CN114996570A

CN114996570A - Information recommendation method and device

Info

Publication number: CN114996570A
Application number: CN202210533418.3A
Authority: CN
Inventors: 王伊凡; 李忆纯; 程兵
Original assignee: Beijing Sankuai Online Technology Co Ltd
Current assignee: Beijing Sankuai Online Technology Co Ltd
Priority date: 2022-05-13
Filing date: 2022-05-13
Publication date: 2022-09-02

Abstract

The specification discloses an information recommendation method and device, which can acquire current scene information and a plurality of historical behavior information corresponding to a user, input the current scene information into a pre-trained scene characteristic extraction network to obtain current scene characteristics, input each historical scene information into a scene characteristic extraction network to obtain each historical scene characteristic, further determine the weight corresponding to each historical behavior information according to the current scene characteristics and each historical scene characteristic, compensate the behavior characteristics corresponding to each historical behavior information according to the weight corresponding to each historical behavior information to obtain compensated characteristics, then recommend information according to the compensated characteristics and the information characteristics corresponding to each piece of the acquired information to be recommended, can mine similar scenes compared with the prior art, and can consider the correlation between the historical scenes corresponding to the historical behavior information and the current scene, thereby making information recommendation more accurate.

Description

Information recommendation method and device

Technical Field

The present disclosure relates to the field of computer technologies, and in particular, to a method and an apparatus for information recommendation.

Background

In practical application, the service platform may recommend various information to the user, such as: the service platform comprises self-media information of each user, comments corresponding to merchants or commodities, multimedia information and the like.

Generally, a service platform may recommend information for a user based on current context information of the user (e.g., a current geographic location of the user, a current time, etc.), and in the prior art, the service platform may filter out historical behaviors of the user under the current context information and recommend information according to the historical behaviors.

However, this approach has certain disadvantages: the scene information of the user may change to some extent, and the screened historical behavior information of the user in the current scene may be less, so that the recommendation information meeting the requirements of the user in the current scene may not be determined more accurately.

Therefore, how to accurately determine the recommendation information meeting the requirements of the user in the current scene is an urgent problem to be solved.

Disclosure of Invention

The present specification provides an information recommendation method and apparatus, which partially solve the above problems in the prior art.

The technical scheme adopted by the specification is as follows:

the present specification provides a method for information recommendation, including:

acquiring current scene information and a plurality of historical behavior information corresponding to a user;

inputting the current scene information into a pre-trained scene feature extraction network to obtain current scene features, and inputting historical scene information contained in each historical behavior information into the scene feature extraction network to obtain each historical scene feature;

determining a weight corresponding to each historical behavior information according to the current scene characteristics and the historical scene characteristics, and compensating the behavior characteristics corresponding to each historical behavior information according to the weight corresponding to each historical behavior information to obtain compensated characteristics;

and recommending information according to the compensated characteristics and the information characteristics corresponding to the acquired information to be recommended.

Optionally, determining a weight corresponding to each historical behavior information according to the current scene characteristics and the historical scene characteristics, specifically including:

inputting the current scene features and the historical scene features into an attention network to obtain attention weights of the current scene features for each historical scene feature;

and for each piece of historical behavior information, taking the attention weight of the current scene characteristic to the historical scene characteristic of the historical behavior information as the weight corresponding to the historical behavior information.

Optionally, the information recommendation is performed according to the compensated features and the information features corresponding to the obtained information to be recommended, and specifically includes:

determining a user comprehensive characteristic corresponding to the user according to the compensated characteristic and the current scene characteristic;

for each piece of information to be recommended, obtaining an information recommendation degree corresponding to the information to be recommended according to the user comprehensive characteristics and the information characteristics corresponding to the information to be recommended;

and recommending information according to the information recommendation degree corresponding to each piece of information to be recommended.

Optionally, training the scene feature extraction network specifically includes:

acquiring a training sample, wherein the marking information corresponding to the training sample is used for representing the operation result of a historical user for historical recommendation information, and the training sample comprises scene information corresponding to the historical user;

inputting the scene information into a scene feature extraction network to be trained to obtain scene features corresponding to the scene information, and inputting the historical recommendation information into an information feature extraction network to obtain information features corresponding to the historical recommendation information;

obtaining a prediction result according to the scene characteristics and the information characteristics;

and training the scene feature extraction network according to the prediction result and the labeling information.

Optionally, the obtaining of the training sample specifically includes:

acquiring a positive sample and a negative sample corresponding to the same historical user, wherein the marking information corresponding to the positive sample is used for indicating that the historical user operates on the historical recommendation information, and the marking information corresponding to the negative sample is used for indicating that the historical user does not operate on the historical recommendation information;

training the scene feature extraction network according to the prediction result and the labeling information, and specifically comprises the following steps:

and training the scene feature extraction network by taking the prediction result of the positive sample corresponding to the historical user larger than the prediction result of the negative sample corresponding to the historical user as a training target.

Optionally, training the scene feature extraction network according to the prediction result and the labeling information, specifically including:

determining a sample weight corresponding to the training sample according to the behavior type corresponding to the marking information of the training sample;

determining a target function corresponding to the training sample according to the prediction result and the labeling information;

and weighting the target function corresponding to the training sample according to the sample weight corresponding to the training sample so as to train the scene feature extraction network.

Optionally, the historical behavior information corresponds to historical behavior sequences, and the historical behavior information included in the same historical behavior sequence belongs to the same behavior type

Obtaining information recommendation degrees corresponding to the information to be recommended according to the compensated features and the information features corresponding to the information to be recommended, and recommending information according to the information recommendation degrees, wherein the information recommendation method specifically comprises the following steps:

determining comprehensive characteristics corresponding to the historical behavior sequence according to the compensated characteristics corresponding to the historical behavior information in the historical behavior sequence;

and recommending information according to the comprehensive characteristics corresponding to the historical behavior sequences and the information characteristics corresponding to the information to be recommended.

Optionally, the context information at least includes spatio-temporal attribute information corresponding to the user, and the spatio-temporal attribute information is used for representing information related to the time and the geographic position where the user is located.

This specification provides an apparatus for information recommendation, including:

the acquisition module is used for acquiring current scene information and a plurality of historical behavior information corresponding to a user;

the input module is used for inputting the current scene information into a pre-trained scene feature extraction network to obtain current scene features, and inputting historical scene information contained in each historical behavior information into the scene feature extraction network to obtain each historical scene feature;

the determining module is used for determining the weight corresponding to each historical behavior information according to the current scene characteristics and the historical scene characteristics, and compensating the behavior characteristics corresponding to each historical behavior information according to the weight corresponding to each historical behavior information to obtain compensated characteristics;

and the recommending module is used for recommending information according to the compensated characteristics and the information characteristics corresponding to the acquired information to be recommended.

The present specification provides a computer-readable storage medium storing a computer program which, when executed by a processor, implements the above-described information recommendation method.

The present specification provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the information recommendation method when executing the program.

The technical scheme adopted by the specification can achieve the following beneficial effects:

it can be seen from the above information recommendation method that the service platform may obtain current scene information and a plurality of historical behavior information corresponding to the user, input the current scene information into a pre-trained scene feature extraction network to obtain current scene features, input historical scene information included in each historical behavior information into the scene feature extraction network to obtain each historical scene feature, further determine a weight corresponding to each historical behavior information according to the current scene features and each historical scene feature, compensate behavior features corresponding to each historical behavior information according to the weight corresponding to each historical behavior information to obtain compensated features, and then perform information recommendation according to the compensated features and the information features corresponding to each acquired information to be recommended.

As can be seen from the above, in the information recommendation method provided in this specification, the historical scene characteristics corresponding to the historical scene information included in each piece of historical behavior information can be determined, determining the current scene characteristics corresponding to the current scene information, determining the weight corresponding to each historical behavior information according to the relationship between the current scene characteristics and each historical scene characteristics, furthermore, the behavior characteristics corresponding to each historical behavior information are compensated through the weight corresponding to each historical behavior information, therefore, by combining the compensated features and the information features of the information to be recommended, how to recommend the information is determined, compared with the prior art, similar scenes can be mined, and the relevance between the historical scene corresponding to the historical behavior information and the current scene can be considered, so that the information recommendation can be more accurately carried out.

Drawings

The accompanying drawings, which are included to provide a further understanding of the specification and are incorporated in and constitute a part of this specification, illustrate embodiments of the specification and together with the description serve to explain the specification and not to limit the specification in a non-limiting sense. In the drawings:

FIG. 1 is a schematic flow chart of a method for information recommendation in this specification;

fig. 2 is a schematic diagram of a model structure corresponding to a scene feature extraction network and an information feature extraction network provided in this specification;

FIG. 3 is a schematic diagram of a model structure of a recommendation model provided in this specification;

FIG. 4 is a schematic diagram of an information recommendation apparatus provided herein;

fig. 5 is a schematic diagram of an electronic device corresponding to fig. 1 provided in the present specification.

Detailed Description

In order to make the objects, technical solutions and advantages of the present disclosure more clear, the technical solutions of the present disclosure will be clearly and completely described below with reference to the specific embodiments of the present disclosure and the accompanying drawings. It is to be understood that the embodiments described are only a few embodiments of the present disclosure, and not all embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present specification without any creative effort belong to the protection scope of the present specification.

The technical solutions provided by the embodiments of the present description are described in detail below with reference to the accompanying drawings.

Fig. 1 is a schematic flow chart of an information recommendation method in this specification, which specifically includes the following steps:

s101: and acquiring current scene information and a plurality of historical behavior information corresponding to the user.

S102: inputting the current scene information into a pre-trained scene feature extraction network to obtain current scene features, and inputting historical scene information contained in each historical behavior information into the scene feature extraction network to obtain each historical scene feature.

In practical application, when a service platform recommends information to a user, it is usually necessary to refer to information related to the current time and information related to a geographical location of the user.

Based on this, the service platform can obtain current scene information and a plurality of historical behavior information corresponding to the user at a certain time (when information recommendation needs to be performed on the user), input the current scene information into a pre-trained scene feature extraction network to obtain current scene features, and input the historical scene information contained in each historical behavior information into the scene feature extraction network to obtain each historical scene feature.

The current scene information and the historical scene information mentioned above are both scene information, and the scene information at least may include spatio-temporal attribute information corresponding to the user, where the spatio-temporal attribute information is used to indicate information related to the time and the geographic location of the user, such as the current time, the current date in the week (such as monday or tuesday), the current city, the current business circle, the current latitude and longitude information, and the like (if the historical scene information is, the date and the geographic location when the historical behavior is required to be executed by the user), and of course, the scene information may also include basic information of the user, such as the gender, the age, the resident geographic location, and the like of the user, and the current scene characteristics and the historical scene characteristics mentioned above are mainly used in this specification to compensate the behavior characteristics corresponding to the historical behavior information, of course, the current scene features may also be used in subsequent information recommendations.

The above-mentioned scene feature extraction network needs to be trained in advance, and a training sample for training the scene feature extraction network can be obtained in advance through the historical behaviors of the user in each service module of the service platform, where each service module mentioned here may include a top page recommendation module, a ranking list module, a search module, and the like, that is, when the scene feature extraction model is trained, all the historical behaviors of each user of the service platform may be combined, and training is not required to be performed only through the historical behaviors of each user in the service module to which the scene feature extraction model is applied.

When the scene feature extraction network is trained, a service platform can obtain a training sample, label information corresponding to the training sample is used for representing an operation result of a historical user for historical recommendation information, the training sample comprises the scene information corresponding to the historical user, the scene information is input into the scene feature extraction network to be trained to obtain a scene feature corresponding to the scene information, the historical recommendation information is input into the information feature extraction network to obtain an information feature corresponding to the historical recommendation information, then, a prediction result is obtained according to the scene feature and the information feature, and the scene feature extraction network is trained according to the prediction result and the label information (the scene feature extraction network and the information feature extraction network can be trained jointly).

The specific training mode may be a conventional training mode, for example, the scene feature extraction network is trained with the objective of minimizing the deviation between the prediction result and the labeled information as an optimization goal.

Of course, the scene feature extraction model may also be trained in other manners, for example, a positive sample and a negative sample corresponding to the same historical user may be obtained, the label information corresponding to the positive sample is used to indicate that the historical user has performed an operation on the historical recommendation information, the label information corresponding to the negative sample is used to indicate that the historical user has not performed an operation on the historical recommendation information, and the scene feature extraction network is trained by using, as a training target, a prediction result of the positive sample corresponding to the historical user that is greater than a prediction result of the negative sample corresponding to the historical user.

That is, for the same user, the prediction result corresponding to the positive sample (e.g. the predicted click rate for the historical recommendation information, the predicted matching degree with the historical recommendation information) of the user needs to be greater than the prediction result corresponding to the negative sample.

It should be further noted that the sample weights corresponding to the training samples obtained through the historical behaviors of different behavior types may also be different, for example, the historical behaviors of the user for the historical recommendation information may include clicking, collecting, placing an order, and the like, and for each historical behavior, the sample weight corresponding to the historical behavior may be set according to actual requirements. Therefore, during training, the sample weight corresponding to the training sample can be determined according to the behavior type corresponding to the labeling information of the training sample, the objective function corresponding to the training sample can be determined according to the prediction result and the labeling information, and the objective function corresponding to the training sample can be weighted according to the sample weight corresponding to the training sample, so that the scene feature extraction network can be trained.

If the training is performed through the second training method in the above-mentioned exemplary training methods, it may be determined that a positive sample and a negative sample of the same historical user are in the same behavior type, and then an objective function of the historical user in the behavior type is determined through the positive sample and the negative sample, and according to the behavior type, a sample weight corresponding to the objective function is determined, and then the objective function is weighted, so as to perform training on the scene feature extraction network.

It should be noted that the model structures of the scene feature extraction network and the information feature extraction network may be specifically as shown in fig. 2.

Fig. 2 is a schematic diagram of a model structure corresponding to a scene feature extraction network and an information feature extraction network provided in this specification.

The scene feature extraction network and the information feature extraction network may both be composed of an embedding layer and an MLP layer, and of course, the two feature extraction networks may also be of other network structures, and are not specifically limited herein, the information on both sides is one-hot encoded and spliced, and each one-hot vector is mapped to a low-dimensional dense vector through the embedding layer. And (3) passing the related embedding through respective MLP layers (comprising a plurality of hidden layers), carrying out multiple non-linearization on the characteristic embedding, learning interactive expression among the characteristics, and obtaining high-order vectorization expression of user scene dimension and information dimension by networks on two sides respectively behind the MLP layers.

S103: and determining the weight corresponding to each historical behavior information according to the current scene characteristics and the historical scene characteristics, and compensating the behavior characteristics corresponding to each historical behavior information according to the weight corresponding to each historical behavior information to obtain the compensated characteristics.

S104: and recommending information according to the compensated characteristics and the information characteristics corresponding to the acquired information to be recommended.

After the service platform acquires the current scene characteristics and the historical scene characteristics, the weight corresponding to each historical behavior information can be determined, the behavior characteristics corresponding to each historical behavior information are compensated according to the weight corresponding to each historical behavior information to obtain compensated characteristics, and then information recommendation is performed according to the compensated characteristics and the information characteristics corresponding to the acquired information to be recommended.

When the weight corresponding to each historical behavior information is determined according to the current scene characteristics and each historical scene characteristics, there may be a variety of ways, for example, the current scene characteristics and each historical scene characteristics may be input into an attention network, an attention weight of the current scene characteristics for each historical scene characteristic is obtained, and for each historical behavior information, the attention weight of the current scene characteristics for the historical scene characteristics of the historical behavior information is used as the weight corresponding to the historical behavior information.

For another example, for each piece of historical behavior information, the similarity between the current scene characteristic and the historical scene characteristic corresponding to the historical behavior information may be used as the weight corresponding to the historical behavior information.

When information recommendation is performed, it is usually necessary to determine which information to be recommended is recommended to a user (or determine the arrangement sequence of the information to be recommended) by combining the features of the user side and the features of the recommendation information side, so that the service platform may determine the comprehensive user features corresponding to the user according to the compensated features and the current scene features, obtain the information recommendation degree corresponding to the information to be recommended according to the comprehensive user features and the information features corresponding to the information to be recommended, and perform information recommendation according to the information recommendation degree corresponding to the information to be recommended.

The information recommendation is performed for the user by combining the current scene characteristics, and considering that the user may be a new user without any historical behavior or that the user may not have any historical behavior in the current scene, the information to be recommended can be matched for the user to a certain extent by the current scene characteristics.

It should be further noted that the form of the historical behavior information may be a plurality of historical behavior sequences, that is, a plurality of pieces of historical behavior information correspond to a plurality of historical behavior sequences, the historical behavior information included in the same historical behavior sequence belongs to the same behavior type, and the service platform may determine, according to the compensated features corresponding to each piece of historical behavior information in the historical behavior sequence, the comprehensive features corresponding to the historical behavior sequence, and perform information recommendation according to the comprehensive features corresponding to each historical behavior sequence and the information features corresponding to the information to be recommended.

That is to say, the behavior types (such as the following order behavior, the collection behavior, the click behavior, and the like) corresponding to different historical behavior sequences are different, and therefore, for each historical behavior sequence, an independent comprehensive feature of each historical behavior sequence can be determined, and information recommendation is performed in combination with the information feature corresponding to the information to be recommended.

After the current scene characteristics and the historical scene characteristics are determined, the subsequent steps can be performed through an overall recommendation model (such as a recall model, a ranking model and the like), that is, the steps of determining the weight of each historical behavior information, determining compensated characteristics and the like can be performed through the recommendation model, and the attention network can be integrated inside the recommendation model.

The model structure of a particular recommendation model may be as shown in fig. 3.

Fig. 3 is a schematic diagram of a model structure of a recommendation model provided in this specification.

For example, the current scene characteristics, the historical scene characteristics, the behavior characteristics corresponding to the historical behavior information, and the information characteristics corresponding to the information to be recommended may be input into the recommendation model, so that the recommendation model determines the weights corresponding to the historical behavior information, obtains the compensated characteristics according to the weights corresponding to the historical behavior information, and determines the user comprehensive characteristics corresponding to the user according to the compensated characteristics and the current scene characteristics, so that the recommendation model performs information recommendation by combining the user comprehensive characteristics and the information characteristics corresponding to the information to be recommended.

As can be seen from the above method, in the information recommendation method provided in this specification, the historical scene features corresponding to the historical scene information included in each piece of historical behavior information can be determined, determining the current scene characteristics corresponding to the current scene information, determining the weight corresponding to each historical behavior information according to the relationship between the current scene characteristics and each historical scene characteristics, furthermore, the behavior characteristics corresponding to each historical behavior information are compensated through the weight corresponding to each historical behavior information, therefore, by combining the compensated features and the information features of the information to be recommended, how to recommend the information is determined, compared with the prior art, similar scenes can be mined, and the relevance of the historical scene of the historical behavior information and the current scene can be considered, so that the information recommendation can be more accurately carried out.

It should be noted that all actions of acquiring signals, information or data in the present application are performed under the premise of complying with the corresponding data protection regulation policy of the country of the location and obtaining the authorization given by the owner of the corresponding device.

Based on the same idea, the present specification further provides a corresponding information recommendation apparatus, as shown in fig. 4.

Fig. 4 is a schematic diagram of an information recommendation apparatus provided in this specification, which specifically includes:

an obtaining module 401, configured to obtain current scene information and a plurality of historical behavior information corresponding to a user;

an input module 402, configured to input the current scene information into a pre-trained scene feature extraction network to obtain a current scene feature, and input historical scene information included in each piece of historical behavior information into the scene feature extraction network to obtain each historical scene feature;

a determining module 403, configured to determine a weight corresponding to each historical behavior information according to the current scene feature and each historical scene feature, and compensate the behavior feature corresponding to each historical behavior information according to the weight corresponding to each historical behavior information, so as to obtain a compensated feature;

and the recommending module 404 is configured to recommend information according to the compensated features and the information features corresponding to the acquired information to be recommended.

Optionally, the determining module 403 is specifically configured to input the current scene feature and the historical scene features into an attention network, so as to obtain an attention weight of the current scene feature for each historical scene feature; and for each piece of historical behavior information, taking the attention weight of the current scene characteristic to the historical scene characteristic of the historical behavior information as the weight corresponding to the historical behavior information.

Optionally, the recommending module 404 is specifically configured to determine, according to the compensated features and the current scene features, a user comprehensive feature corresponding to the user; for each piece of information to be recommended, obtaining information recommendation degrees corresponding to the information to be recommended according to the user comprehensive characteristics and the information characteristics corresponding to the information to be recommended; and recommending information according to the information recommendation degree corresponding to each piece of information to be recommended.

Optionally, the apparatus further comprises:

a training module 405, configured to obtain a training sample, where label information corresponding to the training sample is used to indicate an operation result of a historical user for historical recommendation information, and the training sample includes scenario information corresponding to the historical user; inputting the scene information into a scene feature extraction network to be trained to obtain scene features corresponding to the scene information, and inputting the historical recommendation information into an information feature extraction network to obtain information features corresponding to the historical recommendation information; obtaining a prediction result according to the scene characteristics and the information characteristics; and training the scene feature extraction network according to the prediction result and the labeling information.

Optionally, the training module 405 is specifically configured to obtain a positive sample and a negative sample corresponding to the same historical user, where the labeling information corresponding to the positive sample is used to indicate that the historical user has performed an operation on the historical recommendation information, and the labeling information corresponding to the negative sample is used to indicate that the historical user has not performed an operation on the historical recommendation information; and training the scene feature extraction network by taking the prediction result of the positive sample corresponding to the historical user larger than the prediction result of the negative sample corresponding to the historical user as a training target.

Optionally, the training module 405 is specifically configured to determine a sample weight corresponding to the training sample according to the behavior type corresponding to the labeling information of the training sample; determining a target function corresponding to the training sample according to the prediction result and the labeling information; and weighting the target function corresponding to the training sample according to the sample weight corresponding to the training sample so as to train the scene feature extraction network.

Optionally, the historical behavior information corresponds to historical behavior sequences, and historical behavior information included in the same historical behavior sequence belongs to the same behavior type;

the recommending module 404 is specifically configured to determine, according to the compensated features corresponding to each piece of historical behavior information in the historical behavior sequence, a comprehensive feature corresponding to the historical behavior sequence; and recommending information according to the comprehensive characteristics corresponding to the historical behavior sequences and the information characteristics corresponding to the information to be recommended.

The present specification also provides a computer-readable storage medium storing a computer program, which is operable to execute the above-described information recommendation method.

This specification also provides a schematic block diagram of the electronic device shown in fig. 5. As shown in fig. 5, at the hardware level, the electronic device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include hardware required by other services. The processor reads the corresponding computer program from the nonvolatile memory into the memory and then runs the computer program to realize the service processing method. Of course, besides the software implementation, the present specification does not exclude other implementations, such as logic devices or a combination of software and hardware, and the like, that is, the execution subject of the following processing flow is not limited to each logic unit, and may be hardware or logic devices.

In the 90 s of the 20 th century, improvements in a technology could clearly distinguish between improvements in hardware (e.g., improvements in circuit structures such as diodes, transistors, switches, etc.) and improvements in software (improvements in process flow). However, as technology advances, many of today's process flow improvements have been seen as direct improvements in hardware circuit architecture. Designers almost always obtain the corresponding hardware circuit structure by programming an improved method flow into the hardware circuit. Thus, it cannot be said that an improvement in the process flow cannot be realized by hardware physical modules. For example, a Programmable Logic Device (PLD), such as a Field Programmable Gate Array (FPGA), is an integrated circuit whose Logic functions are determined by programming the Device by a user. A digital system is "integrated" on a PLD by the designer's own programming without requiring the chip manufacturer to design and fabricate application-specific integrated circuit chips. Furthermore, nowadays, instead of manually making an Integrated Circuit chip, such Programming is often implemented by "logic compiler" software, which is similar to a software compiler used in program development and writing, but the original code before compiling is also written by a specific Programming Language, which is called Hardware Description Language (HDL), and HDL is not only one but many, such as abel (advanced Boolean Expression Language), ahdl (alternate Hardware Description Language), traffic, pl (core universal Programming Language), HDCal (jhdware Description Language), lang, Lola, HDL, laspam, hardward Description Language (vhr Description Language), vhal (Hardware Description Language), and vhigh-Language, which are currently used in most common. It will also be apparent to those skilled in the art that hardware circuitry that implements the logical method flows can be readily obtained by merely slightly programming the method flows into an integrated circuit using the hardware description languages described above.

The controller may be implemented in any suitable manner, for example, the controller may take the form of, for example, a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro) processor, logic gates, switches, an Application Specific Integrated Circuit (ASIC), a programmable logic controller, and an embedded microcontroller, examples of which include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320, the memory controller may also be implemented as part of the control logic for the memory. Those skilled in the art will also appreciate that, in addition to implementing the controller as pure computer readable program code, the same functionality can be implemented by logically programming method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Such a controller may thus be considered a hardware component, and the means included therein for performing the various functions may also be considered as a structure within the hardware component. Or even means for performing the functions may be regarded as being both a software module for performing the method and a structure within a hardware component.

The systems, apparatuses, modules or units described in the above embodiments may be specifically implemented by a computer chip or an entity, or implemented by a product with certain functions. One typical implementation device is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

For convenience of description, the above devices are described as being divided into various units by function, respectively. Of course, the functions of the various elements may be implemented in the same one or more software and/or hardware implementations of the present description.

As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.

Computer-readable media, including both permanent and non-permanent, removable and non-removable media, may implement the information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

As will be appreciated by one skilled in the art, embodiments of the present description may be provided as a method, system, or computer program product. Accordingly, the description may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the description may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and so forth) having computer-usable program code embodied therein.

This description may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.

The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.

The above description is only an example of the present specification, and is not intended to limit the present specification. Various modifications and alterations to this description will become apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present specification should be included in the scope of the claims of the present specification.

Claims

1. A method for information recommendation, comprising:

2. The method of claim 1, wherein determining the weight corresponding to each historical behavior information according to the current scene characteristics and the historical scene characteristics specifically comprises:

and regarding each piece of historical behavior information, taking the attention weight of the current scene characteristic to the historical scene characteristic of the historical behavior information as the weight corresponding to the historical behavior information.

3. The method according to claim 1, wherein the information recommendation is performed according to the compensated features and the information features corresponding to the acquired information to be recommended, and specifically includes:

4. The method of claim 1, wherein training the scene feature extraction network specifically comprises:

5. The method of claim 4, wherein obtaining training samples comprises:

6. The method of claim 4, wherein training the scene feature extraction network according to the prediction result and the labeling information comprises:

determining sample weight corresponding to the training sample according to the behavior type corresponding to the marking information of the training sample;

7. The method of claim 1, wherein the historical behavior information corresponds to historical behavior sequences, and historical behavior information included in a same historical behavior sequence belongs to a same behavior type;

according to the compensated features and the information features corresponding to the information to be recommended, information recommendation is performed, and the method specifically includes:

8. The method according to any one of claims 1 to 7, wherein the scene information at least comprises spatiotemporal attribute information corresponding to the user, and the spatiotemporal attribute information is used for representing information related to the time and the geographic position where the user is located.

9. An apparatus for information recommendation, comprising:

10. A computer-readable storage medium, characterized in that the storage medium stores a computer program which, when executed by a processor, implements the method of any of the preceding claims 1 to 8.

11. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the method of any of claims 1 to 8 when executing the program.