CN113641894A

CN113641894A - Information recommendation method and device

Info

Publication number: CN113641894A
Application number: CN202110816739.XA
Authority: CN
Inventors: 高玉龙; 钟超; 游安
Original assignee: Beijing Sankuai Online Technology Co Ltd
Current assignee: Beijing Sankuai Online Technology Co Ltd
Priority date: 2021-07-20
Filing date: 2021-07-20
Publication date: 2021-11-12

Abstract

The specification discloses an information recommendation method and device, which can acquire user characteristics of a user, determine candidate recommendation information for the user under different scene types, extract information characteristics required by the scene type corresponding to the candidate recommendation information for each candidate recommendation information, and determine a feature vector from a preset feature aggregation matrix, wherein the feature aggregation matrix is used for representing each information characteristic required by each scene type and each feature vector under the combination of various user characteristics. The service platform can input the feature vector into a sub-network of a corresponding scene type in a preset sorting model to obtain a recommendation score corresponding to the candidate recommendation information, so that the candidate recommendation information is sorted according to the recommendation score corresponding to the candidate recommendation information to obtain a sorting result, and information is recommended to a user according to the sorting result, thereby improving the information recommendation efficiency and reducing the operation cost of the service platform.

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

With the continuous development of computer technology, a service platform provides recommended services for users, and the service platform can recommend services related to search terms to the users according to the search terms input by the users, or actively push the services which are possibly interested to the users based on the interests and hobbies of the users.

In order to further improve the service experience of the user, the service platform generally needs a mixed recommendation mode to recommend the service to the user. The mixed arrangement mode is that in a recommendation page displayed to a user, not only one type of recommendation information is displayed, but also multiple types of recommendation information are displayed in the same recommendation page in a mixed mode.

For example, a user enters a search term in a search bar: after the commodity A, the business platform can display the link of the commodity A, the link of a merchant selling the commodity A, the comment link related to the commodity A and the activity link related to the promotion activity of the commodity A to the user in one recommendation page, so that more business choices are provided for the user.

In practical application, the service platform generally needs to determine different types of recommendation information in a recommendation page through recommendation models of different scenes, for example, for the above example, the link of the product a is recommended through the recommendation model for recommending the product, the link of the merchant selling the product a is recommended through the recommendation model for recommending the merchant, and so on.

However, for a recommendation model of any scene, based on actual service needs, a recommendation policy of the scene may be adjusted, which is often reflected in a change in features required by the recommendation model of the scene, which results in a need to modify an original recommendation model, thereby greatly increasing the operation cost of a service platform.

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 user characteristics of a user, and determining candidate recommendation information aiming at the user, wherein the candidate recommendation information comprises recommendation information of different scene types;

for each candidate recommendation information, extracting information characteristics required by the scene type corresponding to the candidate recommendation information from the candidate recommendation information;

determining the user characteristics and characteristic vectors corresponding to the information characteristics from a preset characteristic collection matrix, wherein the characteristic collection matrix is used for representing each information characteristic required by each scene type and each characteristic vector under each user characteristic combination;

inputting the feature vector into a sub-network of a scene type corresponding to the candidate recommendation information contained in a preset sequencing model to obtain a recommendation score corresponding to the candidate recommendation information;

and sorting the candidate recommendation information according to the recommendation scores corresponding to the candidate recommendation information to obtain a sorting result, and recommending information to the user according to the sorting result.

Optionally, determining the user feature and a feature vector corresponding to the information feature from a preset feature collection matrix specifically includes:

determining common features required under each scene type from the user features and the information features, and performing feature compensation on the user features and the information features according to a feature compensation rule of the scene type corresponding to the candidate recommendation information to obtain compensation features;

coding the common features and the cross features to obtain feature codes;

and determining the user characteristics and the characteristic vectors corresponding to the information characteristics from a preset characteristic collection matrix according to the characteristic codes.

Optionally, for each scene type, the sub-network corresponding to the scene type in the ranking model includes a feature extraction layer corresponding to the scene type;

for each candidate recommendation information, extracting information features required by the scene type corresponding to the candidate recommendation information from the candidate recommendation information, specifically including:

and inputting the candidate recommendation information and the user characteristics into a characteristic extraction layer contained in a sub-network corresponding to the candidate recommendation information in the ranking model to obtain information characteristics required by the scene type corresponding to the candidate recommendation information.

Optionally, the method further comprises:

determining a feature extraction layer corresponding to the scene type to be newly added;

and constructing a sub-network corresponding to the scene type to be newly added in the sequencing model according to the feature extraction layer and a preset basic sub-network.

Optionally, constructing a sub-network corresponding to the scene type to be newly added in the ranking model according to the feature extraction layer and a preset basic sub-network, and specifically includes:

initializing network parameters of the basic sub-networks according to the network parameters contained in the sub-networks corresponding to the deployed scene types in the sequencing model to obtain initialized basic sub-networks;

and constructing a sub-network corresponding to the scene type to be newly added in the sequencing model according to the initialized basic sub-network and the feature extraction layer.

Optionally, the method further comprises:

after the sub-network corresponding to the scene type to be newly added is deployed, acquiring feedback information of a user for recommendation information corresponding to the scene type to be newly added, wherein the feedback information is used for reflecting the actual browsing condition of the user on the acquired recommendation information;

and training the sequencing model comprising the sub-network corresponding to the scene type to be newly added according to the feedback information.

Optionally, the method further comprises:

determining new features, wherein the new features comprise new user features and new information features under various scene types;

and combining the newly added features with the original features related to the feature collection matrix to update the feature collection matrix, wherein the updated feature collection matrix comprises feature vectors under the combination of the newly added features and the original features, and the original features comprise user features and information features related to the feature collection matrix.

This specification provides an apparatus for information recommendation, including:

the system comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring user characteristics of a user and determining candidate recommendation information aiming at the user, and the candidate recommendation information comprises recommendation information of different scene types;

the extraction module is used for extracting information characteristics required by the scene type corresponding to the candidate recommendation information from the candidate recommendation information aiming at each candidate recommendation information;

the determining module is used for determining the user characteristics and the characteristic vectors corresponding to the information characteristics from a preset characteristic collection matrix, wherein the characteristic collection matrix is used for representing each information characteristic required by each scene type and each characteristic vector under each user characteristic combination;

the scoring module is used for inputting the feature vectors into a sub-network of a scene type corresponding to the candidate recommendation information contained in a preset sequencing model to obtain recommendation scores corresponding to the candidate recommendation information;

and the recommending module is used for sequencing the candidate recommending information according to the recommending scores corresponding to the candidate recommending information to obtain a sequencing result, and recommending information to the user according to the sequencing result.

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 above information recommendation method when executing the program.

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

in the information recommendation method provided in this specification, a user characteristic of a user may be obtained, candidate recommendation information for the user is determined, and each candidate recommendation information includes recommendation information of different scene types, and then, for each candidate recommendation information, an information characteristic required by a scene type corresponding to the candidate recommendation information is extracted from the candidate recommendation information, and the user characteristic and a feature vector corresponding to the information characteristic are determined from a preset characteristic aggregation matrix, where the characteristic aggregation matrix is used to represent each information characteristic required by each scene type and each feature vector under a combination of various user characteristics. The service platform can input the feature vector into a sub-network of a scene type corresponding to the candidate recommendation information contained in a preset ranking model to obtain recommendation scores corresponding to the candidate recommendation information, so that the candidate recommendation information is ranked according to the recommendation scores corresponding to the candidate recommendation information to obtain a ranking result, and information recommendation is performed on the user according to the ranking result.

It can be seen from the above method that the method can determine the feature vectors corresponding to the user features and the information features, that is, the feature vectors corresponding to the candidate recommendation information, through a feature aggregation matrix including the feature vectors under any feature required by each scene type, so that even if the ranking model in the method needs to rank various recommendation information, different scene types may have requirements for updating features, the method only needs to update the feature aggregation matrix, and does not need to modify the original ranking model, thereby improving the information recommendation efficiency and reducing the operation cost of the service platform.

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 structural diagram of a ranking model provided herein;

FIG. 3 is a schematic diagram of an information recommendation apparatus provided herein;

fig. 4 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 a method for information recommendation in this specification, including the following steps:

s101: the method comprises the steps of obtaining user characteristics of a user and determining candidate recommendation information aiming at the user, wherein the candidate recommendation information comprises recommendation information of different scene types.

In practical applications, the service platform may recommend recommendation information in each scene type to the user in one recommendation page, that is, when the service platform recommends information to the user, the user may view recommendation information in a plurality of scene types, for example, may view merchant, commodity comment, merchant comment, theme content (the theme content mentioned here may refer to a content including a series of recommendation information in different scene types related to a set theme, and the set theme may be preset), and the like.

Based on this, the service platform may obtain the user characteristics of the user, and determine each candidate recommendation information for the user, where each candidate recommendation information includes recommendation information of different scene types, where the user characteristics may be various, such as: age, gender, historically viewed merchants/goods, level of consumption, etc. of the user. For example, if the user views the required recommendation information by inputting a search word, the service platform may determine, according to the search word input by the user, related search information as the candidate recommendation information, and for example, the service platform may determine, according to historical preference of the user for various recommendation information provided in the service platform, the candidate recommendation information for the user.

S102: and aiming at each candidate recommendation information, extracting information characteristics required by the scene type corresponding to the candidate recommendation information from the candidate recommendation information.

S103: and determining the user characteristics and the characteristic vectors corresponding to the information characteristics from a preset characteristic collection matrix, wherein the characteristic collection matrix is used for representing each information characteristic required by each scene type and each characteristic vector under each user characteristic combination.

After determining each candidate recommendation information, the service platform may extract, for each candidate recommendation information, information features required by a scene type corresponding to the candidate recommendation information from the candidate recommendation information, and determine, from a preset feature aggregation matrix, user features and feature vectors corresponding to the information features, where the feature aggregation matrix is used to represent each information feature required by each scene type and each feature vector under a combination of each user feature. That is to say, the feature aggregation matrix includes feature values of information features of recommendation information in each scene type, feature values of various user features, and feature values corresponding to user features and information features after combination, and the information features required by the scene type corresponding to the candidate recommendation information determined by the service platform may refer to various features related to the scene type corresponding to the candidate recommendation information, for example, for a commodity, a commodity type, a commodity price, a commodity sales volume, and related information features of a merchant corresponding to the commodity may all be information features corresponding to the commodity, and for a merchant, for example, a merchant type, a merchant evaluation, and a per-capita consumption level corresponding to the merchant may be information features corresponding to the merchant.

Since the feature aggregation matrix includes feature vectors (including feature values of various features) under various features (information features, user features, and the like), the feature vectors corresponding to the user features and the information features can be determined through the feature aggregation matrix, that is, the feature vectors may refer to feature vectors corresponding to the candidate recommendation information, and the feature aggregation matrix is a feature matrix that is constructed by unifying multiple different scene types.

Therefore, the feature vectors corresponding to the candidate recommendation information are determined through the feature aggregation matrix, then the candidate recommendation information is ranked through the feature vectors of the candidate recommendation information, even if the candidate recommendation information corresponds to different scene types, the ranking result of the candidate recommendation information is reasonable because the feature vectors are derived from the unified feature aggregation matrix, the feature aggregation matrix can be constructed in a hashmap embedding mode, the feature vectors in the feature aggregation matrix can be dynamically increased in the mode, therefore, when some features are added, the feature vectors under the features can be directly added into the feature aggregation matrix, and the specific mode of the added features can be mentioned later.

When the service platform specifically determines the user characteristics and the characteristic vectors corresponding to the information characteristics, common characteristics required under each scene type can be determined from the user characteristics and the information characteristics, the user characteristics and the information characteristics are compensated according to the characteristic compensation rule of the scene type corresponding to the candidate recommendation information to obtain compensation characteristics, the common characteristics and the compensation characteristics are encoded to obtain characteristic codes, and the characteristic vectors corresponding to the user characteristics and the information characteristics are determined from a preset characteristic collection matrix according to the characteristic codes.

The common features may refer to features available for candidate recommendation information of each scene type, for example, gender, age, location of the user, current time, current geographic location, and the like of the user. And the compensation feature may refer to a feature that is different for different scene types, for example, for a feature of price, the price of the takeaway product, the merchant in the preferential subject, etc. is different, and for another example, the merchant does not have a feature unique to the product such as the product attribute, but the product has the features.

Therefore, for such features, the service platform needs to compensate the user features and the information features according to the feature compensation rule of the scene type corresponding to the candidate recommendation information to obtain compensation features, for example, for features (referred to as personalized features herein) with different scene types, such as price, the information features and the user features corresponding to the personalized features can be selected from the information features and the user features, and the information features and the user features corresponding to the personalized features are compensated to obtain compensation features. The compensation referred to herein may include multiplication, addition, etc. For another example, for the features unique to different scene types (referred to as scene unique features herein), the operation of performing compensation according to the feature compensation rule may be to select the scene unique features of the scene type corresponding to the candidate recommendation information from the information features and the user features.

After the common features and the compensation features are determined, various features required by the candidate recommendation information are determined, but only required features are determined, and a specific feature vector is not determined, so that the common features and the compensation features need to be encoded to obtain a feature code, the feature code can represent the positions of the common features and the compensation features in a feature collection matrix, that is, a feature value required by the candidate recommendation information can be extracted from the feature collection matrix through the feature code, and a feature vector corresponding to the user features and the information features is obtained. The feature code may be in the form of [0,1,1, 0.. once, 1], that is, the feature code may be a vector or a matrix containing only zero and one, and by multiplying the feature code by the feature aggregation matrix, feature values of various common features and compensation features required by the candidate recommendation information may be obtained, so that a feature vector is conveniently obtained.

S104: and inputting the characteristic vector into a sub-network of a scene type corresponding to the candidate recommendation information contained in a preset sequencing model to obtain a recommendation score corresponding to the candidate recommendation information.

After determining the feature vector corresponding to the candidate recommendation information, the service platform may input the feature vector into a sub-network of a scene type corresponding to the candidate recommendation information included in a preset ranking model, so as to obtain a recommendation score corresponding to the candidate recommendation information. That is, the ranking model may include a sub-network of each scene type, the sub-network of each scene type is used to determine the recommendation score of the candidate recommendation information in the scene type, and the result of the ranking model may be as shown in fig. 2.

Fig. 2 is a schematic structural diagram of a ranking model provided in this specification.

As can be seen from fig. 2, the ranking model includes subnetworks corresponding to the scene types, and each of the subnetworks included in each scene type includes a feature extraction layer corresponding to the scene type, where the service platform may input the candidate recommendation information and the user feature into the feature extraction layer included in the subnetwork corresponding to the candidate recommendation information in the ranking model, so as to obtain the information feature required by the scene type corresponding to the candidate recommendation information.

When a new scene type is required to be added, a subnetwork corresponding to the scene type can be directly added in the sequencing model, specifically, the service platform can determine a feature extraction layer corresponding to the scene type to be added, and construct a subnetwork corresponding to the scene type to be added in the sequencing model according to the feature extraction layer and a preset basic subnetwork, where the basic subnetwork may refer to a basic network common to the subnetworks corresponding to the scene types during initial training, and certainly, the basic subnetwork may also be a subnetwork corresponding to a scene type similar to the scene type to be added.

In order to enable the sub-networks corresponding to the scene types to be newly added to have better accuracy when insufficient training is performed, the service platform may initialize the network parameters of the basic sub-networks according to the network parameters included in the sub-networks corresponding to the deployed scene types in the ranking model to obtain the initialized basic sub-networks, and construct the sub-networks corresponding to the scene types to be newly added in the ranking model according to the initialized basic sub-networks and the feature extraction layer. For example, the initialized network parameter may refer to an average value obtained by averaging network parameters included in the subnetworks corresponding to the scene types, or may refer to a network parameter of a subnet of a scene type similar to the scene type to be newly added.

It should be noted that, the service platform may train the ranking model through various training modes, such as incremental training, streaming training, and the like, where, taking the scene type to be newly added as an example, after deploying the sub-network corresponding to the scene type to be newly added, the service platform may obtain feedback information of the user for recommendation information corresponding to the scene type to be newly added, where the feedback information is used to reflect an actual browsing condition of the user on each obtained recommendation information, and then, may train the ranking model including the sub-network corresponding to the scene type to be newly added according to the feedback information. The feedback information may be used to indicate whether the user browses (or clicks, purchases, etc.) the obtained recommendation information. The service platform may train the ranking model once after acquiring a piece of feedback information on line, or train the ranking model once after acquiring a plurality of pieces of feedback information.

S105: and sorting the candidate recommendation information according to the recommendation scores corresponding to the candidate recommendation information to obtain a sorting result, and recommending information to the user according to the sorting result.

After determining the recommendation scores corresponding to the candidate recommendation information, the service platform may rank the candidate recommendation information to obtain a ranking result, and recommend information to the user according to the ranking result. That is, candidate recommendations with high recommendation scores may be recommended to the user at an earlier time, while candidate recommendations with low recommendation scores may be recommended to the user at a later time, or may not be recommended to the user.

It should be noted that, in actual needs, some features may need to be added, for example, when a scene type needs to be added, features corresponding to the scene type need to be added, and for example, some existing scene types need to be added. Therefore, the service platform can determine new features, the new features comprise new user features and new information features under various scene types, then, the new features can be combined with the original features related in the feature collection matrix to update the feature collection matrix, the updated feature collection matrix comprises feature vectors under the combination of the new features and the original features, and the original features comprise the original user features and the information features related in the feature collection matrix.

That is to say, the service platform needs to add the newly added feature into the feature collection matrix, that is, the feature collection matrix may include various feature values of the newly added feature, and when the newly added feature is added into the feature collection matrix, the service platform may determine a corresponding feature extraction submodel in the ranking model, so that, when a recommendation score of candidate recommendation information including the newly added feature needs to be determined, the service platform may determine the newly added feature through the feature extraction submodel, and may also determine a feature value of the candidate recommendation information under the newly added feature through the feature collection matrix, thereby determining a recommendation score corresponding to the candidate recommendation information, so that, when the newly added feature needs to be added, only the feature collection matrix needs to be updated, and the ranking model does not need to be reconstructed and retrained.

It can be seen from the above method that the method can determine the feature vector corresponding to the user feature and the information feature by a feature aggregation matrix including the feature vector under any feature required by each scene type, namely, the candidate recommendation information corresponds to the feature vector, so that even if the ranking model in the method needs to rank various recommendation information, different scene types may have the requirement of updating the features, the method only needs to update the feature collection matrix without modifying the original ranking model, thereby improving the information recommendation efficiency, reducing the operation cost of the service platform, and the service platform can construct a new sub-network of the scene type, and the sub-network is trained through the training modes such as streaming training, incremental training and the like, so that the complete sequencing model does not need to be trained again.

Based on the same idea, the present specification further provides a corresponding information recommendation apparatus, as shown in fig. 3.

Fig. 3 is a schematic diagram of an information recommendation apparatus provided in this specification, including:

an obtaining module 301, configured to obtain user characteristics of a user, and determine candidate recommendation information for the user, where the candidate recommendation information includes recommendation information of different scene types;

an extracting module 302, configured to extract, for each candidate recommendation information, an information feature required by a scene type corresponding to the candidate recommendation information from the candidate recommendation information;

a determining module 303, configured to determine the user characteristics and characteristic vectors corresponding to the information characteristics from a preset characteristic aggregation matrix, where the characteristic aggregation matrix is used to represent each information characteristic required by each scene type and each characteristic vector under each combination of the user characteristics;

a scoring module 304, configured to input the feature vector into a sub-network of a scene type corresponding to the candidate recommendation information included in a preset ranking model, to obtain a recommendation score corresponding to the candidate recommendation information;

the recommending module 305 is configured to sort the candidate recommendation information according to the recommendation scores corresponding to the candidate recommendation information to obtain a sorting result, and recommend information to the user according to the sorting result.

Optionally, the determining module 303 is specifically configured to determine, from the user features and the information features, common features required in each scene type, and perform feature compensation on the user features and the information features according to a feature compensation rule of a scene type corresponding to the candidate recommendation information to obtain compensation features; coding the common characteristic and the compensation characteristic to obtain a characteristic code; and determining the user characteristics and the characteristic vectors corresponding to the information characteristics from a preset characteristic collection matrix according to the characteristic codes.

Optionally, for each scene type, the sub-network corresponding to the scene type in the ranking model includes a feature extraction layer corresponding to the scene type; the extracting module 302 is specifically configured to input the candidate recommendation information and the user characteristic into a characteristic extraction layer included in a sub-network corresponding to the candidate recommendation information in the ranking model, so as to obtain an information characteristic required by a scene type corresponding to the candidate recommendation information.

Optionally, the apparatus further comprises:

the building module 306 is configured to determine a feature extraction layer corresponding to the scene type to be newly added; and constructing a sub-network corresponding to the scene type to be newly added in the sequencing model according to the feature extraction layer and a preset basic sub-network.

Optionally, the constructing module 306 is specifically configured to initialize the network parameters of the basic sub-network according to the network parameters included in the sub-networks corresponding to the deployed scene types in the sequencing model, so as to obtain an initialized basic sub-network; and constructing a sub-network corresponding to the scene type to be newly added in the sequencing model according to the initialized basic sub-network and the feature extraction layer.

Optionally, the apparatus further comprises:

a training module 307, configured to acquire feedback information of a user for recommendation information corresponding to the to-be-newly-added scene type after deploying the sub-network corresponding to the to-be-newly-added scene type, where the feedback information is used to reflect an actual browsing condition of the user on each acquired recommendation information; and training the sequencing model comprising the sub-network corresponding to the scene type to be newly added according to the feedback information.

Optionally, the building module 306 is further configured to determine new features, where the new features include new user features and new information features in each scene type; and combining the newly added features with the original features related to the feature collection matrix to update the feature collection matrix, wherein the updated feature collection matrix comprises feature vectors under the combination of the newly added features and the original features, and the original features comprise user features and information features related to the feature collection matrix.

The present specification also provides a computer-readable storage medium storing a computer program, which can be used to execute a method of information recommendation provided in fig. 1 above.

This specification also provides a schematic block diagram of an electronic device corresponding to that of figure 1, shown in figure 4. As shown in fig. 4, 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 for other services. The processor reads a corresponding computer program from the nonvolatile memory into the memory and then runs the computer program to implement the information recommendation method described in fig. 1. 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, devices, modules or units illustrated in the above embodiments may be implemented by a computer chip or an entity, or 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, and are described separately. 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 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 the like) having computer-usable program code embodied therein.

The description has been presented with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the description. 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 non-transitory and non-transitory, removable and non-removable media, may implement 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.

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 according to claim 1, wherein determining the feature vector corresponding to the user feature and the information feature from a preset feature aggregation matrix specifically comprises:

coding the common characteristic and the compensation characteristic to obtain a characteristic code;

3. The method of claim 1, wherein for each scene type, the sub-network corresponding to the scene type in the ranking model comprises a feature extraction layer corresponding to the scene type;

4. The method of claim 3, wherein the method further comprises:

5. The method of claim 4, wherein constructing the sub-network corresponding to the scene type to be newly added in the ranking model according to the feature extraction layer and a preset basic sub-network specifically comprises:

6. The method of claim 4, wherein the method further comprises:

7. The method of claim 1, wherein the method further comprises:

8. An apparatus for information recommendation, comprising:

9. 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 7.

10. 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 7 when executing the program.