WO2024016680A1 - Procédé et appareil de recommandation de flux d'informations et produit programme d'ordinateur - Google Patents

Procédé et appareil de recommandation de flux d'informations et produit programme d'ordinateur Download PDF

Info

Publication number
WO2024016680A1
WO2024016680A1 PCT/CN2023/080416 CN2023080416W WO2024016680A1 WO 2024016680 A1 WO2024016680 A1 WO 2024016680A1 CN 2023080416 W CN2023080416 W CN 2023080416W WO 2024016680 A1 WO2024016680 A1 WO 2024016680A1
Authority
WO
WIPO (PCT)
Prior art keywords
factor
weight
user
information
information flow
Prior art date
Application number
PCT/CN2023/080416
Other languages
English (en)
Chinese (zh)
Inventor
邓罗丹
Original Assignee
百度在线网络技术（北京）有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 百度在线网络技术（北京）有限公司 filed Critical 百度在线网络技术（北京）有限公司
Publication of WO2024016680A1 publication Critical patent/WO2024016680A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation

Definitions

  • the present disclosure relates to the field of artificial intelligence technology, specifically to the field of evolutionary strategy technology, and in particular to information flow recommendation methods, devices and model training methods, devices, electronic devices, storage media and computer program products, which can be used in information flow recommendation scenarios.
  • Information flow recommendation is different from advertising. It not only focuses on the click-to-view ratio of resources, but also integrates a series of experience indicators such as reading time, diversity of displayed resources, number of user likes, and number of shares as comprehensive recommendation indicators. Although there are more and more target factors for fusion, different fusion factors have their own applicable scenario restrictions. For example, the primary task of the new user model is to promote activation and attract new users, and goals such as duration and diversity are not the key concerns of the system. How to perform adaptive factor screening for the scenarios faced by users is a common problem in information flow recommendation systems.
  • the present disclosure provides an information flow recommendation method and device, as well as a model training method, device, electronic equipment, storage media and computer program products.
  • an information flow recommendation method which includes: obtaining the characteristic information of the first user in an information flow recommendation scenario; and determining, through a multi-factor fusion parameter network, the third factor corresponding to each factor in the factor set according to the characteristic information.
  • a model training method which includes: obtaining the characteristic information of the second user in an information flow recommendation scenario; and determining the third factor corresponding to each factor in the factor set according to the characteristic information through an initial multi-factor fusion parameter network.
  • an information flow recommendation device including: a first acquisition unit configured to obtain the characteristic information of the first user in an information flow recommendation scenario; a first determination unit configured to obtain the feature information through multi-factor fusion
  • the parameter network determines the first weight corresponding to each factor in the factor set based on the characteristic information, where the factors in the factor set represent the indicator information that needs to be considered in the information flow recommendation process;
  • the second determination unit is configured to pass the gating Screen the network and determine the second weight corresponding to each factor in the factor set according to the characteristic information;
  • the third determination unit is configured to determine the first weight in the factor set suitable for the information flow recommendation scenario based on the first weight and the second weight.
  • the user's target factor; the recommendation unit is configured to determine and push the recommendation result corresponding to the first user in the information flow recommendation scenario to the first user based on the target factor.
  • a model training device including: a second acquisition unit configured to acquire characteristic information of the second user in an information flow recommendation scenario; a fourth determination unit configured to obtain the feature information through initial multi-factor fusion
  • the parameter network determines the first weight corresponding to each factor in the factor set based on the characteristic information, where the factors in the factor set represent the indicator information that needs to be considered in the information flow recommendation process; the fifth determination unit is configured to pass the initial gate
  • the control screening network determines the second weight corresponding to each factor in the factor set according to the characteristic information; the sixth determination unit is configured to determine the third weight in the factor set suitable for the information flow recommendation scenario based on the first weight and the second weight.
  • the seventh determination unit is configured to determine the recommendation results corresponding to the second user in the information flow recommendation scenario according to the target factors;
  • the training unit is configured to use an evolutionary strategy to determine the recommendation results according to the second user's Feedback information, adjust the parameters of the initial multi-factor fusion parameter network and the parameters of the initial gate screening network to obtain the multi-factor fusion parameters after training number network and gated screening network.
  • an electronic device including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by at least one processor, and the instructions are processed by at least one The processor executes, so that at least one processor can execute the method described in any implementation manner of the first aspect and the second aspect.
  • a non-transitory computer-readable storage medium storing computer instructions
  • the computer instructions are used to cause the computer to execute the method described in any implementation manner of the first aspect or the second aspect.
  • a computer program product including: a computer program. When executed by a processor, the computer program implements the method described in any implementation manner of the first aspect or the second aspect.
  • an information flow recommendation method is provided.
  • the first weight of each factor corresponding to the user is determined through a multi-factor fusion parameter network, and the first weight of each factor is determined through a gated screening network.
  • Two weights are used to accurately determine the target factors suitable for users in the information flow recommendation scenario based on the first weight and the second weight, and perform information flow recommendation, thereby improving the accuracy of the recommendation results.
  • FIG. 1 is an exemplary system architecture diagram to which an embodiment of the present disclosure may be applied;
  • Figure 2 is a flow chart of an embodiment of an information flow recommendation method according to the present disclosure
  • Figure 3 is a schematic diagram of an application scenario of the information flow recommendation method according to this embodiment.
  • Figure 4 is a flow chart of yet another embodiment of an information flow recommendation method according to the present disclosure.
  • Figure 5 is a flow chart of one embodiment of a model training method according to the present disclosure.
  • Figure 6 is a structural diagram of an embodiment of an information flow recommendation device according to the present disclosure.
  • Figure 7 is a structural diagram of an embodiment of a model training device according to the present disclosure.
  • FIG. 8 is a schematic structural diagram of a computer system suitable for implementing embodiments of the present disclosure.
  • the collection, storage, use, processing, transmission, provision and disclosure of user personal information are in compliance with relevant laws and regulations and do not violate public order and good customs.
  • Figure 1 shows an exemplary architecture 100 in which the information flow recommendation method and device, and the model training method and device of the present disclosure can be applied.
  • the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105.
  • the communication connections between terminal devices 101, 102, and 103 constitute a topological network, and the network 104 is used to provide a medium for communication links between the terminal devices 101, 102, and 103 and the server 105.
  • Network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, among others.
  • the terminal devices 101, 102, and 103 may be hardware devices or software that support network connection for data interaction and data processing.
  • the terminal devices 101, 102, and 103 are hardware, they can be various electronic devices that support network connection, information acquisition, interaction, display, processing and other functions, including but not limited to smart phones, tablet computers, e-book readers, Laptops and desktop computers and more.
  • the terminal devices 101, 102, and 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 for providing distributed services, or as a single software or software module. There are no specific limitations here.
  • the server 105 may be a server that provides various services. For example, based on the characteristic information of the users corresponding to the terminal devices 101, 102, 103 in the information flow recommendation scenario, the multi-factor fusion parameter network determines the third factor corresponding to each factor of the user. One weight, determine the second weight of each factor through the gated screening network, so as to accurately determine the target factor suitable for the user based on the first weight and the second weight, and perform the background processing server for information flow recommendation. For another example, according to the terminal device 101, The feedback information for the recommendation results provided by 102 and 103 is based on the evolutionary strategy training to obtain the background processing server of the multi-factor fusion parameter network and the gated screening network. As an example, server 105 may be a cloud server.
  • the server can be hardware or software.
  • the server can be implemented as a distributed server cluster composed of multiple servers or as a single server.
  • the server is software, it can be implemented as multiple software or software modules (for example, software or software modules used to provide distributed services), or it can be implemented as a single software or software module. There are no specific limitations here.
  • the information flow recommendation method and model training method provided by the embodiments of the present disclosure can be executed by the server or by the terminal device, or can be executed by the server and the terminal device in cooperation with each other.
  • various parts (for example, each unit) included in the information flow recommendation device and the model training device can be all installed in the server, or they can all be installed in the terminal device, or they can be installed in the server and the terminal device respectively.
  • the number of terminal devices, networks and servers in Figure 1 is only illustrative. Depending on implementation needs, there can be any number of end devices, networks, and servers.
  • the system architecture may only include the electronic device on which the information flow recommendation method and the model training method run (for example, server or terminal device).
  • FIG. 2 is a flow chart of an information flow recommendation method provided by an embodiment of the present disclosure.
  • the process 200 includes the following steps:
  • Step 201 Obtain the characteristic information of the first user in the information flow recommendation scenario.
  • the execution subject of the information flow recommendation method can obtain the information flow recommendation scenario of the first user remotely or locally based on a wired network connection or a wireless network connection. feature information below.
  • Information flow recommendation scenarios can be recommendation scenarios corresponding to various types of information flows. For example, in news applications, the information flow recommendation scenario is to determine the news information flow that the user is interested in; in video applications, the information flow recommendation scenario is to determine the video information flow that the user is interested in.
  • the characteristic information of the first user in the information flow recommendation scenario includes the user characteristics of the first user Scene feature information of information and information flow recommendation scenarios.
  • user characteristic information includes user activity, age, gender, average daily product usage time, number of uses, etc.
  • scene characteristic information includes refresh status, refresh times, refresh time, etc.
  • Step 202 Determine the first weight corresponding to each factor in the factor set according to the feature information through the multi-factor fusion parameter network.
  • the above-mentioned execution subject can determine the first weight corresponding to each factor in the factor set according to the feature information through the multi-factor fusion parameter network.
  • the factors in the factor set represent the indicator information that needs to be considered in the information flow recommendation process.
  • the factor set includes factors such as reading time, diversity of displayed resources, number of user likes, number of shares, etc.
  • the factor set corresponding to each information flow recommendation scenario includes multiple factors for the information flow recommendation scenario.
  • a multi-factor fusion parameter network includes multiple tower networks, each tower network corresponding to a factor in the factor set. At the bottom of the multi-factor fusion parameter network, multiple tower networks share feature information, and multiple tower networks are used to output the first weight of the corresponding factor.
  • Each tower network can be a neural network, including but not limited to convolutional neural network, recurrent neural network and other network models.
  • Step 203 Determine the second weight corresponding to each factor in the factor set according to the characteristic information through the gated screening network.
  • the above-mentioned execution subject can determine the second weight corresponding to each factor in the factor set according to the characteristic information through the gated screening network.
  • Gated screening networks can be implemented based on gated recurrent neural networks.
  • the above execution subject can 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 is consistent with the number of factors included in the factor set, and the factor set
  • the factors included in correspond to the output of the gated screening network.
  • the multiple outputs of the gated screening network are the second weights corresponding to each factor in the factor set.
  • Step 204 Based on the first weight and the second weight, determine the target factor in the factor set that is suitable for the first user in the information flow recommendation scenario.
  • the above execution subject can determine the factor based on the first weight and the second weight.
  • the target factor in the set that is suitable for the first user in the information flow recommendation scenario.
  • the above-mentioned executive body can determine the first weight corresponding to the factor and the second weight corresponding to the factor, and determine the total weight through summation, weighted summation, etc.; and then calculate the total weight.
  • the preset number of factors ranked first are determined as the target factors, or the factors whose total weight is greater than the preset value are determined as the target factors.
  • Step 205 Based on the target factor, determine and push the recommendation result corresponding to the first user in the information flow recommendation scenario to the first user.
  • the above execution subject may determine and push the recommendation result corresponding to the first user in the information flow recommendation scenario to the first user based on the target factor.
  • the above-mentioned execution subject can first obtain the total weight of the target factor based on the first weight and the second weight corresponding to the target factor; then, based on the corresponding target factor and the total weight, The weighted items corresponding to each target factor are obtained to combine the weighted items of each target factor to obtain the multi-target factor fusion formula.
  • the recommendation ranking score of the content to be sorted in the preset content collection to be recommended can be determined through the multi-objective factor fusion formula; the content to be sorted in the content collection to be sorted is sorted based on the recommendation ranking score, so as to The preset number of top-ordered contents to be sorted are pushed to the first user as the recommendation results corresponding to the first user.
  • Figure 3 is a schematic diagram 300 of an application scenario of the information flow recommendation method according to this embodiment.
  • the user 301 issues a startup instruction to a short video application through the terminal device 302.
  • the server 303 first obtains the feature information of the user 301 in the short video information stream recommendation scenario based on the opening instruction; then, through the multi-factor fusion parameter network 304, determines the first weight 305 corresponding to each factor in the factor set based on the feature information, where , the factors in the factor set represent the index information that needs to be considered in the information flow recommendation process; then, through the gated filtering network 306, the second weight 307 corresponding to each factor in the factor set is determined according to the feature information; then, according to the first The weight 305 and the second weight 307 are used to determine the target factor 308 in the factor set that is suitable for the first user in the information flow recommendation scenario; based on the target factor 308, determine and push to the user 301 the target factor 308 corresponding to the first user in the
  • an information flow recommendation method is provided.
  • the first weight of each factor corresponding to the user is determined through a multi-factor fusion parameter network, and the first weight of each factor corresponding to the user is determined through the gated filter.
  • the selection network determines the second weight of each factor, so as to accurately determine the target factors suitable for the user based on the first weight and the second weight, and perform information flow recommendation, which improves the accuracy of the recommendation results.
  • the above execution subject may perform the above step 203 in the following manner:
  • the initial second weight corresponding to each factor in the factor set is determined based on the feature information; then, the initial second weight corresponding to each factor is converted into 0 or 1 through the preset activation function to obtain each factor. The corresponding second weight.
  • the value range of the preset activation function at the top layer of the gated screening network is 0 and 1, thus cleverly transforming the continuous value problem into a 0/1 problem.
  • its preset activation function can be:
  • the second weight output by the gated screening network is 0 or 1, thus cleverly converting the continuous value problem into a 0/1 problem and improving the efficiency of the target factor determination process. Efficiency and convenience.
  • the above execution subject may perform the above step 204 in the following manner:
  • the weight product corresponding to each factor can be easily determined to be the first weight corresponding to the factor or zero; then, factors whose weight product is zero are removed and the weight product is retained Non-zero factors are used to obtain the target factors.
  • the target factor is determined based on the weight product of the first weight and the second weight corresponding to each factor, which further improves the convenience and accuracy of the target factor determination process.
  • the above execution subject may also perform the following operations: first, obtain the first user's feedback information on the recommendation results.
  • the feedback information may be the first user's reflection information on the information flow in the recommended results after obtaining the recommended results.
  • feedback information includes whether to click, whether to view, whether to like, comment and other interactive operations.
  • an evolutionary strategy is adopted to adjust the parameters of the multi-factor fusion parameter network based on feedback information. and the parameters of the gated screening network to perform subsequent user recommendation tasks in the information flow recommendation scenario through the adjusted multi-factor fusion parameter network and gated screening network.
  • Evolutionary strategy algorithms refer to algorithms based on evolutionary theory that can be used to explore parameter perturbations that make the overall return of multi-factor fusion parameter networks and gated screening networks greater. Specifically, based on the feedback information and the preset reward function, the parameters of the multi-factor fusion parameter network and the reward values of the parameters of the gated screening network are determined; based on the principle of maximizing the reward value, the parameters and gating of the multi-factor fusion parameter network are guided. The adjustment process of filtering network parameters; based on the preset evolutionary strategy algorithm, iterates according to the preset number of iterations to generate a new round of parameters and gating of the multi-factor fusion parameter network that satisfies the Gaussian distribution of the mean and variance of each parameter. Filter network parameters.
  • the multi-factor fusion parameter network and gated screening network after adjusting parameters will be used to perform subsequent user recommendation tasks.
  • the above-mentioned executive body uses an evolutionary strategy to adjust the multi-factor fusion parameter network and gated screening network during the application process, which can continuously improve the multi-factor fusion parameter network and gated screening network. Accuracy of gated screening networks.
  • FIG. 4 a schematic process 400 of yet another embodiment of the information flow recommendation method according to the present disclosure is shown, including the following steps:
  • Step 401 Obtain the characteristic information of the first user in the information flow recommendation scenario.
  • Step 402 Determine the first weight corresponding to each factor in the factor set according to the feature information through the multi-factor fusion parameter network.
  • the factors in the factor set represent the indicator information that needs to be considered in the information flow recommendation process.
  • Step 403 Determine the initial second weight corresponding to each factor in the factor set according to the characteristic information through the gated screening network.
  • Step 404 Convert the initial second weight corresponding to each factor to 0 or 1 through a preset activation function to obtain the second weight corresponding to each factor.
  • Step 405 Multiply the first weight and the second weight corresponding to each factor in the factor set to obtain a weight product.
  • Step 406 Determine the target factor in the factor set that is suitable for the first user in the information flow recommendation scenario based on the weight product corresponding to each factor in the factor set.
  • Step 407 Determine and push the information flow recommendation scenario to the first user based on the target factor. The recommendation result corresponding to the first user.
  • the process 400 of the information flow recommendation method in this embodiment specifically illustrates the determination process of the second weight and the determination process of the target factor, further improving the The efficiency and convenience of the target factor determination process improves the accuracy of the recommendation results.
  • FIG. 5 a schematic process 500 of one embodiment of a model training method according to the present disclosure is shown, including the following steps:
  • Step 501 Obtain the characteristic information of the second user in the information flow recommendation scenario.
  • the execution subject of the model training method can obtain the second user's information flow recommendation scenario remotely or locally based on a wired network connection or a wireless network connection. characteristic information.
  • the second user is the user to be recommended for information flow during the training process of the initial multi-factor fusion parameter network and the initial gate screening network.
  • the training process shown in steps 501-506 can be performed.
  • Information flow recommendation scenarios can be recommendation scenarios corresponding to various types of information flows. For example, in news applications, the information flow recommendation scenario is to determine the user's news information flow; in video applications, the information flow recommendation scenario is to determine the user's video information flow.
  • the characteristic information of the second user in the information flow recommendation scenario includes the user characteristic information of the second user and the scene characteristic information of the information flow recommendation scenario.
  • user characteristic information includes user activity, age, gender, average daily product usage time, number of uses, etc.
  • scene characteristic information includes refresh status, refresh times, refresh time, etc.
  • Step 502 Determine the first weight corresponding to each factor in the factor set according to the feature information through the initial multi-factor fusion parameter network.
  • the above-mentioned execution subject can determine the first weight corresponding to each factor in the factor set according to the characteristic information through the initial multi-factor fusion parameter network.
  • the factors in the factor set represent the indicator information that needs to be considered in the information flow recommendation process.
  • the factor set includes factors such as reading time, diversity of displayed resources, number of user likes, number of shares, etc.
  • the factor set corresponding to each information flow recommendation scenario includes multiple factors for the information flow recommendation scenario.
  • the initial multi-factor fusion parameter network includes multiple tower networks, each tower network corresponding to a factor in the factor set. At the bottom of the initial multi-factor fusion parameter network, multiple tower networks share feature information, and multiple tower networks are used to output the first weight of the corresponding factor.
  • Each tower network can be a neural network, including but not limited to convolutional neural network, recurrent neural network and other network models.
  • Step 503 Determine the second weight corresponding to each factor in the factor set according to the characteristic information through the initial gating screening network.
  • the above-mentioned execution subject can 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 can be implemented based on gated recurrent neural networks.
  • the above execution subject can determine the network structure of the initial gating screening network according to the number of factors included in the factor set, so that the number of outputs of the initial gating screening network is consistent with the number of factors included in the factor set, and The factors included in the factor set correspond one-to-one with the output of the initial gated screening network.
  • the multiple outputs of the initial gated screening network are the second weights corresponding to each factor in the factor set.
  • Step 504 Based on the first weight and the second weight, determine the target factor in the factor set that is suitable for the second user in the information flow recommendation scenario.
  • the above execution subject may determine the target factor in the factor set that is suitable for the second user in the information flow recommendation scenario based on the first weight and the second weight.
  • the above-mentioned executive body can determine the first weight corresponding to the factor and the second weight corresponding to the factor, and determine the total weight through summation, weighted summation, etc.; and then calculate the total weight.
  • the preset number of factors ranked first are determined as the target factors, or the factors whose total weight is greater than the preset value are determined as the target factors.
  • Step 505 Determine the recommendation result corresponding to the second user in the information flow recommendation scenario according to the target factor.
  • the above execution subject can determine the recommendation result corresponding to the second user in the information flow recommendation scenario according to the target factor.
  • the above-mentioned execution subject can first obtain the total weight of the target factor based on the first weight and the second weight corresponding to the target factor; then, based on the corresponding target factor and the total weight, The weighted items corresponding to each target factor are obtained to combine the weighted items of each target factor to obtain the multi-target factor fusion formula.
  • the recommendation ranking score of the content to be sorted in the preset content collection to be recommended can be determined through the multi-objective factor fusion formula; the content to be sorted in the content collection to be sorted is sorted based on the recommendation ranking score, so as to The preset number of top-ordered contents to be sorted are pushed to the second user as the recommendation results corresponding to the second user.
  • Step 506 Use an evolutionary strategy to adjust the parameters of the initial multi-factor fusion parameter network and the parameters of the initial gate screening network based on the second user's feedback information on the recommendation results to obtain the trained multi-factor fusion parameter network and gate screening. network.
  • the above-mentioned execution subject can adopt an evolutionary strategy to adjust the parameters of the initial multi-factor fusion parameter network and the parameters of the initial gated screening network according to the feedback information of the second user on the recommendation results to obtain the multi-factor fusion after training.
  • Parametric networks and gated screening networks are used to adjust the parameters of the initial multi-factor fusion parameter network and the parameters of the initial gated screening network according to the feedback information of the second user on the recommendation results to obtain the multi-factor fusion after training.
  • the parameters of the initial multi-factor fusion parameter network and the reward values of the parameters of the initial gated screening network are determined; based on the principle of maximizing the reward value, the parameters of the initial multi-factor fusion parameter network are guided. and the adjustment process of the parameters of the initial gated screening network; based on the preset evolutionary strategy algorithm, iterate according to the preset number of iterations to generate a new round of multi-factor fusion parameter network that satisfies the Gaussian distribution of the mean and variance of each parameter.
  • Parameters and parameters of the gated screening network, and the adjusted multi-factor fusion parameter network and gated screening network are used as the initial multi-factor fusion parameter network and initial gated screening network for the next round of training.
  • the trained multi-factor fusion parameter network and gated screening network are obtained.
  • the preset end condition may be, for example, that the number of iterations exceeds a preset times threshold, the training time exceeds a preset time threshold, etc.
  • the screening of target factors based on the gated screening network effectively improves the evolutionary efficiency of the evolutionary strategy, and at the same time triggers the automatic screening of target factors that are beneficial to the whole from the perspective of global optimization.
  • the above execution subject may perform the above step 503 in the following manner:
  • each factor in the factor set is determined based on the characteristic information.
  • the initial second weight corresponding to the sub-factor; then, the initial second weight corresponding to each factor is converted into 0 or 1 through the preset activation function to obtain the second weight corresponding to each factor.
  • the value range of the preset activation function at the top layer of the initial gated screening network is 0 and 1, thus cleverly transforming the continuous value problem into a 0/1 problem.
  • its preset activation function can be:
  • the second weight output by the initial gate screening network is 0 or 1, thus cleverly transforming the continuous value problem into a 0/1 problem, further improving the evolution of the evolutionary strategy. efficiency.
  • the above execution subject may perform the above step 504 in the following manner:
  • the weight product corresponding to each factor can be conveniently determined to be the first weight corresponding to the factor or zero; then, factors whose weight product is zero are removed and the weights are retained Multiply the non-zero factors to get the target factor.
  • the target factor is determined based on the weight product of the first weight and the second weight corresponding to each factor, which further improves the convenience and accuracy of the target factor determination process during model training.
  • the present disclosure provides an embodiment of an information flow recommendation device.
  • the device embodiment corresponds to the method embodiment shown in Figure 2.
  • the device is specifically Can be used in various electronic devices.
  • the information flow recommendation device 600 includes: a first acquisition unit 601, configured to obtain the characteristic information of the first user in an information flow recommendation scenario; a first determination unit 602, configured to fuse parameters through multiple factors
  • the network determines the first weight corresponding to each factor in the factor set according to the characteristic information, where the factors in the factor set represent the index information that needs to be considered in the information flow recommendation process;
  • the second determination unit 603 is configured to pass the gating Screen the network and determine the second weight corresponding to each factor in the factor set based on the feature information;
  • the third determination unit 604 is
  • the recommendation unit 605 is configured to determine, based on the first weight and the second weight, the target factor in the factor set that is suitable for the first user in the information flow recommendation scenario; the recommendation unit 605 is configured to determine and push the information flow to the first user based on the target factor. Recommendation results corresponding to the first user in the recommendation scenario.
  • the second determination unit 603 is further configured to: determine the initial second weight corresponding to each factor in the factor set according to the characteristic information through the gated screening network; The activation function converts the initial second weight corresponding to each factor into 0 or 1 to obtain the second weight corresponding to each factor.
  • the third determination unit 604 is further configured to: multiply the first weight and the second weight corresponding to each factor in the factor set to obtain a weight product; according to the factor set The weight product corresponding to each factor in determines the target factor in the factor set that is suitable for the first user in the information flow recommendation scenario.
  • the above device further includes: a feedback unit (not shown in the figure) configured to obtain the first user's feedback information on the recommendation results; an evolution unit (not shown in the figure) ), is configured to adopt an evolutionary strategy to adjust the parameters of the multi-factor fusion parameter network and the parameters of the gated screening network based on feedback information, so as to perform subsequent user recommendations in the information flow through the adjusted multi-factor fusion parameter network and gated screening network. Recommended tasks in scenarios.
  • an information flow recommendation device is provided.
  • the first weight of each factor corresponding to the user is determined through a multi-factor fusion parameter network
  • the second weight of each factor is determined through a gated screening network. weight to accurately determine the target factors applicable to the user based on the first weight and the second weight, and perform information flow recommendation, thereby improving the accuracy of the recommendation results.
  • the present disclosure provides an embodiment of a model training device.
  • the device embodiment corresponds to the method embodiment shown in Figure 5.
  • the device can specifically Used in various electronic equipment.
  • the model training device 700 includes: a second acquisition unit 701, configured to acquire the characteristic information of the second user in the information flow recommendation scenario; a fourth determination unit 702, configured to use the initial multi-factor fusion parameters The network determines the first weight corresponding to each factor in the factor set according to the characteristic information, where the factors in the factor set represent the index information that needs to be considered in the information flow recommendation process; the fifth determination unit 703 is configured to pass the initial gate Control the screening network and determine the second weight corresponding to each factor in the factor set based on the feature information; the sixth determination unit 704, configured to determine the target factor in the factor set suitable for the second user in the information flow recommendation scenario based on the first weight and the second weight; the seventh determination unit 705, configured to determine the information flow recommendation scenario based on the target factor The recommendation results corresponding to the second user under; the training unit 706 is configured to adopt an evolutionary strategy and adjust the parameters of the initial multi-factor fusion parameter network and the parameters of the initial gate screening network according to the feedback information of the second user on the recommendation results,
  • the fifth determination unit 703 is further configured to: determine the initial second weight corresponding to each factor in the factor set according to the characteristic information through the initial gating screening network; Assume that the activation function converts the initial second weight corresponding to each factor into 0 or 1 to obtain the second weight corresponding to each factor.
  • the sixth determination unit 704 is further configured to: multiply the first weight and the second weight corresponding to each factor in the factor set to obtain a weight product; according to the factor set The weight product corresponding to each factor in determines the target factor in the factor set that is suitable for the second user in the information flow recommendation scenario.
  • the target factor is determined based on the weight product of the first weight and the second weight corresponding to each factor, which further improves the convenience and accuracy of the target factor determination process during model training.
  • the present disclosure also provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores information that can be executed by the at least one processor.
  • the instruction is executed by at least one processor, so that when executed by at least one processor, the information flow recommendation method and the model training method described in any of the above embodiments can be implemented.
  • the present disclosure also provides a readable storage medium that stores computer instructions.
  • the computer instructions are used to enable the computer to implement the information flow recommendation described in any of the above embodiments when executed. Methods, model training methods.
  • Embodiments of the present disclosure provide a computer program product. When executed by a processor, the computer program can implement the information flow recommendation method and model training method described in any of the above embodiments.
  • FIG. 8 illustrates a schematic of an example electronic device 800 that may be used to implement embodiments of the present disclosure.
  • Electronic devices are intended to refer to various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers.
  • Electronic devices may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices.
  • the components shown herein, their connections and relationships, and their functions are examples only and are not intended to limit implementations of the disclosure described and/or claimed herein.
  • the device 800 includes a computing unit 801 that can execute according to a computer program stored in a read-only memory (ROM) 802 or loaded from a storage unit 808 into a random access memory (RAM) 803 Various appropriate actions and treatments. In the RAM 803, various programs and data required for the operation of the device 800 can also be stored.
  • Computing unit 801, ROM 802 and RAM 803 are connected to each other via bus 804.
  • An input/output (I/O) interface 805 is also connected to bus 804.
  • the I/O interface 805 includes: an input unit 806, such as a keyboard, a mouse, etc.; an output unit 807, such as various types of displays, speakers, etc.; a storage unit 808, such as a magnetic disk, optical disk, etc. ; and 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 through computer networks such as the Internet and/or various telecommunications networks.
  • Computing unit 801 may be a variety of general and/or special purpose processing components having 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, digital signal processing processor (DSP), and any appropriate processor, controller, microcontroller, etc.
  • the computing unit 801 performs various methods and processes described above, such as the information flow recommendation method.
  • the information flow recommendation method may be implemented as a computer software program that is tangibly embodied in a machine-readable medium, such as the storage unit 808.
  • part or all of the computer program may be loaded and/or installed onto device 800 via ROM 802 and/or communication unit 809.
  • the computer program When the computer program is loaded into the RAM 803 and executed by the computing unit 801, one or more steps of the information flow recommendation method described above may be performed.
  • the computing unit 801 may be configured to perform the information flow recommendation method in any other suitable manner (eg, by means of firmware).
  • Various implementations of the systems and techniques described above may be implemented in digital electronic circuits System, integrated circuit system, field programmable gate array (FPGA), application specific integrated circuit (ASIC), application specific standard product (ASSP), system on chip (SOC), load programmable logic device (CPLD), computer hardware, Implemented in firmware, software, and/or a combination thereof.
  • FPGA field programmable gate array
  • ASIC application specific integrated circuit
  • ASSP application specific standard product
  • SOC system on chip
  • CPLD load programmable logic device
  • computer hardware Implemented in firmware, software, and/or a combination thereof.
  • These various embodiments may include implementation in one or more computer programs executable and/or interpreted on a programmable system including at least one programmable processor, the programmable processor
  • the processor which may be a special purpose or general purpose programmable processor, may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
  • An output device may be a special purpose or general purpose programmable processor, may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
  • An output device may be a special purpose or general purpose programmable processor, may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the 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 codes may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that the program codes, when executed by the processor or controller, cause the functions specified in the flowcharts and/or block diagrams/ The operation is implemented.
  • 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.
  • a machine-readable medium may be a tangible medium that may 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.
  • Machine-readable media may include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any suitable combination of the foregoing.
  • machine-readable storage media would include one or more wire-based electrical connections, laptop disks, hard drives, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the above.
  • RAM random access memory
  • ROM read only memory
  • EPROM or flash memory erasable programmable read only memory
  • CD-ROM portable compact disk read-only memory
  • magnetic storage device or any suitable combination of the above.
  • the systems and techniques described herein may be implemented on a computer having a display device (eg, a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user ); and a keyboard and pointing device (eg, a mouse or a trackball) through which a user can provide input to the computer.
  • a display device eg, a CRT (cathode ray tube) or LCD (liquid crystal display) monitor
  • a keyboard and pointing device eg, a mouse or a trackball
  • Other kinds of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback). feedback); and input from the user can be received in any form (including acoustic input, speech input, or tactile input).
  • the systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., A user's computer having a graphical user interface or web browser through which the user can interact with implementations of the systems and technologies described herein), or including such backend components, middleware components, or any combination of front-end components in a computing system.
  • the components of the system may be interconnected by any form or medium of digital data communication (eg, a communications network). Examples of communication networks include: local area network (LAN), wide area network (WAN), and the Internet.
  • Computer systems may include clients and servers.
  • Clients and servers are generally remote from each other and typically interact over a communications network.
  • the relationship of client and server is created by computer programs running on corresponding computers and having a client-server relationship with each other.
  • the server can be a cloud server, also known as cloud computing server or cloud host. It is a host product in the cloud computing service system to solve the management difficulties existing in traditional physical host and virtual private server (VPS, Virtual Private Server) services. Large, weak business scalability; it can also be a server of a distributed system, or a server combined with a blockchain.
  • an information flow recommendation method is provided.
  • the first parameter corresponding to each factor of the user is determined through a multi-factor fusion parameter network.
  • the weight determines the second weight of each factor through the gated screening network to accurately determine the target factors suitable for the user based on the first weight and the second weight, and performs information flow recommendation, which improves the accuracy of the recommendation results.

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

L'invention concerne un procédé et un appareil de recommandation de flux d'informations, un dispositif électronique, un support de stockage et un produit de programme, se rapportant au domaine technique de l'intelligence artificielle, en particulier, au domaine technique des stratégies d'évolution. La solution de mise en œuvre spécifique consiste : à acquérir des informations de caractéristiques d'un premier utilisateur dans un scénario de recommandation de flux d'informations (201) ; à déterminer, au moyen d'un réseau de paramètres de fusion multi-facteurs selon les informations de caractéristiques, des premiers poids correspondant à des facteurs dans un ensemble de facteurs (202), les facteurs dans l'ensemble de facteurs représentant des informations d'indice qui doivent être considérées dans un processus de recommandation de flux d'informations ; à déterminer, au moyen d'un réseau de criblage à porte selon les informations de caractéristique, des seconds poids correspondant aux facteurs dans l'ensemble de facteurs (203) ; à déterminer, en fonction des premiers poids et des seconds poids, un facteur cible dans l'ensemble de facteurs approprié pour le premier utilisateur dans le scénario de recommandation de flux d'informations (204) ; et selon le facteur cible, à déterminer et à pousser un résultat de recommandation correspondant au premier utilisateur dans le scénario de recommandation de flux d'informations au premier utilisateur (205). Le procédé améliore la précision du résultat de recommandation.
PCT/CN2023/080416 2022-07-20 2023-03-09 Procédé et appareil de recommandation de flux d'informations et produit programme d'ordinateur WO2024016680A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210857765.1 2022-07-20
CN202210857765.1A CN115203564A (zh) 2022-07-20 2022-07-20 信息流推荐方法、装置及计算机程序产品

Publications (1)

Publication Number Publication Date
WO2024016680A1 true WO2024016680A1 (fr) 2024-01-25

Family

ID=83582969

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/080416 WO2024016680A1 (fr) 2022-07-20 2023-03-09 Procédé et appareil de recommandation de flux d'informations et produit programme d'ordinateur

Country Status (2)

Country Link
CN (1) CN115203564A (fr)
WO (1) WO2024016680A1 (fr)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115203564A (zh) * 2022-07-20 2022-10-18 百度在线网络技术（北京）有限公司 信息流推荐方法、装置及计算机程序产品

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111918136A (zh) * 2020-07-04 2020-11-10 中信银行股份有限公司 一种兴趣的分析方法及装置、存储介质、电子设备
CN113836406A (zh) * 2021-09-10 2021-12-24 北京小米移动软件有限公司 信息流推荐方法及装置
CN114265979A (zh) * 2021-12-17 2022-04-01 北京百度网讯科技有限公司 确定融合参数的方法、信息推荐方法和模型训练方法
CN114463091A (zh) * 2022-01-29 2022-05-10 北京沃东天骏信息技术有限公司 信息推送模型训练和信息推送方法、装置、设备和介质
US20220215032A1 (en) * 2020-02-13 2022-07-07 Tencent Technology (Shenzhen) Company Limited Ai-based recommendation method and apparatus, electronic device, and storage medium
CN115203564A (zh) * 2022-07-20 2022-10-18 百度在线网络技术（北京）有限公司 信息流推荐方法、装置及计算机程序产品

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220215032A1 (en) * 2020-02-13 2022-07-07 Tencent Technology (Shenzhen) Company Limited Ai-based recommendation method and apparatus, electronic device, and storage medium
CN111918136A (zh) * 2020-07-04 2020-11-10 中信银行股份有限公司 一种兴趣的分析方法及装置、存储介质、电子设备
CN113836406A (zh) * 2021-09-10 2021-12-24 北京小米移动软件有限公司 信息流推荐方法及装置
CN114265979A (zh) * 2021-12-17 2022-04-01 北京百度网讯科技有限公司 确定融合参数的方法、信息推荐方法和模型训练方法
CN114463091A (zh) * 2022-01-29 2022-05-10 北京沃东天骏信息技术有限公司 信息推送模型训练和信息推送方法、装置、设备和介质
CN115203564A (zh) * 2022-07-20 2022-10-18 百度在线网络技术（北京）有限公司 信息流推荐方法、装置及计算机程序产品

Also Published As

Publication number Publication date
CN115203564A (zh) 2022-10-18

Similar Documents

Publication Publication Date Title
JP7331171B2 (ja) 画像認識モデルをトレーニングするための方法および装置、画像を認識するための方法および装置、電子機器、記憶媒体、並びにコンピュータプログラム
WO2020094060A1 (fr) Procédé et appareil de recommandation
US20220374776A1 (en) Method and system for federated learning, electronic device, and computer readable medium
WO2019047790A1 (fr) Procédé et système de génération de caractéristiques combinées d'échantillons d'apprentissage automatique
WO2023109059A1 (fr) Procédé de détermination de paramètre de fusion, procédé de recommandation d'informations et procédé d'apprentissage de modèle
CN113301442B (zh) 确定直播资源的方法、设备、介质及程序产品
CN110909165A (zh) 数据处理方法、装置、介质及电子设备
US20230083116A1 (en) Federated learning method and system, electronic device, and storage medium
CN114329201A (zh) 深度学习模型的训练方法、内容推荐方法和装置
WO2023206771A1 (fr) Procédé et appareil de modélisation d'environnement faisant appel à un graphe de flux de décision, et dispositif électronique
US20210157819A1 (en) Determining a collection of data visualizations
WO2024016680A1 (fr) Procédé et appareil de recommandation de flux d'informations et produit programme d'ordinateur
WO2023207028A1 (fr) Procédé et appareil de récupération d'image et produit-programme d'ordinateur
US20220391780A1 (en) Method of federated learning, electronic device, and storage medium
WO2023093015A1 (fr) Procédé et appareil de filtrage de données, dispositif et support de stockage
CN112528159B (zh) 一种特征质量评估方法、装置、电子设备及存储介质
CN114065864A (zh) 联邦学习方法、联邦学习装置、电子设备以及存储介质
WO2024040869A1 (fr) Procédé d'entraînement de modèle multitâche, procédé de recommandation d'informations, appareil et dispositif
WO2023185125A1 (fr) Procédé et appareil de traitement de données de ressource produit, dispositif électronique et support de stockage
US20220374775A1 (en) Method for multi-task scheduling, device and storage medium
CN115185606A (zh) 业务配置参数的获得方法、装置、设备及存储介质
CN113792876A (zh) 骨干网络的生成方法、装置、设备以及存储介质
CN113806541A (zh) 情感分类的方法和情感分类模型的训练方法、装置
CN113010782A (zh) 需求量获取方法、装置、电子设备以及计算机可读介质
US20240037410A1 (en) Method for model aggregation in federated learning, server, device, and storage medium

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23841758

Country of ref document: EP

Kind code of ref document: A1