factor in the factor set according to the feature information;

a sixth determining unit, configured to determine, according to the first weight and the second weight, a target factor in the factor set that is applicable to the second user in the information flow recommendation scenario;

a seventh determining unit, configured to determine, according to the target factor, a recommendation result corresponding to the second user in the information flow recommendation scenario;

and the training unit is configured to adjust the parameters of the initial multi-factor fusion parameter network and the parameters of the initial gating screening network according to the feedback information of the second user on the recommendation result by adopting an evolution strategy so as to obtain the trained multi-factor fusion parameter network and the trained gating screening network.

13. The apparatus of claim 12, wherein the fifth determining unit is further configured to:

determining initial second weights corresponding to the factors in the factor set according to the characteristic information through the initial gating screening network; and converting the initial second weight corresponding to each factor into 0/1 through a preset activation function to obtain the second weight corresponding to each factor.

14. The apparatus of claim 12 or 13, wherein the sixth determining unit is further configured to:

multiplying a first weight and a second weight corresponding to each factor in the factor set to obtain a weight product; and determining a target factor, which is applicable to the second user under the information flow recommendation scene, in the factor set according to the weight product corresponding to each factor in the factor set.

15. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7.

16. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method of any one of claims 1-7.

17. A computer program product, comprising: computer program which, when being executed by a processor, carries out the method according to any one of claims 1-7.