CN113032588B

CN113032588B - Multimedia recommendation system, method, device and equipment

Info

Publication number: CN113032588B
Application number: CN201911357766.4A
Authority: CN
Inventors: 牛亚男; 冷德维
Original assignee: Beijing Dajia Internet Information Technology Co Ltd
Current assignee: Beijing Dajia Internet Information Technology Co Ltd
Priority date: 2019-12-25
Filing date: 2019-12-25
Publication date: 2024-03-08
Anticipated expiration: 2039-12-25
Also published as: CN113032588A

Abstract

The disclosure provides a multimedia recommendation system, a method, a device and equipment, and belongs to the technical field of Internet. The system comprises: recommendation means and feature service means; the recommendation device is used for acquiring user characteristics corresponding to the user information in the recommendation processing request and sending an information identifier of the multimedia information to the feature service device; the feature service device is used for reading the multimedia features corresponding to the information identifiers; and the recommending device is also used for splicing the user characteristics and the multimedia characteristics, and acquiring the matching degree of the user information and each piece of multimedia information according to the obtained spliced characteristics. Because the multimedia recommendation system comprises the recommendation device and the feature service device, the recommendation device is responsible for acquiring the information identification and the matching degree, and the feature service device is responsible for storing and reading the features, compared with the method that the feature storage, the feature reading and the matching degree acquisition are carried out by one device, the efficiency of the multimedia recommendation system for recommending the information is improved.

Description

Multimedia recommendation system, method, device and equipment

Technical Field

The disclosure relates to the technical field of internet, and in particular relates to a multimedia recommendation system, a method, a device and equipment.

Background

Today, with the development of internet technology, entertainment applications on terminals are increasing. For example, a short video application is installed on the terminal, through which a user can watch a short video. To increase the exposure to short videos, the short video application may also recommend short videos to the user that some users may be interested in. When the short video is used for recommending short videos to users, the probability of interest of the users in a plurality of short videos to be recommended can be determined by means of the multimedia recommendation system, and the short videos of interest of the users are recommended to the users according to the probability of interest of the users in each short video.

Disclosure of Invention

The embodiment of the disclosure provides a multimedia recommendation system, a method, a device and equipment, which can solve the problem of low information recommendation efficiency of the multimedia recommendation system in the related technology. The technical scheme is as follows:

in a first aspect, a multimedia recommendation system is provided, the multimedia recommendation system comprising: recommendation means and feature service means;

the recommending device is used for acquiring user characteristics corresponding to the user information after receiving a recommending processing request carrying the user information and a plurality of multimedia information, and sending an information identifier of the multimedia information to the characteristic service device;

The feature service device is used for reading the multimedia features corresponding to the information identifiers of the multimedia information and returning the read multimedia features to the recommending device;

the recommendation device is also used for splicing the user characteristics and the read multimedia characteristics to obtain splicing characteristics, and the matching degree of the user information and each multimedia information is obtained according to the splicing characteristics.

In one possible implementation, the recommending device includes M recommending units, where M is an integer greater than 1;

each recommending unit is used for acquiring the user characteristics of the corresponding user information, sending the information identification of part of the multimedia information in the plurality of multimedia information to the characteristic service device, and ensuring that no intersection exists between the part of the multimedia information corresponding to any two recommending units;

each recommending unit is further used for splicing the user characteristics with the multimedia characteristics of the corresponding multimedia information returned by the characteristic service device to obtain spliced characteristics; determining the matching degree of the user information and the corresponding multimedia information according to the splicing characteristics;

the recommending device is also used for combining the user information obtained by each recommending unit with the matching degree of the corresponding multimedia information and then returning the user information to the sender of the recommending processing request.

In another possible implementation manner, the feature service apparatus includes N feature service units, where N is an integer greater than 1;

each feature service unit is used for storing part of the multimedia features, and no intersection exists between the part of the multimedia features stored by any two feature service units;

the feature service unit is further configured to read a multimedia feature corresponding to the information identifier of the received multimedia information from the stored part of the multimedia features, and return the read multimedia feature to the corresponding recommendation unit.

In another possible implementation manner, the recommending device further includes a request processing unit;

the request processing unit is configured to receive the recommendation processing request, allocate the user information and the plurality of multimedia information in the recommendation processing request to the M recommendation units, and no intersection exists between the multimedia information allocated by any two recommendation units.

In another possible implementation manner, each recommending unit is further configured to send the matching degree between the user information and each multimedia information in the part of multimedia information to the request processing unit;

The request processing unit is further configured to sort the plurality of multimedia information according to the matching degree of the user information and each multimedia information in the plurality of media body information, and output the sorted plurality of multimedia information.

In another possible implementation manner, the multimedia recommendation system further includes: an off-line feature training device;

the offline feature training device is used for training multimedia features and synchronizing the multimedia features to the N feature service units.

In a possible implementation manner, the offline feature training device is further configured to determine, when adding a multimedia feature, a feature service unit in which the multimedia feature is stored, and store the multimedia feature to a corresponding feature service unit.

In a second aspect, there is provided a multimedia recommendation method, the method being applied in a recommendation device, the method comprising:

receiving a recommendation processing request carrying user information and a plurality of multimedia information, acquiring user characteristics corresponding to the user information, and sending an information identifier of the multimedia information to a characteristic service device;

and splicing the user characteristics and the multimedia characteristics read by the characteristic service device to obtain splicing characteristics, and acquiring the matching degree of the user information and each multimedia information according to the splicing characteristics.

the information identification for transmitting the multimedia information to the feature service device comprises the following steps:

distributing the multimedia information to the M recommendation units, wherein no intersection exists between the multimedia information distributed by any two recommendation units, and the M recommendation units are used for respectively sending the information identification of the multimedia information to the feature service device;

the step of splicing the user features and the multimedia features read by the feature service device to obtain splicing features, and obtaining the matching degree of the user information and each multimedia information according to the splicing features comprises the following steps:

the M recommending units are used for respectively splicing the user characteristics and the multimedia characteristics of the corresponding multimedia information returned by the characteristic service device to obtain spliced characteristics; and acquiring the matching degree of the user information and the corresponding multimedia information according to the splicing characteristics.

In another possible implementation manner, the obtaining, according to the splicing feature, a matching degree between the user information and each multimedia information includes:

And inputting the spliced characteristics into a probability prediction model to obtain the matching degree of the user information and each piece of multimedia information.

In another possible implementation manner, the assigning the plurality of multimedia information to the M recommendation units includes:

the multimedia information is evenly distributed to the M recommending units; or distributing the plurality of multimedia information to the M recommendation units according to the load balancing principle according to the load of each recommendation unit.

In another possible implementation manner, after the obtaining the matching degree between the user information and each multimedia information, the method further includes:

and sequencing the plurality of multimedia information according to the matching degree of the user information and each multimedia information, and outputting the sequenced plurality of multimedia information.

In a third aspect, there is provided another multimedia recommendation method, the method being applied in a feature service apparatus, the method comprising:

receiving an information identifier of each piece of multimedia information sent by a recommending device;

reading the multimedia features corresponding to the information identifiers of the multimedia information according to the information identifiers of the multimedia information;

And returning the read multimedia features to the recommending device.

In one possible implementation manner, the feature service apparatus includes N feature service units, where N is an integer greater than 1; each feature service unit stores part of the multimedia features, and no intersection exists between any two feature service units;

the reading the multimedia features corresponding to the information identifiers of the multimedia information according to the information identifiers of each multimedia information comprises the following steps:

reading multimedia features corresponding to the information identification of the multimedia information from the stored partial multimedia features through the N feature service units respectively;

the returning the read multimedia features to the recommending device comprises the following steps:

and respectively returning the read multimedia features to the recommending device through the N feature service units.

In a fourth aspect, there is provided another multimedia recommendation method applied in an online feature training apparatus, the method comprising:

training multimedia features, and synchronizing the multimedia features into a feature service device; the feature service device comprises N feature service units, wherein N is an integer greater than 1; each feature service unit stores part of the multimedia features, and no intersection exists between any two feature service units;

When adding the multimedia features, determining a feature service unit in which the multimedia features are stored;

and storing the multimedia features to the determined feature service unit.

In a possible implementation manner, when adding a multimedia feature, determining a feature service unit in which the multimedia feature is stored includes:

when the multimedia features are added, carrying out hash operation on the multimedia features to obtain hash values of the multimedia features; and mapping the N according to the hash value to obtain the feature service unit in which the multimedia features are stored.

In a fifth aspect, there is provided a recommendation device, the device comprising:

a recommendation processing request receiving module configured to execute a recommendation processing request carrying user information and a plurality of multimedia information;

the user characteristic acquisition module is configured to acquire user characteristics corresponding to the user information;

an information identification transmission module configured to perform information identification for transmitting the multimedia information to the feature service apparatus;

and the matching degree acquisition module is configured to splice the user characteristics and the multimedia characteristics read by the characteristic service device to obtain splicing characteristics, and acquire the matching degree of the user information and each multimedia information according to the splicing characteristics.

the information identification sending module is further configured to perform allocation of the plurality of multimedia information to the M recommendation units, and no intersection exists between the multimedia information allocated by any two recommendation units, and send information identifications of the multimedia information to the feature service device through the M recommendation units respectively;

the matching degree obtaining module is configured to perform splicing of the user features and the multimedia features of the corresponding multimedia information returned by the feature service device through the M recommending units respectively to obtain spliced features; and acquiring the matching degree of the user information and each piece of multimedia information according to the splicing characteristics.

In another possible implementation manner, the matching degree obtaining module is configured to input the spliced feature into a probability prediction model to obtain the matching degree of the user information and each piece of multimedia information.

In another possible implementation manner, the information identifier sending module is configured to perform an average allocation of the plurality of multimedia information to the M recommendation units; or distributing the plurality of multimedia information to the M recommendation units according to the load balancing principle according to the load of each recommendation unit.

In another possible implementation, the apparatus further includes:

and the multimedia information output module is configured to execute the steps of sequencing the plurality of multimedia information according to the matching degree of the user information and each multimedia information and outputting the sequenced plurality of multimedia information.

In a sixth aspect, there is provided a feature service apparatus, the apparatus comprising:

an information identification receiving module configured to perform information identification of each multimedia information transmitted by the recommendation device;

a multimedia feature reading module configured to perform reading of multimedia features corresponding to the information identifications of the respective multimedia information according to the information identifications of the respective multimedia information;

and the multimedia feature sending module is configured to return the read multimedia features to the recommending device.

the multimedia feature reading module is configured to execute reading multimedia features corresponding to the information identification of the multimedia information from the stored partial multimedia features through the N feature service units respectively;

The multimedia feature sending module is configured to execute the return of the read multimedia features to the recommending device through the N feature service units.

In a seventh aspect, there is provided an offline feature training apparatus, the apparatus comprising:

a feature training module configured to perform training of multimedia features, synchronizing the multimedia features into a feature service device; the feature service device comprises N feature service units, wherein N is an integer greater than 1; each feature service unit stores part of the multimedia features, and no intersection exists between any two feature service units;

a feature service unit determination module configured to perform a feature service unit that determines that a multimedia feature is stored when the multimedia feature is added;

and the feature storage module is configured to store the multimedia features to the determined feature service unit.

In a possible implementation manner, the feature service unit determining module is configured to perform hash operation on the multimedia feature when the multimedia feature is added, so as to obtain a hash value of the multimedia feature; and mapping the N according to the hash value to obtain the feature service unit in which the multimedia features are stored.

In an eighth aspect, there is provided a multimedia recommendation device, the device comprising: one or more processors; volatile or non-volatile memory for storing the one or more processor-executable instructions; wherein the one or more processors are configured to perform the multimedia recommendation method of any of the above.

A ninth aspect provides a non-transitory computer readable storage medium, wherein instructions are stored on the computer readable storage medium, the instructions when executed by a processor of a server implementing the multimedia recommendation method of any one of the above.

The technical scheme provided by the embodiment of the disclosure has the beneficial effects that:

in an embodiment of the present disclosure, a multimedia recommendation system includes: recommendation means and feature service means; the recommending device is used for acquiring user characteristics corresponding to the user information after receiving a recommending processing request carrying the user information and a plurality of multimedia information, and sending an information identifier of the multimedia information to the characteristic service device; the feature service device is used for reading the multimedia features corresponding to the information identifiers of the multimedia information and returning the read multimedia features to the recommending device; the recommendation device is also used for splicing the user characteristics and the read multimedia characteristics to obtain splicing characteristics, and the matching degree of the user information and each multimedia information is obtained according to the splicing characteristics. Because the multimedia recommendation system comprises the recommendation device and the feature service device, the recommendation device is responsible for acquiring the information identification of the multimedia information and the matching degree of the user information and each multimedia information, and the feature service device is responsible for storing the features and reading the features according to the information identification, compared with the method that the features are stored, the features are read and the matching degree is acquired through one device, the efficiency of the multimedia recommendation system for recommending the information is improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings required for the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present disclosure, and other drawings may be obtained according to these drawings without inventive effort for a person of ordinary skill in the art.

FIG. 1 is a schematic illustration of one implementation environment provided by embodiments of the present disclosure;

FIG. 2 is a schematic diagram of a multimedia recommendation system provided in an embodiment of the present disclosure;

FIG. 3 is a schematic diagram of a multimedia recommendation system provided in an embodiment of the present disclosure;

FIG. 4 is a schematic diagram of a multimedia recommendation system provided in an embodiment of the present disclosure;

FIG. 5 is a schematic diagram of a multimedia recommendation system provided in an embodiment of the present disclosure;

FIG. 6 is a flowchart of a multimedia recommendation method provided by an embodiment of the present disclosure;

FIG. 7 is a flowchart of a multimedia recommendation method provided by an embodiment of the present disclosure;

FIG. 8 is a flowchart of a multimedia recommendation method provided by an embodiment of the present disclosure;

FIG. 9 is a block diagram of a recommendation device provided by an embodiment of the present disclosure;

FIG. 10 is a block diagram of a feature service apparatus provided by an embodiment of the present disclosure;

FIG. 11 is a block diagram of an offline feature training apparatus provided by an embodiment of the present disclosure;

fig. 12 is a schematic structural diagram of a multimedia recommendation device according to an embodiment of the present disclosure.

Reference numerals denote:

10-terminal, 20-multimedia server, 30-multimedia recommendation system, 31-recommendation device, 32-feature service device, 33-offline feature training device, 311-recommendation unit, 312-request processing unit, 321-feature service unit.

Detailed Description

For the purposes of clarity, technical solutions and advantages of the present disclosure, the following further details the embodiments of the present disclosure with reference to the accompanying drawings.

Fig. 1 is a schematic diagram of an application scenario provided in an embodiment of the present disclosure. Referring to fig. 1, the application scenario includes a terminal 10, a multimedia server 20, and a multimedia recommendation system 30. Wherein the terminal 10 and the multimedia server 20 are connected through a wireless or wired network. In addition, a multimedia application for providing a multimedia service is installed on the terminal 10, and a user corresponding to the terminal 10 can implement functions such as data transmission and message interaction through the multimedia application. The multimedia server 20 is connected to the multimedia recommendation system 30. The terminal 10 may be a computer, a mobile phone, a tablet computer, or other electronic devices. The multimedia application may be a short video application or an audio playing application, etc.

When the terminal 10 logs in the multimedia server 20 through the multimedia application, the multimedia server 20 may recommend multimedia information for the terminal 10 by means of the multimedia recommendation system 30. Correspondingly, the multimedia server 20 is configured to send a recommendation processing request to the multimedia recommendation system 30, where the recommendation processing request carries multimedia information of the user. The multimedia information includes at least a plurality of multimedia information and may further include user information.

The user information may include at least one of a sex, an age, a living area, a occupation, a school, a specialty, an academic, a favorite subject, a favorite food, a favorite star, a favorite cosmeceutical owner, a favorite movie type, a favorite variety, a favorite song type, a favorite animal, a favorite tourist attraction, a favorite brand, a favorite sports item, and the like of the user.

The multimedia information may include at least one of video information and audio information. The video information may include short video. The audio information may include at least one of songs, audio novels, commentary, phase, etc., and the present disclosure is described with respect to multimedia information as short video.

The multimedia recommendation system 30 receives the recommendation processing request sent by the multimedia server 20, and obtains the matching degree of the user information and each multimedia information according to the user information and the multimedia information. In one possible implementation, the multimedia recommendation system 30 sends the matching degree of the user information and each multimedia information to the multimedia server 20. The multimedia server 20 recommends multimedia information for the user according to the matching degree of the user information and each multimedia information.

In another possible implementation, the multimedia recommendation system 30 sorts the plurality of multimedia information according to the matching degree of the user information and each multimedia information, and sends the sorted plurality of multimedia information to the multimedia server 20. The multimedia server 20 recommends multimedia information for the user according to each of the sorted multimedia information.

For example, when the multimedia application is a short video application, the multimedia server 20 is a short video application server, accordingly. When the user views a short video through the short video application on the terminal 10, the short video application server transmits a recommendation processing request to the multimedia recommendation system 30. The multimedia recommendation system 30 obtains the matching degree of the user information and each short video according to the user information and the plurality of short videos in the recommendation processing request, and the multimedia recommendation system 30 may further sort the plurality of short videos according to the matching degree of the user information and each short video, and send the sorted plurality of short videos to the short video application server. After receiving the sequenced plurality of short videos, the short video application server sequentially recommends the sequenced plurality of short videos to the terminal 10.

Fig. 2 is a schematic diagram of a multimedia recommendation system according to an embodiment of the disclosure. Referring to fig. 2, the multimedia recommendation system includes: the recommendation device is connected with the feature service device;

The recommendation device is used for acquiring user characteristics corresponding to the user information after receiving recommendation processing requests carrying the user information and a plurality of multimedia information, and sending information identifiers of the multimedia information to the characteristic service device;

In an embodiment of the present disclosure, a multimedia recommendation system includes: recommendation means and feature service means; the recommendation device is used for acquiring user characteristics corresponding to the user information after receiving recommendation processing requests carrying the user information and a plurality of multimedia information, and sending information identifiers of the multimedia information to the characteristic service device; the feature service device is used for reading the multimedia features corresponding to the information identifiers of the multimedia information and returning the read multimedia features to the recommending device; the recommendation device is also used for splicing the user characteristics and the read multimedia characteristics to obtain splicing characteristics, and the matching degree of the user information and each multimedia information is obtained according to the splicing characteristics. Because the multimedia recommendation system comprises the recommendation device and the feature service device, the recommendation device is responsible for acquiring the information identification of the multimedia information and the matching degree of the user information and each multimedia information, and the feature service device is responsible for storing the features and reading the features according to the information identification, compared with the method that the features are stored, the features are read and the matching degree is acquired through one device, the efficiency of the multimedia recommendation system for recommending the information is improved.

In one possible implementation manner, the manner in which the recommending device obtains the user characteristics corresponding to the user information may be: the recommending means abstract user characteristics of the user from the user information.

Wherein, when the user information includes at least one of gender, age, living area, occupation, school, specialty, academic, favorite discipline, favorite food, favorite star, favorite cosmeceutical owner, favorite movie type, favorite variety, favorite song type, favorite animal, favorite tourist attraction, favorite brand, favorite sports, etc., the user feature corresponding to the user information includes at least one of gender, age, area, occupation, school, professional, academic, food, star, cosmetic blogger, film and television type, synthetic art, song, animal, tourist attraction, brand, sports, etc.

The recommending means is also for transmitting the user characteristics of the user to the characteristics service means.

The recommending device is also used for acquiring the information identification of each piece of multimedia information according to the plurality of pieces of multimedia information. The implementation mode can be as follows: the recommending means abstracts an information identification of the multimedia information from each of the multimedia information. The information identifier of the multimedia information may include one or more of gender information identifier, age information identifier, region information identifier, occupation information identifier, school information identifier, professional information identifier, academic information identifier, subject information identifier, food information identifier, star information identifier, beauty blog main information identifier, film and television type information identifier, variety of nature information identifier, song information identifier, animal information identifier, tourist attraction information identifier, brand information identifier, sports item information identifier, and the like. Of course, the information identifier may also include other information identifiers, such as building information identifiers, cultural information identifiers, etc., and the disclosure does not limit the content and number of the information identifiers of the multimedia information.

For example, when the multimedia information is a short video, and the content of the short video is a video clip of a sports star participating in a certain process, the recommendation device may obtain, according to the multimedia information, an information identifier of the multimedia information in the following manner: the recommending device acquires gender information identification of the sports stars in the short video: "Man", age information identification: "45 years", region information identification: "China", professional information identification: "Table tennis coach", professional information identification: "table tennis", academic information identification: "university", star information identification: zhang Sang, variety information identification: "sports star" and the like, and "men", "45 years", "china", "table tennis coaches", "table tennis", "university", "Zhang san", "sports star" are identified as information of the short video.

The feature service device is also used for receiving the information identification of each piece of multimedia information sent by the recommending device and reading the multimedia features corresponding to the information identifications of the pieces of multimedia information.

The feature service device is also used for storing the multimedia features, namely, storing the corresponding relation between the information identification and the feature value of each multimedia feature. Correspondingly, the mode of the feature service device reading the multimedia features corresponding to the information identifiers of the multimedia information is as follows: and the feature service device reads the feature value corresponding to each information identifier from the corresponding relation between the information identifier and the feature value according to the information identifier of each multimedia information.

The feature service device is also used for storing the corresponding relation between the user features and the feature values. The feature service device is also used for receiving the user features sent by the recommending device, and reading the feature values corresponding to the user features from the corresponding relations between the user features and the feature values according to the received user features.

In the embodiment of the disclosure, since the multimedia recommendation system includes the recommendation device and the feature service device, the feature service device is responsible for storing the features and returning the features according to the information identifier, that is, the feature service device is used as an independent module for storing and returning the feature values in the multimedia recommendation system, so that the feature service device can be used by not only the multimedia recommendation system but also other multimedia recommendation systems. In addition, the recommending device does not need to store the characteristics and read the characteristic values, so that the space of the recommending device is saved, and compared with the method that the characteristics are read and the information matching degree is obtained through one module, the recommending device saves 2/3 of memory and is convenient to optimize.

In one possible implementation, the feature service device is configured to, when reading the feature value of the user feature and the feature value of the multimedia feature of each multimedia information, return the read feature value of the user feature and the feature value of the multimedia feature of each multimedia information to the recommendation device. The recommendation device is used for splicing the characteristic value of the user characteristic with the characteristic value of the multimedia characteristic of each piece of multimedia information to obtain a spliced characteristic, and determining the matching degree of the user information and each piece of multimedia information according to the spliced characteristic.

The recommending device comprises a probability prediction model, and correspondingly, the mode of determining the matching degree of the user information and each piece of multimedia information by the recommending device according to the splicing characteristics can be as follows: the recommendation device inputs the spliced characteristics into the probability prediction model to obtain probability values of matching of the user information output by the probability prediction model and each piece of multimedia information, and the corresponding probability values are used as the matching degree of the user information and the corresponding pieces of multimedia information.

The probabilistic predictive model may be a DNN (Deep Neural Networks, deep neural network) model, a CNN (Convolutional Neural Networks, convolutional neural network) model, an RNN (Recurrent Neural Networks, cyclic neural network) model, or other neural network model, etc., and the disclosure is not limited to which model is specifically used as the probabilistic predictive model.

The recommendation device is also used for outputting the matching degree of the user information and each piece of multimedia information to the multimedia server. Therefore, the multimedia server can recommend the multimedia information according to the matching degree of the user information and each piece of multimedia information. When the multimedia information is short video, the multimedia server can recommend short video for the user according to the matching degree of the user information and each multimedia information, thereby improving the browsing amount of the user and increasing the exposure of the short video.

In one possible implementation manner, the multimedia recommendation system further comprises an offline feature training device, and the offline feature training device is connected with the feature service device. And the offline feature training device is used for training the multimedia features and synchronizing the multimedia features to the feature service device.

Fig. 3 is a schematic diagram of a multimedia recommendation system according to an embodiment of the disclosure. Referring to fig. 3, the recommending apparatus includes M recommending units, M being an integer greater than 1;

each recommending unit is connected with the characteristic service device;

each recommending unit is used for acquiring the user characteristics of the corresponding user information and the information identification of part of the multimedia information in the plurality of multimedia information; and sending the information identification of part of the multimedia information in the plurality of multimedia information to the feature service device, wherein no intersection exists between the part of the multimedia information corresponding to any two recommendation units.

Each recommending unit is also used for splicing the user characteristics with the multimedia characteristics of the corresponding multimedia information returned by the characteristic service device to obtain spliced characteristics; and determining the matching degree of the user information and the corresponding multimedia information according to the splicing characteristics.

Each recommending unit is further used for sending the user characteristics of the corresponding user information to the characteristic service device.

The number of recommended units may be configured as needed, and in the embodiment of the present disclosure, the number of recommended units is illustrated as 8. For example, referring to FIG. 3, the number of recommended units is 8, recommended unit 0, recommended unit 1, recommended unit 7, respectively. The recommendation unit is an refer-ence card (estimated slicing) obtained by slicing the recommendation device.

The number of the multimedia information processed by each recommendation unit can be the same or different; also, the amount of multimedia information processed by each recommending unit may be allocated by an independent request processing unit. Correspondingly, referring to fig. 4, the recommending apparatus further includes a request processing unit; the request processing unit is connected with each recommending unit; the request processing unit is used for receiving the recommendation processing request, distributing the user information and the multimedia information in the recommendation processing request to the M recommendation units, and enabling no intersection to exist between the multimedia information distributed by any two recommendation units.

The mode that the request processing unit distributes the user information and the plurality of multimedia information in the recommendation processing request to the M recommendation units for processing is as follows:

In one possible implementation, the request processing unit equally distributes the user information and the plurality of multimedia information in the recommendation processing request to the M recommendation units for processing according to the number of the plurality of multimedia information and the number of the recommendation units.

For example, if the number of multimedia information is 16 and the number of recommendation units is 8, the request processing unit distributes the user information and the plurality of multimedia information in the recommendation processing request to the M recommendation units for processing according to the number of multimedia information and the number of recommendation units in an average manner that: the request processing unit assigns user information and 2 pieces of multimedia information to each recommending unit.

In another possible implementation manner, the request processing unit distributes the user information and the plurality of multimedia information according to the load information of each recommending unit, and further distributes the user information and the plurality of multimedia information to the M recommending units for processing according to the load of each recommending unit and the load balancing principle. Correspondingly, the request processing unit acquires the load information of each recommending unit before distributing the user information and the multimedia information, determines one recommending unit with the minimum load from each recommending unit according to the load information of each recommending unit, and distributes the user information and the multimedia information to the recommending unit for processing. Or, according to the load information of each recommending unit, the quantity of the multimedia information distributed by each recommending unit is determined, so that the quantity of the multimedia information distributed by the recommending unit with large load is small, and the quantity of the multimedia information distributed by the recommending unit with small load is large.

Wherein the amount of multimedia information may be any amount. In the embodiment of the present disclosure, taking the number of multimedia information as 8 and the number of recommendation units as 8 as an example, correspondingly, the implementation manner of each recommendation unit obtaining the user characteristics of the user information and the information identifier of the part of the multimedia information in the plurality of multimedia information and sending the user characteristics and the information identifier of the part of the multimedia information in the plurality of multimedia information to the feature service device may be:

the recommending unit 0 obtains the user characteristics according to the user information, obtains the information identification of the first multimedia information according to the first multimedia information, and sends the user characteristics and the information identification of the first multimedia information to the characteristic service device. The recommending unit 1 acquires the user characteristics from the user information, acquires the information identification of the second multimedia information from the second multimedia information, transmits the user characteristics and the information identification of the second multimedia information to the feature service apparatus, and the like.

And each recommending unit is used for splicing the characteristic value of the user characteristic returned by the characteristic service device with the characteristic value of the multimedia characteristic of the corresponding multimedia information.

Each recommending unit splices the characteristic value of the user characteristic with the characteristic value of the multimedia characteristic corresponding to the multimedia information, and the mode of obtaining the spliced characteristic corresponding to each piece of multimedia information in part of the multimedia information is as follows:

and for any one of the recommendation units and any one of the part of the multimedia information corresponding to the recommendation units, the recommendation units splice the characteristic value of each user characteristic corresponding to each user characteristic of the user with the characteristic value of each corresponding multimedia characteristic of the multimedia information to obtain spliced characteristics corresponding to the multimedia information. The present disclosure is illustrated by way of example with respect to recommendation unit 0.

For example, the number of user features is 3, respectively: sex characteristics: "Man", occupational features: "athlete", professional characteristics: the characteristic value of the user characteristic corresponding to the badminton and the male is A, the characteristic value of the user characteristic corresponding to the athlete is B, and the characteristic value of the user characteristic corresponding to the badminton is C. The recommendation unit 0 corresponds to a piece of multimedia information, the sex information of which identifies that the characteristic value corresponding to "man" is "a", the professional information identifies that the characteristic value corresponding to "athlete" is "b", and the professional information identifies that the characteristic value corresponding to "badminton" is "c". The recommendation unit 0 splices the A and the a, splices the B and the B, and splices the C and the C to obtain splicing characteristics corresponding to the multimedia information.

In a possible implementation manner, each recommending unit comprises a probability prediction model, and each recommending unit is used for inputting the splicing characteristic into the probability prediction model and outputting the matching degree of the user and each piece of multimedia information in the corresponding part of multimedia information.

Correspondingly, each recommendation unit inputs the splicing characteristics into the probability prediction model, and the mode for outputting the matching degree of the user and each piece of multimedia information in the corresponding part of multimedia information is as follows: and each recommending unit inputs the splicing characteristic corresponding to each piece of the multimedia information in the part of the multimedia information into a probability prediction model to obtain a probability value of matching the user information output by the probability prediction model with each piece of the multimedia information in the corresponding part of the multimedia information, and the corresponding probability value is used as the matching degree of the user information and each piece of the multimedia information in the corresponding part of the multimedia information.

Taking the recommending unit 0 as an example, the recommending unit 0 splices the A and the a, splices the B and the B, and splices the C and the C, so that the splicing characteristics corresponding to the obtained multimedia information are input into the probability prediction model, and the matching degree of the user information output by the probability prediction model and the multimedia information is obtained.

In the embodiment of the disclosure, the recommending device comprises a plurality of recommending units, each recommending unit is only responsible for acquiring the information identification of part of the multimedia information, so that the plurality of multimedia information is distributed to each recommending unit, and the plurality of recommending units synchronously process, thereby improving the service throughput of the recommending device and reducing the service time consumption.

In a possible implementation, each recommendation unit is further configured to send the matching degree of the user information and each multimedia information in the corresponding part of multimedia information to the request processing unit. The request processing unit is also used for receiving the matching degree of the user information sent by each recommending unit and each piece of multimedia information in the part of multimedia information and outputting the matching degree of the user information and each piece of multimedia information.

In one possible implementation, the request processing unit is further configured to sort the plurality of multimedia information according to the matching degree between the user information and each multimedia information, and output the sorted plurality of multimedia information. The manner in which the request processing unit sorts the plurality of multimedia information may be: the request processing unit arranges the multimedia information with high matching degree to the front and the multimedia information with low matching degree to the back according to the matching degree of the user information and each multimedia information. For example, if the matching degree is a probability value that the user information matches the multimedia information, the request processing unit ranks the multimedia information with a large corresponding probability value to the front and ranks the multimedia information with a small corresponding probability value to the rear.

In the embodiment of the disclosure, the request processing unit sorts the plurality of multimedia information according to the matching degree of the user information and each multimedia information by receiving the matching degree of the user information sent by each recommending unit and each multimedia information in the corresponding part, and outputs the sorted plurality of multimedia information, thereby facilitating the multimedia server to recommend the multimedia information according to the output of the multimedia recommending system.

It should be noted that in the present disclosure, the recommendation device processes the full amount of information in the recommendation processing request, that is, all of the user information and the multimedia information, and each recommendation unit processes only a part of the full amount of information, and each recommendation unit is responsible for processing the information without crossing, so that the information to be processed in the recommendation processing request is distributed to a plurality of recommendation units, and the plurality of recommendation units synchronously process, thereby greatly improving the speed of information processing and improving the multimedia recommendation efficiency of the multimedia recommendation system.

Referring to fig. 5, the feature service apparatus includes N feature service units, N being an integer greater than 1; each feature service unit is connected with each recommendation unit.

and the feature service unit is also used for reading the multimedia features corresponding to the information identifiers of the received multimedia information from the stored partial multimedia features and returning the read multimedia features to the corresponding recommendation units.

The number of feature service units and the number of recommendation units may be the same or different. In addition, the number of the recommending units and the feature service units may be configured as required, and in the embodiment of the present disclosure, the recommending units and the feature service units are the same in number and are both 8, that is, M and N are both equal to 8. For example, with continued reference to FIG. 5, the number of recommended units is 8, recommended unit 0, recommended unit 1, recommended unit 7, respectively. The number of the feature service units is 8, namely feature service unit 0, feature service unit 1 and feature service unit 7. The feature service unit is a Parameter card (Parameter slicing) obtained by slicing the feature service apparatus.

Each recommending unit is connected with each feature service unit, the recommending unit 0 is respectively connected with the feature service unit 0, the feature service unit 1, the feature service unit 7, the recommending unit 1 is respectively connected with the feature service unit 0, the feature service unit 1, the feature service unit 7, and the like.

In the embodiment of the disclosure, by storing part of the features in the full multimedia features by each feature service unit, and no repeated features exist between the features stored by each feature service unit and the features stored by other feature service units, the online feature service device can store more features, thereby improving the accuracy of multimedia recommendation and ensuring that online benefits are the same as offline training benefits. For example, the offline feature training device trains 340 billions of offline features, in the prior art, only 60 billions of offline features can be stored, and in the present disclosure, when the number of feature service units is 8, each feature service unit only needs to store 42.5 billions of features, so that the number of online features is the same as the number of offline features, and the online benefits and the offline benefits are consistent.

In one possible implementation manner, when the feature service apparatus includes a plurality of feature service units, an implementation manner in which each recommendation unit sends the user feature and the information identifier of the corresponding part of the plurality of multimedia information to the feature service apparatus may be: and each recommending unit sends the user characteristics and the information identification of the part of the multimedia information corresponding to the recommending unit to each characteristic service unit.

For example, in the embodiment of the present disclosure, the number of recommended units is 8, and the number of feature service units is 8, where the recommended unit 0 sends the obtained user features and the information identifiers of the corresponding part of the multimedia information to the feature service unit 0, the feature service unit 1, & lt & gtthe & lt/& gtthe feature service unit 7, respectively. The recommending unit 1 sends the obtained user characteristics and the information identifications of the corresponding part of the multimedia information to the characteristic service unit 0, the characteristic service unit 1, the characteristic service unit 7 and the like.

The feature service unit is used for receiving the information identification of each piece of multimedia information in the corresponding piece of multimedia information sent by each recommendation unit, reading the multimedia feature corresponding to the information identification of the received multimedia information from the stored part of multimedia features according to the received information identification of each piece of multimedia information, and returning the read multimedia feature to the corresponding recommendation unit.

In one possible implementation, each feature service unit stores a correspondence of an information identification of each of the partial multimedia features to a feature value. Correspondingly, the feature service unit reads the multimedia features corresponding to the information identification of the received multimedia information from the stored partial multimedia features according to the information identification of each received multimedia information, wherein the mode of reading the multimedia features corresponding to the information identification of the received multimedia information is as follows: and the feature service unit reads the feature value of the multimedia feature of each piece of multimedia information from the corresponding relation between the information identifier and the feature value according to the received information identifier of each piece of multimedia information.

The feature service unit is also used for storing the corresponding relation between part of the user features and the feature values. The feature service unit is further configured to receive the corresponding user feature sent by each recommendation unit, and read, according to the received user feature, a feature value corresponding to the user feature from a correspondence between the user feature and the feature value.

Correspondingly, for any information identifier corresponding to the multimedia information, if the feature service unit stores the feature value of the multimedia feature corresponding to the information identifier, returning the feature value of the multimedia feature corresponding to the information identifier to the recommendation unit for sending the information identifier; and if the feature service unit stores the feature value corresponding to the user feature, returning the feature value of the user feature to the recommendation unit for transmitting the user feature.

It should be noted that, the correspondence between the information identifier and the feature value may be a key-value pair. Wherein, the key is the information identification of the multimedia feature corresponding to the key-value pair, and the value is the feature value of the multimedia feature corresponding to the key-value pair.

In one possible implementation manner, the offline feature training device is further configured to determine, when adding a multimedia feature, a feature service unit in which the multimedia feature is stored, and store the multimedia feature to a corresponding feature service unit. Embodiments of this process may be:

When the multimedia features are added, the offline feature training device carries out hash operation on the information marks of the multimedia features to obtain hash values of the multimedia features, maps the number N of the feature service units according to the hash values to obtain feature service units with stored multimedia features, and sends the multimedia features to the corresponding feature service units.

In one possible implementation manner, the feature service unit stores the corresponding multimedia features may be: the feature service unit encodes the multimedia feature into a predefined format, thereby converting the multimedia feature into a key-value pair (key-value pair), and storing the key-value pair.

In the embodiment of the disclosure, when the multimedia features are added, the offline feature training device determines the feature service unit in which the multimedia features are stored, only the feature service unit stores the multimedia features, and each feature training module does not store repeated features, so that the feature quantity of the online feature service device is enlarged, and the online service index is improved. And the off-line feature training device only needs to transmit features to the feature service unit, but does not need to transmit features to the recommendation unit, so that the pressure of data transmission is reduced by 83.5%.

Fig. 6 is a flowchart of a multimedia recommendation method provided in an embodiment of the present disclosure. Referring to fig. 6, the multimedia recommendation method may be applied to a recommendation apparatus, and the embodiment includes:

in step 601, a recommendation processing request carrying user information and a plurality of multimedia information is received, a user feature corresponding to the user information is obtained, and an information identifier of the multimedia information is sent to a feature service device.

In step 602, the user features and the multimedia features read by the feature service device are spliced to obtain spliced features, and the matching degree of the user information and each multimedia information is obtained according to the spliced features.

In the embodiment of the disclosure, the recommendation device obtains the matching degree of the user information and the multimedia information based on the multimedia features read by the feature service device by sending the information identification of the multimedia information to the feature service device, so that the recommendation device does not need to store the multimedia features and read the features, the capacity of the recommendation device is greatly saved, and the efficiency of the recommendation device for obtaining the matching degree of the user information and each multimedia information can be improved.

Transmitting an information identification of the multimedia information to the feature service apparatus, comprising:

distributing the multimedia information to M recommendation units, wherein no intersection exists between the multimedia information distributed by any two recommendation units, and the M recommendation units are used for respectively transmitting the information identification of the multimedia information to the feature service device;

splicing the user characteristics and the multimedia characteristics read by the characteristic service device to obtain splicing characteristics, and acquiring the matching degree of the user information and each multimedia information according to the splicing characteristics, wherein the method comprises the following steps:

the method comprises the steps that the M recommending units are used for respectively splicing the user characteristics and the multimedia characteristics of corresponding multimedia information returned by the characteristic service device to obtain spliced characteristics; and acquiring the matching degree of the user information and the corresponding multimedia information according to the splicing characteristics.

In another possible implementation manner, obtaining the matching degree of the user information and each multimedia information according to the splicing characteristic includes:

In another possible implementation, the assigning the plurality of multimedia information to the M recommendation units includes:

Equally distributing the multimedia information to M recommending units; or distributing the plurality of multimedia information to the M recommending units according to the load balancing principle according to the load of each recommending unit.

In another possible implementation manner, after obtaining the matching degree between the user information and each piece of multimedia information, the method further includes:

Fig. 7 is a flowchart of a multimedia recommendation method provided in an embodiment of the present disclosure. Referring to fig. 7, the multimedia recommendation method may be applied to a feature service apparatus, the method including:

in step 701, an information identification of each multimedia information transmitted by the recommending means is received.

In step 702, according to the information identifier of each multimedia message, the multimedia features corresponding to the information identifiers of the multimedia messages are read.

In step 703, the read multimedia features are returned to the recommending means.

In the embodiment of the disclosure, the feature service device reads the corresponding multimedia features according to the information identification of the multimedia information sent by the recommending device and returns the read multimedia features to the recommending device, so that the recommending device does not need to store and read the features, the capacity of the recommending device is saved, and the information processing efficiency of the recommending device is improved. On the other hand, the feature service device reads the features according to the information identifier, and then returns the features to the sender of the information identifier, namely, independent feature reading is realized, so that the feature reading service can be provided for different recommendation devices.

reading the multimedia features corresponding to the information identifiers of the multimedia information according to the information identifiers of each multimedia information, wherein the method comprises the following steps:

reading multimedia features corresponding to the information identification of the multimedia information from the stored partial multimedia features through N feature service units respectively;

returning the read multimedia features to the recommending device, comprising:

Fig. 8 is a flowchart of a multimedia recommendation method provided in an embodiment of the present disclosure. Referring to fig. 8, the multimedia recommendation method may be applied to an online feature training apparatus, and the method includes:

training multimedia features, synchronizing the multimedia features into a feature service device, step 801; the feature service device comprises N feature service units, wherein N is an integer greater than 1; each feature service unit stores part of the multimedia features, and no intersection exists between any two feature service units storing part of the multimedia features.

In step 802, when a multimedia feature is added, a feature service unit in which the multimedia feature is stored is determined.

In step 803, the multimedia features are stored to the determined feature service unit.

In the embodiment of the disclosure, the offline feature training device synchronizes the multimedia features into the online feature service device by training the multimedia features, so that the storage amount of the online features is continuously increased. On the other hand, the off-line feature training device determines the feature service unit in which the multimedia features are stored when the multimedia features are added, and stores the multimedia features into the determined feature service unit, so that repeated features are not stored among a plurality of feature service units of the on-line feature service device, the storage capacity of the on-line features is greatly improved, and the accuracy of information recommendation when a subsequent multimedia recommendation system performs multimedia recommendation based on the on-line multimedia features is ensured.

In one possible implementation, when adding a multimedia feature, determining a feature service unit in which the multimedia feature is stored includes:

when the multimedia features are added, carrying out hash operation on the multimedia features to obtain hash values of the multimedia features; and mapping N according to the hash value to obtain a feature service unit with the stored multimedia features.

Fig. 9 is a block diagram of a recommendation device provided in an embodiment of the present disclosure. Referring to fig. 9, the apparatus includes:

the recommendation processing request receiving module 901 is configured to execute a recommendation processing request carrying user information and a plurality of multimedia information.

The user feature acquisition module 902 is configured to perform acquisition of a user feature corresponding to the user information.

The information identification transmitting module 903 is configured to perform information identification for transmitting the multimedia information to the feature service apparatus.

The matching degree obtaining module 904 is configured to perform splicing of the user feature and the multimedia feature read by the feature service device, obtain a spliced feature, and obtain the matching degree of the user information and each multimedia information according to the spliced feature.

the information identification sending module is further configured to distribute the plurality of multimedia information to the M recommending units, no intersection exists between the multimedia information distributed by any two recommending units, and the M recommending units are used for respectively sending the information identification of the multimedia information to the feature service device;

the matching degree obtaining module 904 is configured to perform splicing of the user features and the multimedia features of the corresponding multimedia information returned by the feature service device through the M recommending units respectively, so as to obtain spliced features; and acquiring the matching degree of the user information and each piece of multimedia information according to the splicing characteristics.

In another possible implementation, the matching degree obtaining module 904 is configured to perform inputting the stitching feature into the probabilistic predictive model to obtain the matching degree of the user information and each multimedia information.

In another possible implementation manner, the information identifier sending module is configured to perform an average distribution of the plurality of multimedia information to the M recommendation units; or distributing the plurality of multimedia information to the M recommending units according to the load balancing principle according to the load of each recommending unit.

In another possible implementation, the apparatus further includes:

Fig. 10 is a block diagram of a feature service apparatus provided by an embodiment of the present disclosure. Referring to fig. 10, the apparatus includes:

the information identification receiving module 1001 is configured to perform information identification of each multimedia information transmitted by the receiving recommendation device.

The multimedia feature reading module 1002 is configured to perform reading of multimedia features corresponding to the information identifications of the respective multimedia information according to the information identifications of each multimedia information.

The multimedia feature transmitting module 1003 is configured to perform returning of the read multimedia features to the recommending apparatus.

a multimedia feature reading module 1002 configured to perform reading of multimedia features corresponding to information identifications of multimedia information from the partial multimedia features stored therein by the N feature service units, respectively;

the multimedia feature transmission module 1003 is configured to perform return of the read multimedia features to the recommending apparatus through the N feature service units, respectively.

Fig. 11 is a block diagram of an offline feature training apparatus provided by an embodiment of the present disclosure. Referring to fig. 11, the apparatus includes:

a feature training module 1101 configured to perform training of multimedia features, synchronizing the multimedia features into feature serving means; the feature service device comprises N feature service units, wherein N is an integer greater than 1; each feature service unit stores part of the multimedia features, and no intersection exists between any two feature service units storing part of the multimedia features.

The feature service unit determining module 1102 is configured to perform a feature service unit that determines that the multimedia feature is stored when the multimedia feature is added.

The feature storage module 1103 is configured to perform the storing of the multimedia features to the determined feature service units.

In a possible implementation manner, the feature service unit determining module 1102 is configured to perform hash operation on the multimedia feature when the multimedia feature is added, so as to obtain a hash value of the multimedia feature; and mapping N according to the hash value to obtain a feature service unit with the stored multimedia features.

Any combination of the above-mentioned optional solutions may be adopted to form an optional embodiment of the present disclosure, which is not described herein in detail.

It should be noted that: the recommendation device, the feature service device and the offline feature training device provided in the above embodiment are only exemplified by the division of the above functional modules when performing multimedia recommendation, and in practical application, the above functional allocation may be performed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules, so as to complete all or part of the functions described above.

Fig. 12 is a schematic structural diagram of a multimedia recommendation device provided in an embodiment of the disclosure, where the multimedia recommendation device 1200 may have a relatively large difference due to different configurations or performances, and may include one or more processors (central processing units, CPU) 1201 and one or more memories 1202, where the memory 1202 stores at least one instruction, and the at least one instruction is loaded and executed by the processor 1201 to implement the methods provided in the foregoing method embodiments. Of course, the multimedia recommendation device 1200 may also have a wired or wireless network interface, a keyboard, an input/output interface, and other components for implementing the functions of the device, which are not described herein.

In an exemplary embodiment, a computer readable storage medium, such as a memory comprising instructions executable by a processor in a terminal to perform the multimedia recommendation method of the above embodiment is also provided. For example, the computer readable storage medium may be ROM, random Access Memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.

It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program for instructing relevant hardware, where the program may be stored in a computer readable storage medium, and the storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.

The foregoing description of the preferred embodiments of the present disclosure is not intended to limit the disclosure, but rather to enable any modification, equivalent replacement, improvement or the like, which fall within the spirit and principles of the present disclosure.

Claims

1. A multimedia recommendation system, the multimedia recommendation system comprising: the device comprises a recommending device, an off-line feature training device and a feature service device, wherein the recommending device is connected with the feature service device;

the feature service device comprises N feature service units, wherein N is an integer greater than 1;

the feature service unit is configured to read, according to the information identifier of each multimedia information, a feature value corresponding to each information identifier from a corresponding relationship between the information identifier of each multimedia feature and a feature value in the stored partial multimedia features, and return the read feature value of each multimedia feature to the recommendation device, where the feature value of each multimedia feature is a character-like multimedia feature, and is used to indicate specific content of the corresponding multimedia feature;

the feature service device is further configured to receive the user features sent by the recommendation device, read feature values corresponding to the user features from the stored correspondence between the user features and the feature values according to the received user features, and return the read feature values of the user features to the recommendation device, where the feature values corresponding to each user feature are character-typed user features, and are used for indicating specific content of the corresponding user feature;

The recommendation device is also used for splicing the characteristic value of the user characteristic with the read characteristic value of the multimedia characteristic to obtain a spliced characteristic, and acquiring the matching degree of the user information and each multimedia information according to the spliced characteristic;

the offline feature training device is used for training multimedia features and synchronizing the multimedia features to the N feature service units;

the offline feature training device is further configured to determine a feature service unit in which the multimedia feature is stored when the multimedia feature is added, and store the multimedia feature to the corresponding feature service unit.

2. The system of claim 1, wherein the recommending means comprises M recommending units, M being an integer greater than 1;

3. The system according to claim 2, wherein the feature service unit is further configured to return the read multimedia features to the corresponding recommendation unit.

4. The system of claim 2, wherein the recommending means further comprises a request processing unit;

5. The system of claim 4, wherein each recommending unit is further configured to send the matching degree between the user information and each multimedia information in the partial multimedia information to the request processing unit;

the request processing unit is further configured to sort the plurality of multimedia information according to the matching degree of the user information and each multimedia information in the plurality of multimedia information, and output the sorted plurality of multimedia information.

6. A multimedia recommendation method, wherein the method is applied to a recommendation device in a multimedia recommendation system, the multimedia recommendation system further comprises a feature service device and an off-line feature training device, the recommendation device is connected with the feature service device, and the method comprises:

the feature service device comprises N feature service units, wherein N is an integer greater than 1; each feature service unit is used for storing part of multimedia features, no intersection exists between the part of multimedia features stored by any two feature service units, the feature service units are used for reading a feature value corresponding to each information identifier from the corresponding relation between the information identifier of each multimedia feature in the stored part of multimedia features and the feature value according to the information identifier of each multimedia feature, and returning the read feature value of each multimedia feature to the recommending device, wherein the feature value of each multimedia feature is a character-typed multimedia feature and is used for indicating the specific content of the corresponding multimedia feature;

The feature service device is used for receiving the user features sent by the recommendation device, reading feature values corresponding to the user features from the corresponding relation between the stored user features and the feature values according to the received user features, returning the read feature values of the user features to the recommendation device, wherein the feature values corresponding to each user feature are the character user features and are used for indicating the specific content of the corresponding user feature;

splicing the characteristic value of the user characteristic with the characteristic value of the multimedia characteristic read by the characteristic service device to obtain a spliced characteristic, and acquiring the matching degree of the user information and each multimedia information according to the spliced characteristic;

the offline feature training device is further used for determining a feature service unit in which the multimedia features are stored when the multimedia features are added, and storing the multimedia features into the corresponding feature service units.

7. The method of claim 6, wherein the recommending means comprises M recommending units, M being an integer greater than 1;

the step of splicing the characteristic value of the user characteristic with the characteristic value of the multimedia characteristic read by the characteristic service device to obtain a spliced characteristic, and obtaining the matching degree of the user information and each multimedia information according to the spliced characteristic comprises the following steps:

the characteristic values of the user characteristics and the characteristic values of the multimedia characteristics of the corresponding multimedia information returned by the characteristic service device are spliced through the M recommendation units respectively to obtain spliced characteristics; and acquiring the matching degree of the user information and the corresponding multimedia information according to the splicing characteristics.

8. The method of claim 6, wherein the obtaining the matching degree between the user information and the corresponding multimedia information according to the stitching feature comprises:

9. The method of claim 7, wherein said assigning said plurality of multimedia information to said M recommendation units comprises:

the multimedia information is evenly distributed to the M recommending units; or distributing the plurality of multimedia information to the M recommending units according to the load balancing principle according to the load of each recommending unit.

10. The method according to any one of claims 7-9, wherein after the obtaining the matching degree between the user information and each multimedia information, the method further comprises:

11. The multimedia recommending method is characterized by being applied to a feature service device in a multimedia recommending system, wherein the multimedia recommending system further comprises a recommending device and an off-line feature training device, the recommending device is connected with the feature service device, and the feature service device comprises N feature service units, and N is an integer larger than 1; each feature service unit is used for storing part of the multimedia features, and no intersection exists between the part of the multimedia features stored by any two feature service units; the method comprises the following steps:

Receiving an information identifier of each piece of multimedia information sent by a recommending device, wherein the recommending device is used for acquiring user characteristics corresponding to the user information after receiving a recommending processing request carrying the user information and a plurality of pieces of multimedia information, and sending the information identifier of the multimedia information to the characteristic service device;

reading the characteristic value corresponding to the information identification of each multimedia information from the corresponding relation between the information identification and the characteristic value of each multimedia feature in the stored part of multimedia features according to the information identification of each multimedia information through the N feature service units, wherein the characteristic value of each multimedia feature is a character-like multimedia feature and is used for indicating the specific content of the corresponding multimedia feature;

respectively returning the read characteristic value of each multimedia characteristic to the recommending device through the N characteristic service units;

receiving user characteristics sent by the recommending device, reading characteristic values corresponding to the user characteristics from the stored corresponding relation between the user characteristics and the characteristic values according to the received user characteristics, and returning the read characteristic values of the user characteristics to the recommending device, wherein the characteristic values corresponding to each user characteristic are character-like user characteristics and are used for indicating specific contents of the corresponding user characteristics;

The recommending device is used for splicing the characteristic values of the user characteristics and the read characteristic values of the multimedia characteristics to obtain splicing characteristics, and the matching degree of the user information and each multimedia information is obtained according to the splicing characteristics;

12. A multimedia recommendation method, wherein the method is applied to an offline feature training device in a multimedia recommendation system, the multimedia recommendation system further comprises a feature service device and a recommendation device, the recommendation device is connected with the feature service device, and the method comprises:

training multimedia features, and synchronizing the multimedia features into a feature service device; the feature service device comprises N feature service units, wherein N is an integer greater than 1; each feature service unit stores part of multimedia features, and no intersection exists between any two of the part of multimedia features stored by the feature service units, wherein the multimedia features comprise the corresponding relation between the information identification of each multimedia feature and the feature value;

storing the multimedia features to the determined feature service unit;

the feature service device is used for receiving the information identification and the user features of each piece of multimedia information sent by the recommendation device, and the recommendation device is used for acquiring the user features corresponding to the user information after receiving the recommendation processing request carrying the user information and a plurality of pieces of multimedia information, and sending the information identification and the user features of the multimedia information to the feature service device;

the feature service unit is used for reading the feature value corresponding to the information identifier of each multimedia information from the corresponding relation between the information identifier of each multimedia feature and the feature value in the stored part of multimedia features according to the information identifier of each multimedia information, and returning the read feature value of each multimedia feature to the recommendation device, wherein the feature value of each multimedia feature is a character-like multimedia feature and is used for indicating the specific content of the corresponding multimedia feature;

the feature service device is further configured to receive the user features sent by the recommendation device, read feature values corresponding to the user features from the stored correspondence between the user features and the feature values according to the received user features, and return the read feature values of the user features to the recommendation device, where each feature value corresponding to the user features is a character user feature and is used for indicating specific content of the corresponding user feature;

The recommendation device is used for splicing the characteristic value of the user characteristic with the read characteristic value of the multimedia characteristic to obtain a spliced characteristic, and the matching degree of the user information and each multimedia information is obtained according to the spliced characteristic.

13. The method of claim 12, wherein the determining a feature service unit in which the multimedia feature is stored when the multimedia feature is added comprises:

14. A recommendation device, wherein the recommendation device is located in a multimedia recommendation system, the multimedia recommendation system further comprises a feature service device and an off-line feature training device, the recommendation device is connected with the feature service device, and the device comprises:

a recommendation processing request receiving module configured to perform receiving a recommendation processing request carrying user information and a plurality of multimedia information;

The matching degree acquisition module is configured to splice the characteristic value of the user characteristic with the characteristic value of the multimedia characteristic read by the characteristic service device to obtain a spliced characteristic, and acquire the matching degree of the user information and each multimedia information according to the spliced characteristic;

15. The apparatus of claim 14, wherein the recommending means comprises M recommending units, M being an integer greater than 1;

the matching degree obtaining module is configured to perform splicing of the feature values of the user features and the feature values of the multimedia features of the corresponding multimedia information returned by the feature service device through the M recommending units respectively to obtain spliced features; and acquiring the matching degree of the user information and the corresponding multimedia information according to the splicing characteristics.

16. The apparatus of claim 14, wherein the matching degree obtaining module is configured to perform inputting the stitching feature into a probabilistic predictive model to obtain the matching degree of the user information and each multimedia information.

17. The apparatus of claim 15, wherein the information identity transmission module is configured to perform an average allocation of the plurality of multimedia information to the M recommendation units; or distributing the plurality of multimedia information to the M recommending units according to the load balancing principle according to the load of each recommending unit.

18. The apparatus according to any one of claims 15-17, wherein the apparatus further comprises:

19. The feature service device is characterized by being positioned in a multimedia recommendation system, the multimedia recommendation system further comprises a recommendation device and an off-line feature training device, the recommendation device is connected with the feature service device, and the feature service device comprises N feature service units, wherein N is an integer greater than 1; each feature service unit is used for storing part of the multimedia features, and no intersection exists between the part of the multimedia features stored by any two feature service units; the device comprises:

The information identification receiving module is configured to execute receiving the information identification of each piece of multimedia information sent by the recommending device, and the recommending device is used for acquiring the user characteristics corresponding to the user information after receiving the recommending processing request carrying the user information and a plurality of pieces of multimedia information and sending the information identification of the multimedia information to the characteristic service device;

the multimedia feature reading module is configured to read a feature value corresponding to the information identifier of each multimedia feature from the corresponding relation between the stored information identifier of each multimedia feature and the feature value of each multimedia feature according to the information identifier of each multimedia feature through the N feature service units, wherein the feature value of each multimedia feature is a character-like multimedia feature and is used for indicating the specific content of the corresponding multimedia feature;

a module for performing the steps of: receiving user characteristics sent by the recommending device, reading characteristic values corresponding to the user characteristics from the stored corresponding relation between the user characteristics and the characteristic values according to the received user characteristics, and returning the read characteristic values of the user characteristics to the recommending device, wherein the characteristic values corresponding to each user characteristic are character-like user characteristics and are used for indicating specific contents of the corresponding user characteristics;

The multimedia feature sending module is configured to return the feature value of each read multimedia feature to the recommending device, and the recommending device is used for splicing the feature value of the user feature and the feature value of the read multimedia feature to obtain a spliced feature, and the matching degree of the user information and each multimedia information is obtained according to the spliced feature;

20. An off-line feature training device, characterized in that, the off-line feature training device is located in a multimedia recommendation system, the multimedia recommendation system further includes a feature service device and a recommendation device, the recommendation device is connected with the feature service device, the device includes:

a feature training module configured to perform training of multimedia features, synchronizing the multimedia features into a feature service device; the feature service device comprises N feature service units, wherein N is an integer greater than 1; each feature service unit stores part of multimedia features, and no intersection exists between any two of the part of multimedia features stored by the feature service units, wherein the multimedia features comprise the corresponding relation between the information identification of each multimedia feature and the feature value;

a feature storage module configured to perform storing the multimedia features to the determined feature service unit;

21. The apparatus according to claim 20, wherein the feature service unit determining module is configured to perform a hash operation on the multimedia feature when adding the multimedia feature, to obtain a hash value of the multimedia feature; and mapping the N according to the hash value to obtain the feature service unit in which the multimedia features are stored.

22. A multimedia recommendation device, the device comprising: one or more processors;

Volatile or non-volatile memory for storing the one or more processor-executable instructions;

wherein the one or more processors are configured to perform the multimedia recommendation method of any one of claims 6 to 10, 11, 12 to 13.

23. A non-transitory computer readable storage medium having instructions stored thereon, which when executed by a processor of a server, implement the multimedia recommendation method of any one of claims 6 to 10, 11, 12 to 13.