CN113538031A

CN113538031A - Method and device for training multimedia resource allocation model and multimedia resource allocation

Info

Publication number: CN113538031A
Application number: CN202011293616.4A
Authority: CN
Inventors: 郭永辉; 姜磊; 谭斌; 杨滔; 黄东波
Original assignee: Tencent Technology Beijing Co Ltd
Current assignee: Tencent Technology Beijing Co Ltd
Priority date: 2020-11-18
Filing date: 2020-11-18
Publication date: 2021-10-22

Abstract

The application belongs to the technical field of data processing and discloses a method and a device for training a multimedia resource allocation model and allocating multimedia resources. Therefore, the multimedia resource allocation can be adjusted in real time, the accuracy of multimedia resource allocation is improved, and the play shortage of multimedia resources is reduced.

Description

Method and device for training multimedia resource allocation model and multimedia resource allocation

Technical Field

The present application relates to the field of data processing technologies, and in particular, to a method and an apparatus for training a multimedia resource allocation model and allocating multimedia resources.

Background

With the development of internet technology, playing media such as social applications and video applications are increasing. In order to promote a product, a combined delivery mode or a single-screen delivery mode is generally adopted, and multimedia resources such as advertisements for promoting the product are delivered through a playing medium.

The joint delivery is to deliver the same multimedia resource in different playing media at the same time, and the single-screen delivery is to deliver the same multimedia resource in one playing medium.

However, how to precisely adjust the allocation of each multimedia resource so that the playing amount of the multimedia resource reaches the predetermined playing amount is a problem to be solved.

Disclosure of Invention

The embodiment of the application provides a method and a device for training a multimedia resource allocation model and allocating multimedia resources, which are used for improving the accuracy of multimedia resource allocation and reducing the play shortage of multimedia resources when multimedia resources allocated to a play medium are adjusted.

In one aspect, a method for training a multimedia resource allocation model is provided, including:

generating a playing state parameter set according to the playing amount index and the played amount of the acquired at least one multimedia resource in each playing medium at the first time;

determining a corresponding multimedia resource playing adjustment mode according to the multimedia resource allocation model and the playing state parameter set, wherein the multimedia resource playing adjustment mode is used for adjusting the playing amount indexes of at least two playing media;

determining the estimated playing amount after the multimedia resource playing adjustment mode is executed according to the multimedia resource distribution model and the playing state parameter set;

acquiring the actual play shortage after executing the multimedia resource play adjustment mode;

determining the loss of the playing amount according to the estimated playing amount and the actual playing shortage;

and adjusting parameters of the multimedia resource distribution model according to the play loss.

In one aspect, a multimedia resource allocation method is provided, where a multimedia resource allocation model provided by any one of the above multimedia resource allocation model training methods includes:

acquiring a play state parameter set of at least one multimedia resource in a plurality of play media;

acquiring a target multimedia resource playing adjustment mode according to the multimedia resource allocation model and the playing state parameter set;

adjusting the playing amount index of at least one multimedia resource in a plurality of playing media according to the playing adjustment mode of the target multimedia resource;

and according to the adjusted play amount index, performing multimedia resource allocation adjustment.

In one aspect, an apparatus for training a multimedia resource allocation model is provided, including:

the generating unit is used for generating a playing state parameter set according to the playing amount index and the played amount of the acquired at least one multimedia resource in each playing medium at the first time;

a first determining unit, configured to determine a corresponding multimedia resource playing adjustment manner according to the multimedia resource allocation model and the playing state parameter set, where the multimedia resource playing adjustment manner is used to adjust the playing amount indexes of at least two playing media;

the pre-estimation unit is used for determining the pre-estimated playing amount after the multimedia resource playing adjustment mode is executed according to the multimedia resource distribution model and the playing state parameter set;

the acquisition unit is used for acquiring the actual play shortage after the multimedia resource play adjustment mode is executed;

the second determining unit is used for determining the loss of the playing amount according to the estimated playing amount and the actual playing shortage;

and the adjusting unit is used for adjusting parameters of the multimedia resource distribution model according to the play loss.

Preferably, the first determination unit is further configured to:

acquiring resource identification information corresponding to at least one multimedia resource, playing medium identification information corresponding to a playing medium and playing type identification information corresponding to a playing type of the multimedia resource;

performing data characteristic conversion on the acquired resource identification information, playing medium identification information and playing type identification information to acquire corresponding resource identification vectors, playing medium identification vectors and playing type identification vectors;

and adding the obtained resource identification vector, the playing medium identification vector and the playing type identification vector to the playing state parameter set.

Preferably, the first determination unit is further configured to:

and sequentially carrying out Hash operation, feature extraction and dimension conversion on the acquired resource identification information, playing medium identification information and playing type identification information to acquire corresponding resource identification vectors, playing medium identification vectors and playing type identification vectors.

Preferably, the first determination unit is configured to:

determining the play shortage of at least one multimedia resource according to the play amount index and the played amount of the at least one multimedia resource in each play medium;

acquiring a first multimedia resource playing adjustment mode correspondingly set by the playing shortage;

determining a corresponding second multimedia resource playing adjustment mode when the estimated playing amount is maximum according to the playing amount index, the played amount and the multimedia resource playing type of at least one multimedia resource in each playing medium respectively;

acquiring a first probability correspondingly set by a first multimedia resource playing adjustment mode and a second probability correspondingly set by a second multimedia resource playing adjustment mode;

and selecting a multimedia resource playing adjustment mode corresponding to the playing state parameter set from the first multimedia resource playing adjustment mode and the second multimedia resource playing adjustment mode according to the first probability and the second probability.

Preferably, the first determination unit is configured to:

determining a second multimedia resource playing adjustment mode set corresponding to the first time according to the playing amount index, the played amount and the multimedia resource playing type of at least one multimedia resource in each playing medium respectively;

respectively determining the estimated playing amount after each second multimedia resource playing adjustment mode in the second multimedia resource playing adjustment mode set is executed;

determining the maximum pre-estimated playing amount in each pre-estimated playing amount corresponding to the second multimedia resource playing adjustment mode set;

and determining a second multimedia resource playing adjustment mode corresponding to the maximum estimated playing amount.

Preferably, the second determination unit is configured to:

acquiring a play state parameter set corresponding to at least one multimedia resource at a second time, wherein a preset interval duration is arranged between the second time and the first time;

determining the maximum estimated playing amount corresponding to the second time according to the playing state parameter set of the second time;

determining a target playing amount according to the actual playing shortage and the maximum estimated playing amount corresponding to the second time, wherein the target playing amount is in negative correlation with the actual playing shortage and is in positive correlation with the maximum estimated playing amount corresponding to the second time;

and determining the play amount loss according to the estimated play amount corresponding to the first time and the target play amount.

Preferably, the adjusting unit is further configured to:

obtaining the number of times of model training;

judging whether the training times of the model are higher than the preset training times or not, if so, obtaining a trained multimedia resource distribution model;

otherwise, taking the sum of the first time and the preset interval duration as a new first time, and according to the new first time, executing the step of generating the playing state parameter set according to the playing amount index and the played amount of the acquired at least one multimedia resource in each playing medium at the first time.

In one aspect, a multimedia resource allocation apparatus is provided, where a multimedia resource allocation model provided by any one of the above multimedia resource allocation model training methods includes:

the multimedia resource playing device comprises an acquisition unit, a playing state parameter set acquisition unit and a playing state parameter setting unit, wherein the acquisition unit is used for acquiring a playing state parameter set of at least one multimedia resource in a plurality of playing media;

the obtaining unit is used for obtaining a target multimedia resource playing adjustment mode according to the multimedia resource distribution model and the playing state parameter set;

the first adjusting unit is used for adjusting the playing quantity index of at least one multimedia resource in a plurality of playing media according to the playing adjusting mode of the target multimedia resource;

and the second adjusting unit is used for carrying out multimedia resource allocation adjustment according to the adjusted play amount index.

In one aspect, a control device is provided, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the steps of the method for training a multimedia resource allocation model and allocating multimedia resources.

In one aspect, a computer readable storage medium is provided, on which a computer program is stored, which when executed by a processor implements the steps of any of the above methods for training a multimedia resource allocation model and for allocating multimedia resources.

In one aspect, a computer program product or computer program is provided that includes computer instructions stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method provided in any one of the above-mentioned training of the multimedia resource allocation model and various alternative implementation manners of multimedia resource allocation.

In the method and the device for training the multimedia resource allocation model and allocating the multimedia resources, the multimedia resource allocation model is adopted, the multimedia resource play adjustment mode is determined according to the play state parameter set, the estimated play amount after the multimedia resource play adjustment mode is executed is determined according to the play state parameter set, and the parameters of the multimedia resource allocation model are adjusted according to the estimated play amount and the actual play shortage, so that the trained multimedia resource allocation model is obtained. Furthermore, a trained multimedia resource allocation model is adopted to determine a target multimedia resource playing adjustment mode corresponding to the current playing state parameter set, and the playing amount index of each multimedia resource in each playing medium is adjusted according to the target multimedia resource playing adjustment mode so as to perform multimedia resource allocation adjustment. Therefore, the multimedia resources distributed by each playing medium can be adjusted in real time according to the acquired playing state parameter set, the multimedia resource distribution accuracy is improved, and the playing shortage of the multimedia resources is reduced.

Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the application. The objectives and other advantages of the application may be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:

fig. 1 is a schematic diagram of a multimedia resource allocation system according to an embodiment of the present application;

fig. 2a is a first diagram illustrating an example of multimedia resource delivery according to an embodiment of the present disclosure;

fig. 2b is a diagram illustrating an example of multimedia resource delivery according to an embodiment of the present disclosure;

fig. 3 is a flowchart illustrating an implementation of a method for training a multimedia resource allocation model according to an embodiment of the present disclosure;

FIG. 4 is a diagram illustrating dimension conversion according to an embodiment of the present disclosure;

FIG. 5 is a block diagram illustrating a multimedia resource allocation model according to an embodiment of the present invention;

fig. 6 is a flowchart illustrating an implementation of a multimedia resource allocation method according to an embodiment of the present disclosure;

FIG. 7a is a schematic diagram of an application scenario in an embodiment of the present application;

FIG. 7b is a diagram illustrating a play deficiency comparison example according to an embodiment of the present disclosure;

FIG. 8a is a training apparatus of a multimedia resource allocation model according to an embodiment of the present disclosure;

FIG. 8b is a schematic structural diagram of an apparatus for allocating multimedia resources according to an embodiment of the present disclosure;

fig. 9 is a schematic structural diagram of a control device in an embodiment of the present application.

Detailed Description

In order to make the purpose, technical solution and beneficial effects of the present application more clear and more obvious, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

First, some terms referred to in the embodiments of the present application will be described to facilitate understanding by those skilled in the art.

The terminal equipment: may be a mobile terminal, a fixed terminal, or a portable terminal such as a mobile handset, station, unit, device, multimedia computer, multimedia tablet, internet node, communicator, desktop computer, laptop computer, notebook computer, netbook computer, tablet computer, personal communication system device, personal navigation device, personal digital assistant, audio/video player, digital camera/camcorder, positioning device, television receiver, radio broadcast receiver, electronic book device, gaming device, or any combination thereof, including the accessories and peripherals of these devices, or any combination thereof. It is also contemplated that the terminal device can support any type of interface to the user (e.g., wearable device), and the like.

A server: the cloud server can be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, and can also be a cloud server for providing basic cloud computing services such as cloud service, a cloud database, cloud computing, cloud functions, cloud storage, network service, cloud communication, middleware service, domain name service, security service, big data and artificial intelligence platform and the like.

Cloud storage: the distributed cloud storage system (hereinafter referred to as a storage system) refers to a storage system which integrates a large number of storage devices (storage devices are also referred to as storage nodes) of different types in a network through application software or application interfaces to cooperatively work through functions of cluster application, grid technology, distributed storage file systems and the like, and provides data storage and service access functions to the outside.

At present, a storage method of a storage system is as follows: logical volumes are created, and when created, each logical volume is allocated physical storage space, which may be the disk composition of a certain storage device or of several storage devices. The application program stores data on a certain logical volume, namely, the data is stored on a file system, the file system divides the data into a plurality of parts, each part is an object, the object not only contains the data but also contains additional information such as data identification, the file system writes each object into a physical storage space of the logical volume respectively, and the file system records storage location information of each object, so that when the application program requests to access the data, the file system can enable the application program to access the data according to the storage location information of each object.

The process of allocating physical storage space for the logical volume by the storage system specifically includes: physical storage space is divided in advance into stripes according to a set of capacity measures of objects stored in a logical volume (the measures usually have a large margin with respect to the capacity of the actual objects to be stored) and Redundant Array of Independent Disks (RAID), and one logical volume can be understood as one stripe, thereby allocating physical storage space to the logical volume.

A database: in short, it can be regarded as an electronic file cabinet, i.e. a place for storing electronic files, and a user can add, query, update, delete, etc. to the data in the files. A "database" is a collection of data that is stored together in a manner that can be shared by multiple users, has as little redundancy as possible, and is independent of the application.

A database management system: the computer software system designed for managing the database generally has basic functions of storage, interception, safety guarantee, backup and the like. The database management system may be categorized according to the database models it supports, such as relational, extensible markup language, or according to the types of computers supported, such as server clusters, mobile phones; or classified according to the Query Language used, e.g., Structured Query Language (SQL), XQuery; or by performance impulse emphasis, e.g., maximum size, maximum operating speed; or other classification schemes. Regardless of the manner of classification used, some database management systems are capable of supporting multiple query languages across categories, for example, simultaneously.

Hash operation: the method is characterized in that an input with an arbitrary length is converted into an output with a fixed length through a hash algorithm. This transformation is a kind of compression mapping, and the space of the hash value is usually much smaller than the space of the input.

One-bit-efficient coding (One-Hot): an N-bit status register is used to encode N states, each state being represented by its own independent register bit and having only one bit active at any time, N being a positive integer. One-Hot encoding can convert a value into a binary vector, which is zero except for the index of the integer, which is labeled 1.

Deep enhanced networks (Deep Q networks, DQN): and taking a deep network (deep network) as a decision maker for storing state and action selection, and continuously training and optimizing through real-time feedback, thereby realizing output of corresponding action according to the current input state. DQN is a model combining deep learning and reinforcement learning.

Deep Neural Networks (Deep Neural Networks, DNN): it can be understood as a neural network with many hidden layers.

The Width and depth (W & D) model is a model used for classification and regression. The core idea is that the memory capacity of the linear model and the generalization capacity of the DNN model are combined, and the parameters of the two models are optimized simultaneously in the training process, so that the optimal prediction capacity of the whole model is achieved.

The design concept of the embodiment of the present application is described below.

With the development of internet technology, playing media such as social applications and video applications are increasing. In order to promote a product, multimedia resources such as advertisements for promoting the product are usually delivered through a playing medium.

For example, in order to maximize benefits, an operator of a video website obtains revenue by delivering advertisements before playing videos, and adopts a general Guaranteed Delivery (Guaranteed Delivery) mode, that is, pays a fee for ensuring that the playing amount of the advertisements reaches the playing amount index agreed in a contract. Therefore, the multimedia resource allocation needs to be adjusted so that the playing amount of the multimedia resource reaches the playing amount index.

In the conventional technology, before the multimedia resource is released, an empty playing amount inventory is estimated according to long-term historical playing data, a playing amount index of the multimedia resource is formulated according to the estimated empty playing amount inventory, and the multimedia resource is distributed when the multimedia resource is released. However, the stock of the free playing amount is usually estimated after long-term training, and the estimation result has deviation and cannot be dynamically adjusted in real time, so that the multimedia resource allocation accuracy is low and the playing shortage is large.

Therefore, a technical solution for training a multimedia resource allocation model and allocating multimedia resources is needed to improve the accuracy of multimedia resource allocation and reduce the play shortage of multimedia resources when adjusting multimedia resource allocation.

In view of the above analysis and consideration, the present application provides a data processing scheme, in which a multimedia resource allocation model is used to determine a multimedia resource play adjustment manner according to a play state parameter set, determine an estimated play amount after the multimedia resource play adjustment manner is executed according to the play state parameter set, and perform parameter adjustment on the multimedia resource allocation model according to the estimated play amount and an actual play shortage, so as to obtain a trained multimedia resource allocation model. Furthermore, a trained multimedia resource allocation model is adopted to determine a target multimedia resource playing adjustment mode corresponding to the current playing state parameter set, and the playing amount index of each multimedia resource in each playing medium is adjusted according to the target multimedia resource playing adjustment mode so as to perform multimedia resource allocation adjustment.

To further illustrate the technical solutions provided by the embodiments of the present application, the following detailed description is made with reference to the accompanying drawings and the detailed description. Although the embodiments of the present application provide method steps as shown in the following embodiments or figures, more or fewer steps may be included in the method based on conventional or non-inventive efforts. In steps where no necessary causal relationship exists logically, the order of execution of the steps is not limited to that provided by the embodiments of the present application. The method can be executed in sequence or in parallel according to the method shown in the embodiment or the figure when the method is executed in an actual processing procedure or a device.

The terms "first," "second," and the like in the description and in the claims, as well as in the drawings, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It will be appreciated that the data so used may be interchanged under appropriate circumstances such that the embodiments described herein may be practiced otherwise than as specifically illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

Fig. 1 is a schematic diagram of a multimedia resource allocation system according to the present application. The multimedia resource allocation system includes: a control device 101, a plurality of play media servers 102, and a plurality of user terminals 103.

The control apparatus 101: the multimedia resource allocation model can be a server or a terminal device, and is used for periodically generating a play state parameter set according to the acquired play state parameters, and training the multimedia resource allocation model according to the periodically acquired play state parameter set to obtain a trained multimedia resource allocation model; the method is also used for determining a target multimedia resource playing adjustment mode according to a playing state parameter set acquired in real time or periodically and a trained multimedia resource allocation model, and adjusting the playing quantity index of the multimedia resource in each playing medium according to the target multimedia resource playing adjustment mode, so that real-time adjustment can be performed according to real-time feedback.

The number of multimedia resources may be one or more. The multimedia asset may be an advertisement. The multimedia resource playing types comprise joint multimedia resources and single-screen multimedia resources.

The joint delivery of the multimedia resources is a mode of joint delivery, and the multimedia resources are delivered in a plurality of playing media simultaneously. For example, the co-delivered multimedia asset may be a co-delivered advertisement. The single-screen multimedia resource is a multimedia resource which is delivered in a single-screen delivery mode and is delivered in only one playing medium. For example, a single screen multimedia asset may be a single screen advertisement.

It should be noted that the stock of the free playing amount of the playing media can be determined according to the click rate of the page for playing the multimedia resource in the playing media, the stock of the free playing amount is continuously changed, that is, the stock of the free playing amount of the playing media at different times may be different, the stock of the free playing amount of different pages in the playing media is also different, and the stock of the free playing amount in the future can only be estimated through the stock of the historical free playing amount, and the estimated stock of the free playing amount is usually inaccurate. The single-screen multimedia resource is played through only one page in the playing medium. The joint multimedia resources can be played through one or more pages in each corresponding playing medium respectively. The maximum playing amount of the multimedia resources which can be played and the number of the multimedia resources which can be released can be determined in each playing medium according to the free playing amount inventory.

For example, referring to fig. 2a, an exemplary diagram of multimedia resource delivery is shown. The playing media are as follows: video applications, news applications, browser applications, and social applications. The joint advertisement A is a joint multimedia resource and is simultaneously delivered in a video application program, a news application program, a browser application program and a social application program. The single-screen advertisement is a single-screen multimedia resource, such as a single-screen advertisement A, a single-screen advertisement B, a single-screen advertisement C and a single-screen advertisement D, and is only put in a video application program. Single screen advertisement E is only delivered in the news application. There is an inventory of free play volume in news applications, browser applications, and social applications, while there is no inventory of free play volume in video applications. Fig. 2b shows an example of multimedia resource delivery. And according to the multimedia resource distribution model and the playing state parameter set, determining a target multimedia resource playing adjustment mode, and adjusting the playing amount indexes of the joint advertisement A in the video application program and the social application program.

The playback media server 102: a server that provides a simple and manageable mechanism for accessing system resources for applications.

The playing media are media for playing multimedia resources, and may be different application programs or different websites. For example, the playing media may be a video application, a news application, a browser application, a social application, and the like. The different applications may be of the same type or of different types. Different playback media may correspond to one playback media server 102 or different playback media servers 102. The playback media server 102 and the control device 101 corresponding to each playback media may be integrated into one device, or may be different devices. In practice, different playback media typically correspond to different playback media servers 102. Therefore, in the embodiment of the present application, only the control device 101 and the playback media server 102 corresponding to each playback media are taken as different devices for example.

For example, the WeChat application corresponds to a WeChat Server, the QQ application corresponds to a QQ Server, and the music application corresponds to a music Server.

Each playing media server 102 is configured to play a multimedia resource according to the playing amount index and the playing type of the multimedia resource in the corresponding playing media of each multimedia resource issued by the control device 101, report playing state parameters, such as the played amount of the multimedia resource in the playing media, acquired in real time to the control device 101, receive a playing amount index adjustment instruction returned by the control device 101 according to the playing state parameters, and perform multimedia resource allocation adjustment according to the received playing amount index adjustment instruction.

For example, the multimedia asset is a syndicated ad a. The playing media of the jointly-delivered advertisement A are as follows: video applications and news applications. The playback media servers 102 for the video application and the news application are video servers and news servers. The control device 101 issues advertisement delivery messages to the video server and the news server, respectively, so that the joint delivery advertisement a is played on the home page of the video application program and the home page of the news application program, respectively.

When the user opens the home page of the video application or the home page of the news application, the associated advertisement a is played. The video server and the news server count the played amount of the joint advertisement A in real time or periodically. The control device 101 determines the play quantity shortage according to the played quantity and the play quantity index of the associated advertisement a in the video application program and the news application program respectively, and adjusts the play quantity indexes in the video application program and the news application program respectively according to the play quantity shortage by adopting a multimedia resource distribution model, so that the video server and the news server adjust the play resource occupation ratio of the associated advertisement a in the first page of the video application program and the first page of the news application program respectively.

The user terminal 103: the terminal device can be provided with a client corresponding to the application program, so that the user can play the multimedia resource through the client.

Referring to fig. 3, a flowchart of an implementation of a method for training a multimedia resource allocation model according to the present application is shown. The method comprises the following specific processes:

step 300: the control equipment generates a playing state parameter set according to the playing amount index and the played amount of the acquired at least one multimedia resource in each playing medium at the first time.

Specifically, the control device obtains a play amount index, a played amount, and play type identification information corresponding to a play type of the multimedia resource at each play medium for at least one multimedia resource at a first time, and generates a play state parameter set according to the obtained play amount index, the played amount, and the play type identification information corresponding to the play type of the multimedia resource.

Further, the control device may further obtain resource identification information corresponding to at least one multimedia resource and playing media identification information corresponding to a playing media.

The number of multimedia resources may be one or more. The multimedia resource playing types comprise joint delivery and single-screen delivery. And if the number of the multimedia resources is only one, the playing type of the multimedia resources is the joint delivery. The number of playing media is at least two.

Wherein, the control device can execute the following steps for each multimedia resource and each corresponding playing media respectively:

s3001: the control device obtains resource identification information corresponding to a multimedia resource, playing medium identification information corresponding to a playing medium, and playing amount index and playing type identification information corresponding to the playing type of the multimedia resource in the playing medium.

In one embodiment, the playing media server and the control device each store media resource identification information corresponding to a multimedia resource, playing media identification information corresponding to a playing media, and playing type identification information corresponding to a playing type of the multimedia resource in the playing media.

The playing media server corresponding to the playing media acquires the played amount of the multimedia resource in the playing media in real time or periodically, and reports the acquired played amount to the control equipment in real time or periodically.

It should be noted that the played amount of the multimedia resource can be determined according to the click rate of the page for playing the multimedia resource.

For example, the video server is arranged to play an advertisement C in an entertainment program page B of the video application program, and when the user opens the entertainment program page B in the video application client, the video server provides the entertainment program page B containing the advertisement C to the video application client according to the received access request of the entertainment program page B, and adds one to the played amount. And the video application client presents the comprehensive program page B to the user and simultaneously plays the advertisement C.

Further, the played amount may also be determined according to the browsing duration of the multimedia resource playing page, which is not limited herein.

S3002: the control device performs data feature conversion on the acquired resource identification information, playing medium identification information and playing type identification information to acquire corresponding resource identification vectors, playing medium identification vectors and playing type identification vectors.

Specifically, the control device performs hash operation, feature extraction, and dimension conversion on the acquired resource identification information, playing medium identification information, and playing type identification information in sequence to obtain corresponding resource identification vectors, playing medium identification vectors, and playing type identification vectors.

The playing amount index and the played amount are numerical values, and the corresponding data type is a continuity variable. The data type of the asset Identification information, the playing media Identification information, and the playing type Identification information is an Identification (ID) type.

Because the ID space is large and sparse, the hash operation is firstly carried out on each identification information of the ID type to obtain a corresponding hash value with a fixed length, so that the data storage space is saved and the data processing speed is improved.

Optionally, One-Hot encoding may be adopted during feature extraction, and the hash value is converted into a binary vector with only One 1. Wherein the obtained binary vector is a discrete vector.

Optionally, during dimension conversion, an embedded network (Embedding) network may be used. Embedding is used to convert discrete variables into continuous vectors. The control device can perform dimension conversion on each obtained binary vector through Embedding to obtain each continuous type vector, namely a resource identification vector, a playing medium identification vector and a playing type identification vector.

For example, referring to fig. 4, a diagram of dimension conversion is shown. The control device performs hash operation and feature extraction on the resource identification information of the advertisement A by adopting a hash algorithm and One-Hot coding to obtain a binary vector 01000. The control device adopts Embedding to carry out dimension conversion on 01000 to obtain a continuous type vector [0.191, 0.063, 0.526 ].

Therefore, the playing state parameters of each ID type can be subjected to feature conversion to obtain corresponding continuous variable vectors.

S3002: the control device adds the playing amount index, the played amount, the resource identification vector, the playing medium identification vector and the playing type identification vector of the multimedia resource in the playing medium at the first time to the playing state parameter subset corresponding to the first time.

Wherein, a play state parameter subset corresponding to a multimedia resource and a play medium at a first time t comprises: the resource identification vector of the multimedia resource, the playing media identification vector of the playing media, the playing type identification vector corresponding to the playing type of the multimedia resource in the playing media, and the playing amount index and the played amount of the multimedia resource in the playing media at the moment t.

Where t represents a first time, e.g., t is 5 points.

It should be noted that, the control device periodically generates the play state parameter set according to the preset interval duration, and therefore, the first time is continuously updated according to the preset interval duration.

Further, the control device may combine the play state parameter subsets corresponding to the multimedia resources and the play media to form a play state parameter set.

Further, the control device may further perform the following steps for each multimedia resource, respectively: determining the total playing quantity index and the played quantity of the multimedia resource in each playing medium, generating a corresponding playing state parameter subset according to the playing quantity index and the played quantity of the multimedia resource in each playing medium, and forming a playing state parameter set according to the playing state parameter subset corresponding to each multimedia resource.

Wherein, a play state parameter subset corresponding to a multimedia resource comprises: the resource identification vector of the multimedia resource, and the playing media identification vector, playing amount index and played amount corresponding to each playing media.

It should be noted that, in the embodiment of the present application, only the playback volume indicator, the played volume, the resource identifier vector, the playback media identifier vector, and the playback type identifier vector are taken as the playback state parameters for illustration, and in practical applications, the playback state parameters may also include other playback-related parameters, which are not limited herein.

In this way, the play state parameter set at the current time can be obtained.

Step 301: and the control equipment determines a corresponding multimedia resource playing adjustment mode according to the multimedia resource allocation model and the playing state parameter set.

Specifically, when step 301 is executed, the following steps may be adopted:

s3011: the control device determines the play shortage of at least one multimedia resource according to the play amount index and the played amount of at least one multimedia resource in each play medium.

Specifically, the control device executes the following steps for each multimedia resource and each corresponding playing medium, respectively: determining the difference value between the playing quantity index and the played quantity of a multimedia resource in a playing medium, and obtaining the playing shortage of the multimedia resource in the playing medium.

Then, the control device determines the sum of the play shortage of at least one multimedia resource in each play medium to obtain the total play shortage.

For example, if the play shortage of the joint advertisement D in the news application is 1000, the play shortage of the video application is 5000, and the play shortage of the game application is-2000, the total play shortage of the joint advertisement D in each playing medium is 5000+1000-2000 ═ 4000.

Thus, the total play shortage of each multimedia resource on each playing medium can be determined.

S3012: the control equipment acquires a first multimedia resource playing adjustment mode correspondingly set by the playing shortage.

Specifically, before executing S3012, the control device sets in advance a corresponding relationship between the play shortage and the first multimedia resource play adjustment manner.

Optionally, the corresponding relationship may be set according to experience or experiment.

The multimedia resource playing adjustment mode is used for adjusting the playing quantity indexes of at least two playing media.

Specifically, only the multimedia resource that uses the joint delivery method can be adjusted in different playing media, so the multimedia resource playing adjustment method is specifically used for adjusting the playing amount index of the joint multimedia resource in at least two playing media.

That is, the playing amount index of the joint multimedia resource in one or more playing media needs to be increased, and the playing amount index of the joint multimedia resource in other one or more playing media needs to be adjusted downward, so that the total playing amount index of the joint multimedia resource in each playing media is unchanged after adjustment.

When the playing amount index is adjusted, the adjustment amount may be a preset specified adjustment amount, or may be determined according to a specified index ratio of the playing amount index, which is not limited herein.

In practical application, both the specified index ratio and the specified adjustment quantity may be set according to a practical application scenario, or may be set according to experience or experiments, for example, the specified index ratio is 2%, and for example, the specified adjustment quantity is 1000, which is not limited herein.

For example, in FIG. 2a, co-delivered ad A is delivered simultaneously in a video application, a news application, a browser application, and a social application. Assuming that the designated adjustment amount is 1000, the first multimedia asset playing adjustment manner may be: the control equipment reduces the playing quantity index of the joint advertisement A in the video application program by 1000, and increases the playing quantity index of the joint advertisement A in the social application program by 1000.

For another example, in fig. 2a, assuming that the total playing amount index of the syndicated advertisement a in the video application, the news application, the browser application, and the social application is 10000, and the designated index accounts for 10%, the adjustment amount is 10000 × 10%, which is 1000, and the first multimedia asset playing adjustment manner may be: the control equipment reduces the playing quantity index of the joint advertisement A in the video application program by 1000, and increases the playing quantity index of the joint advertisement A in the social application program by 1000.

It should be noted that, if the playing volume index of the playing media needing to be adjusted downward is lower than the adjustment quantity needing to be adjusted downward, it indicates that the adjustment quantity is larger, and the adjustment quantity is adjusted, so that the playing volume index of the playing media needing to be adjusted downward is not lower than the adjustment quantity needing to be adjusted downward.

In one embodiment, the minimum playing amount index of the playing media to be adjusted downward may be obtained, and the adjustment amount may be obtained by multiplying the minimum playing amount index by the specified index.

In practical applications, the number of playing media to be adjusted in the first multimedia resource playing adjustment manner and the adjustment number of the playing amount index of each playing media may be set according to practical application scenarios, which is not limited herein.

S3013: the control device determines a corresponding second multimedia resource playing adjustment mode when the estimated playing amount is maximum according to the playing amount index, the played amount and the multimedia resource playing type of at least one multimedia resource in each playing medium.

Specifically, the control device determines the maximum estimated playing amount and a second multimedia resource playing adjustment mode corresponding to the maximum estimated playing amount according to the multimedia resource distribution model and the playing state parameter set.

Fig. 5 is a schematic diagram of a multimedia resource allocation model. The multimedia resource distribution model comprises an input and output module, an estimation module, a target module, a loss module and a storage module.

Wherein, the input/output module: and the system is used for inputting the playing state parameter set St at the moment t and outputting the determined multimedia resource playing adjustment mode at.

Further, a play state parameter set St +1 at time t +1 may also be input.

A prediction module: used for determining the multimedia resource playing adjustment mode at and the estimated playing amount Qt after executing the multimedia resource playing adjustment mode at.

A target module: the method is used for determining the actual playing shortage Rt after the multimedia resource playing adjustment mode at is executed, determining the maximum estimated playing quantity Qt +1 at the moment t +1 according to the playing state parameter set St +1 at the second time, namely the moment t +1, and determining the target playing quantity Yt according to the actual playing shortage Rt and the maximum estimated playing quantity Qt + 1.

Alternatively, the multimedia resource allocation model may be DQN. The prediction module and the target module can adopt different depth networks, for example, a DNN network, a Linear Regression (Linear Regression) network of one layer or a W & D model.

A loss module: the system is used for determining the playing quantity loss L (w), and returning the playing quantity loss L (w) to the estimation module, so that the estimation module adjusts parameters of the estimation module according to the playing quantity loss L (w), and copies the adjusted parameters to the target module, so that the target module performs parameter synchronous updating. Wherein w is a model parameter.

A storage module: the method is used for data storage, and can store a playing state parameter set, a playing adjustment mode of each multimedia resource, actual playing shortage and the like at each moment in the training process.

Optionally, the storage module may store in a cloud storage or database storage manner, and may manage the database through a data management system.

Further, before the model training, the multimedia resource allocation model is initialized in advance by using the acquired play state parameter set and the randomly selected multimedia resource play adjustment mode.

Further, the multimedia resource allocation model may also use a proportional, integral, and differential (PID) model, which is not limited herein.

When determining the second multimedia resource playing adjustment mode corresponding to the maximum estimated playing amount, the following steps may be adopted:

step a: the control device adopts an estimation module to determine a second multimedia resource playing adjustment mode set corresponding to the first time according to the playing amount index, the played amount and the multimedia resource playing type of at least one multimedia resource in each playing medium respectively.

Specifically, the second multimedia resource playing adjustment mode set is a set of at least one second multimedia resource playing adjustment mode.

In one embodiment, the second set of multimedia asset playing adjustment manners is a preset set, that is, the second set of multimedia asset playing adjustment manners is fixed.

In one embodiment, the jointly-cast multimedia resources are screened out according to the playing types of the multimedia resources, and a plurality of second multimedia resource playing adjustment modes are determined according to the playing amount indexes and the played amount of the jointly-cast multimedia resources in each playing medium. I.e. the second set of multimedia asset playing adjustment modes is constantly changing.

The specific operation manner of the second multimedia resource playing adjustment manner is referred to the first multimedia resource playing adjustment manner in S3012, and is not described herein again.

For example, in fig. 2a, it is assumed that the total playing amount index of the syndicated advertisement a in the video application, the news application, the browser application and the social application is 10000, and the specified index accounts for 10%, that is, the adjustment amount is 10000 × 10%, which is 1000.

The second multimedia resource playing adjustment mode 1 is: the control equipment reduces the playing quantity index of the joint advertisement A in the video application program by 1000, and increases the playing quantity index of the joint advertisement A in the social application program by 1000. The second multimedia resource playing adjustment mode 2 is as follows: the control equipment reduces the playing quantity index of the jointly cast advertisement A in the video application program by 2000, and increases the playing quantity index of the jointly cast advertisement A in the social application program by 1000, and increases the playing quantity index of the jointly cast advertisement A in the news application program by 1000. … … the second multimedia asset playing adjustment mode n is … …. Wherein n is a positive integer.

In practical applications, the second multimedia resource playing adjustment manner included in the second multimedia resource playing adjustment manner set may be set according to a practical application scene, and is not limited herein.

Therefore, each second multimedia resource playing adjustment mode which can be executed can be determined according to the playing state parameter set at the current moment.

Step b: and the control equipment adopts an estimation module to respectively determine the estimated playing amount after each second multimedia resource playing adjustment mode in the second multimedia resource playing adjustment mode set is executed.

Step c: and the control equipment determines the maximum pre-estimated playing quantity in the pre-estimated playing quantities corresponding to the second multimedia resource playing adjustment mode set.

Optionally, the maximum estimated playback amount corresponding to the time t may be represented as:

(St，a，w)；

wherein At is a second multimedia resource playing adjustment mode set corresponding to the time t, a is a second multimedia resource playing adjustment mode, and w is a model parameter.

Step d: and the control equipment determines a second multimedia resource playing adjustment mode corresponding to the maximum estimated playing amount.

Thus, the second multimedia resource playing adjustment mode executed when the maximum estimated playing amount is determined can be determined.

S3014: the control equipment acquires a first probability set corresponding to the first multimedia resource playing adjustment mode and a second probability set corresponding to the second multimedia resource playing adjustment mode.

Before executing S3014, the control device may set corresponding probabilities for the first multimedia asset playing adjustment manner and the second multimedia asset playing adjustment manner in advance, and a sum of the first probability and the second probability is 1.

The first probability and the second probability may be set empirically, and in practical applications, the first probability and the second probability may also be set according to practical application scenarios, for example, the first probability is 0.1, and the second probability is 0.9. And are not intended to be limiting herein.

S3015: and the control equipment selects a multimedia resource playing adjustment mode corresponding to the playing state parameter set from the first multimedia resource playing adjustment mode and the second multimedia resource playing adjustment mode according to the first probability and the second probability.

That is, the probability that the control device selects the first multimedia resource playing adjustment mode is the first probability, and the probability that the control device selects the second multimedia resource playing adjustment mode is the second probability. And the control equipment takes the selected multimedia resource playing adjustment mode as a multimedia resource playing adjustment mode corresponding to the playing state parameter set.

In this way, the multimedia resource playing adjustment mode corresponding to the playing state parameter set is selected from the first multimedia resource playing adjustment mode correspondingly set to the playing shortage and the second multimedia resource playing adjustment mode corresponding to the time when the predicted playing quantity is maximum and determined by adopting the prediction algorithm of the prediction module, and the playing shortage and the prediction algorithm are comprehensively considered, so that the accuracy of determining the multimedia resource playing adjustment mode is improved.

Step 302: and the control equipment determines the estimated playing amount after the multimedia resource playing adjustment mode is executed according to the multimedia resource distribution model and the playing state parameter set.

Specifically, the control device inputs the playing state parameter set and the determined multimedia resource playing adjustment mode to the pre-estimation module, and outputs the pre-estimated playing amount.

That is, through the estimation module, when the current time is predicted to be the play state parameter set St, after the multimedia resource play adjustment mode at is executed, the estimated play amount Qt of each multimedia resource in each play medium is estimated.

Step 303: the control equipment acquires the actual play shortage after the multimedia resource play adjustment mode is executed.

Specifically, the control device sends the determined multimedia resource playing adjustment mode at to the corresponding playing media server through the input/output module, so that each playing media server executes the received multimedia resource playing adjustment mode at, and determines the actual playing shortage after the multimedia resource playing adjustment mode is executed according to the played amount and the playing amount index returned by each playing media server.

Therefore, the actual playing shortage at the time t can be obtained through real-time feedback.

Step 304: and the control equipment determines the loss of the playing amount according to the estimated playing amount and the actual playing shortage.

Specifically, when step 304 is executed, the following steps may be adopted:

s3041: and the control equipment acquires a play state parameter set corresponding to at least one multimedia resource at a second time.

Specifically, the control device receives the playing state parameters uploaded by each playing media server, and obtains a playing state parameter set St +1 corresponding to the second time according to the playing state parameters corresponding to the second time.

The second time is separated from the first time by a preset interval duration.

In practical applications, the preset interval duration may be set according to a practical application scenario, for example, the preset interval duration is 1 hour, which is not limited herein.

S3042: and the control equipment determines the maximum estimated playing amount corresponding to the second time according to the playing state parameter set of the second time.

Specifically, the control device inputs the play state parameter set at the second time to the target module, and outputs the maximum estimated play amount corresponding to the second time.

Optionally, the maximum estimated playback amount corresponding to the time t +1 may be represented as:

wherein At +1 is a second multimedia resource playing adjustment mode set corresponding to the time t +1, Q is a maximum estimated playing amount, a represents a second multimedia resource playing adjustment mode in At +1, and w is a model parameter.

The specific step of determining the maximum estimated playback volume corresponding to the second time may be referred to as S3013, which is not described herein again.

S3043: and the control equipment determines the target playing quantity according to the actual playing shortage and the maximum estimated playing quantity corresponding to the second time.

The target playing quantity is in negative correlation with the actual playing shortage, and is in positive correlation with the maximum estimated playing quantity corresponding to the second time.

In one embodiment, the target playback volume may be determined using the following formula:

where Yt is a target playback volume At time t, At +1 is a second multimedia resource playback adjustment mode set corresponding to time t +1, Q is an estimated playback volume, a represents a multimedia resource playback adjustment mode in A t +1, w is a model parameter, Rt represents an actual playback deficiency At time t, and γ is a weight coefficient, which may be set according to an actual application scenario, and is not limited herein.

S3044: and the control equipment determines the play amount loss according to the estimated play amount corresponding to the first time and the target play amount.

Specifically, the control device inputs the estimated playing amount and the target playing amount corresponding to the first time to the loss module to obtain the playing amount loss.

The play loss is negatively correlated with the estimated play amount corresponding to the first time and positively correlated with the target play amount.

In one embodiment, when determining the loss of the playback volume, the following formula may be used:

L(w)＝E(Yt-Qt)²；

wherein, l (w) is the playing volume loss, w is the model parameter, E () represents the expectation function, Yt is the target playing volume at time t, Qt represents the estimated playing volume corresponding to time t.

In this way, the play amount loss can be determined.

Step 305: and the control equipment adjusts parameters of the multimedia resource distribution model according to the play loss.

Specifically, the control device inputs the play amount loss output by the loss module to the estimation module, the estimation module adjusts the model parameters according to the play amount loss, and periodically synchronizes the adjusted model parameters to the target module.

Step 306: the control device determines whether the obtained training times of the model are higher than a preset training time, if so, step 307 is executed, otherwise, step 308 is executed.

Specifically, the initial value of the number of model training times is set to a specified value, e.g., 0. If the adjustment of the model parameters is determined to be finished, the control device adds one to the number of model training times, and judges whether the updated number of model training times is higher than the preset number of training times, if so, the step 307 is executed, otherwise, the step 308 is executed.

In practical applications, the preset training times may be set according to practical application scenarios, for example, the preset training times is T × M, which is not limited herein.

Wherein, T is the time period of single round training, and M is the training round.

Step 307: the control device obtains a trained multimedia resource allocation model.

Thus, the multimedia resource allocation model can be applied to allocate the multimedia resources.

Step 308: the control device takes the sum of the first time and the preset interval duration as a new first time, and performs step 300.

That is, time t is updated to time t + 1.

Thus, periodic adjustments and training can be performed based on real-time feedback.

Referring to fig. 6, a flowchart of an implementation of a multimedia resource allocation method provided in the present application is shown. The method comprises the following specific processes:

step 600: the control device obtains a play state parameter set of at least one multimedia resource in a plurality of play media.

Specifically, the control device periodically obtains a play state parameter set of at least one multimedia resource in a plurality of play media.

Step 601: and the control equipment acquires a target multimedia resource playing adjustment mode according to the multimedia resource allocation model and the acquired playing state parameter set.

Step 602: the control equipment adjusts the playing amount index of at least one multimedia resource in a plurality of playing media according to the playing adjustment mode of the target multimedia resource.

Step 603: and the control equipment performs multimedia resource allocation adjustment according to the adjusted play amount index.

Specifically, the control device sends the adjusted play amount index to the corresponding play media server, so that the play media server puts in the multimedia resource according to the adjusted play amount index.

Fig. 7a is a schematic diagram of an application scenario. Each playing media server, i.e., the video server, the news server, the browser server … …, the social server, etc., respectively obtains the playing status parameters, such as the total playing amount of each multimedia resource at each client through the tracking proxy application, and reports the obtained playing status parameters to the control device in real time or periodically. And the control equipment performs model training according to the received playing state parameters and performs multimedia resource allocation through the trained multimedia resource allocation model.

Fig. 7b is a diagram showing a playback shortage comparison example. The ordinate is the play deficit rate, i.e. the ratio of the play deficit in the play amount index. Each application program corresponds to two bar graphs, the first bar graph represents the play deficit rate before the multimedia resource allocation is adjusted, and the second bar graph represents the play deficit rate after the multimedia resource allocation is adjusted. Obviously, after the multimedia resource allocation method provided by the application is adopted, the playing shortage rate of each multimedia resource is reduced, so that the whole playing shortage rate is also reduced.

In the embodiment of the application, the played amount and the play shortage of each multimedia resource in each playing medium can be obtained in real time, and the playing index of each multimedia resource in each playing medium is adjusted in real time according to the play shortage by adopting a multimedia resource allocation model, so that the real-time feedback and adjustment of the playing of the multimedia resource are realized, the accuracy of multimedia resource allocation is improved, the total play shortage of the multimedia resource is reduced, and the play amount of the multimedia resource can reach the expected playing target.

Based on the same inventive concept, the embodiment of the present application further provides a training device for a multimedia resource allocation model and a multimedia resource allocation device, and because the principle of solving the problem of the above device and equipment is similar to that of a training method for a multimedia resource allocation model and a multimedia resource allocation method, the implementation of the above device can refer to the implementation of the method, and repeated details are omitted.

Fig. 8a is a schematic structural diagram of a training apparatus for a multimedia resource allocation model according to an embodiment of the present application. A training device of a multimedia resource allocation model comprises:

a generating unit 811, configured to generate a play status parameter set according to the play amount index and the played amount of the obtained at least one multimedia resource at each play medium at the first time;

a first determining unit 812, configured to determine a corresponding multimedia resource playing adjustment manner according to the multimedia resource allocation model and the playing state parameter set, where the multimedia resource playing adjustment manner is used to adjust the playing amount indexes of at least two playing media;

an estimation unit 813, configured to determine an estimated playing amount after executing the multimedia resource playing adjustment mode according to the multimedia resource allocation model and the playing state parameter set;

an obtaining unit 814, configured to obtain an actual play shortage after the multimedia resource play adjustment mode is executed;

a second determining unit 815, configured to determine a play amount loss according to the estimated play amount and the actual play shortage;

an adjusting unit 816 is configured to perform parameter adjustment on the multimedia resource allocation model according to the play amount loss.

Preferably, the first determining unit 812 is further configured to:

Preferably, the first determining unit 812 is configured to:

Preferably, the second determining unit 815 is configured to:

Preferably, the adjusting unit 816 is further configured to:

obtaining the number of times of model training;

Fig. 8b is a schematic structural diagram of a multimedia resource allocation apparatus according to an embodiment of the present application. A multimedia resource allocation model training and multimedia resource allocation device comprises:

an obtaining unit 821, configured to obtain a playing state parameter set of at least one multimedia resource in multiple playing media;

an obtaining unit 822, configured to obtain a target multimedia resource playing adjustment manner according to the multimedia resource allocation model and the playing state parameter set;

a first adjusting unit 823, configured to adjust a playing amount indicator of at least one multimedia resource in multiple playing media according to a playing adjustment manner of a target multimedia resource;

a second adjusting unit 824, configured to perform multimedia resource allocation adjustment according to the adjusted play amount index.

Fig. 9 shows a schematic configuration of a control device 9000. Referring to fig. 9, a control apparatus 9000 includes: a processor 9010, a memory 9020, a power supply 9030, a display unit 9040, and an input unit 9050.

The processor 9010 is a control center of the control device 9000, connects respective components by various interfaces and lines, and executes various functions of the control device 9000 by running or executing software programs and/or data stored in the memory 9020, thereby monitoring the control device 9000 as a whole.

In the embodiment of the present application, the processor 9010, when calling the computer program stored in the memory 9020, executes the method for training the multimedia resource allocation model and allocating the multimedia resource provided in the embodiment shown in fig. 3 and 6.

Optionally, processor 9010 may include one or more processing units; preferably, the processor 9010 may integrate an application processor, which mainly handles operating systems, user interfaces, applications, etc., and a modem processor, which mainly handles wireless communications. It is to be understood that the modem processor may not be integrated into the processor 9010. In some embodiments, the processor, memory, and/or memory may be implemented on a single chip, or in some embodiments, they may be implemented separately on separate chips.

The memory 9020 may mainly include a program storage area and a data storage area, where the program storage area may store an operating system, various applications, and the like; the storage data area may store data created according to the use of the control apparatus 9000, and the like. Further, the memory 9020 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage device.

The control device 9000 further comprises a power supply 9030 (e.g., a battery) for supplying power to each component, and the power supply may be logically connected to the processor 9010 via a power management system, so as to implement functions of managing charging, discharging, and power consumption via the power management system.

The display unit 9040 can be used to display information input by a user or information provided to the user, and various menus of the control device 9000, and the like, and in the embodiment of the present invention, is mainly used to display a display interface of each application in the control device 9000 and objects such as texts and pictures displayed in the display interface. The display unit 9040 may include a display panel 9041. The Display panel 9041 may be configured to have a Liquid CryStal Display (LCD), an Organic Light-Emitting Diode (OLED), or the like.

The input unit 9050 may be configured to receive information such as numbers or characters input by a user. The input unit 9050 may include a touch panel 9051 and other input devices 9052. Among other things, the touch panel 9051, also referred to as a touch screen, may collect touch operations by a user (e.g., operations by a user on or near the touch panel 9051 using any suitable object or accessory such as a finger, a touch pen, etc.).

Specifically, the touch panel 9051 may detect a touch operation of the user, detect signals generated by the touch operation, convert the signals into touch point coordinates, send the touch point coordinates to the processor 9010, receive a command sent from the processor 9010, and execute the command. In addition, the touch panel 9051 may be implemented by using various types such as a resistive type, a capacitive type, an infrared ray, and a surface acoustic wave. Other input devices 9052 may include, but are not limited to, one or more of a physical keyboard, function keys (such as volume control keys, power on/off keys, etc.), a trackball, a mouse, a joystick, and the like.

Of course, the touch panel 9051 may cover the display panel 9041, and when the touch panel 9051 detects a touch operation on or near the touch panel 9051, the touch panel is transmitted to the processor 9010 to determine the type of the touch event, and then the processor 9010 provides a corresponding visual output on the display panel 9041 according to the type of the touch event. Although in fig. 9 the touch panel 9051 and the display panel 9041 are implemented as two separate components to control the input and output functions of the device 9000, in some embodiments the touch panel 9051 may be integrated with the display panel 9041 to implement the input and output functions of the device 9000.

The control device 9000 may also include one or more sensors, such as a pressure sensor, a gravitational acceleration sensor, a proximity light sensor, and/or the like. Of course, the control device 9000 may also include other components such as a camera, as needed in a particular application, and since these components are not the components used with importance in the embodiments of the present application, they are not shown in fig. 9 and will not be described in detail.

Those skilled in the art will appreciate that fig. 9 is merely an example of a control device and is not intended to be limiting and may include more or fewer components than those shown, or some components may be combined, or different components.

Those of ordinary skill in the art will understand that: all or part of the steps for implementing the method embodiments may be implemented by hardware related to program instructions, and the program may be stored in a computer readable storage medium, and when executed, the program performs the steps including the method embodiments; and the aforementioned storage medium includes: a mobile storage device, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes. Alternatively, the integrated unit of the present invention may be stored in a computer-readable storage medium if it is implemented in the form of a software functional module and sold or used as a separate product. Based on such understanding, the technical solutions of the embodiments of the present invention may be substantially implemented or portions thereof that contribute to the prior art may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the methods described in the embodiments of the present invention. And the aforementioned storage medium includes: a removable storage device, a ROM, a RAM, a magnetic or optical disk, or various other media that can store program code.

Embodiments of the present application also provide a computer program product or computer program comprising computer instructions stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, so that the computer device executes the method for training the multimedia resource allocation model and controlling the multimedia resource allocation in any of the above method embodiments.

Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a general hardware platform, and certainly can also be implemented by hardware. Based on such understanding, the above technical solutions substantially or partially contributing to the related art may be embodied in the form of a software product, which may be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes several instructions for enabling a control device (which may be a personal computer, a server, or a network device, etc.) to execute the methods of the various embodiments or some parts of the embodiments.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions in the embodiments of the present application.

Claims

1. A method for training a multimedia resource allocation model, comprising:

determining a corresponding multimedia resource playing adjustment mode according to a multimedia resource distribution model and the playing state parameter set, wherein the multimedia resource playing adjustment mode is used for adjusting the playing amount indexes of at least two playing media;

according to a multimedia resource distribution model and the play state parameter set, determining the estimated play amount after the multimedia resource play adjustment mode is executed;

acquiring the actual play shortage after the multimedia resource play adjustment mode is executed;

and adjusting parameters of the multimedia resource distribution model according to the play amount loss.

2. The method of claim 1, wherein before determining the corresponding adjustment manner for playing the multimedia resource according to the multimedia resource allocation model and the set of playing state parameters, the method further comprises:

acquiring resource identification information corresponding to the at least one multimedia resource, playing media identification information corresponding to a playing media and playing type identification information corresponding to a playing type of the multimedia resource;

and adding the obtained resource identification vector, the obtained play medium identification vector and the obtained play type identification vector to the play state parameter set.

3. The method of claim 2, wherein performing data feature transformation on the obtained resource identification information, playing media identification information, and playing type identification information to obtain corresponding resource identification vector, playing media identification vector, and playing type identification vector comprises:

4. The method of claim 2, wherein determining the corresponding multimedia asset play adjustment mode according to the multimedia asset allocation model and the play state parameter set comprises:

determining the play shortage of the at least one multimedia resource according to the play amount index and the played amount of the at least one multimedia resource in each play medium;

determining a corresponding second multimedia resource playing adjustment mode when the estimated playing amount is maximum according to the playing amount index, the played amount and the multimedia resource playing type of the at least one multimedia resource in each playing medium respectively;

acquiring a first probability corresponding to the first multimedia resource playing adjustment mode and a second probability corresponding to the second multimedia resource playing adjustment mode;

5. The method as claimed in claim 4, wherein determining the second multimedia asset playing adjustment mode corresponding to the maximum predicted playing amount according to the playing amount indicator, the played amount and the multimedia asset playing type of the at least one multimedia asset in each playing medium respectively comprises:

determining a second multimedia resource playing adjustment mode set corresponding to the first time according to the playing amount index, the played amount and the multimedia resource playing type of the at least one multimedia resource in each playing medium respectively;

6. The method of claim 5, wherein determining the playback volume loss based on the estimated playback volume and the actual playback volume deficit comprises:

acquiring a play state parameter set corresponding to the at least one multimedia resource at a second time, wherein a preset interval duration is arranged between the second time and the first time;

7. The method of any of claims 1-6, wherein after performing parameter adjustments to the multimedia resource allocation model based on the playout amount loss, further comprising:

obtaining the number of times of model training;

judging whether the training times of the model are higher than preset training times or not, and if so, obtaining a trained multimedia resource distribution model;

otherwise, taking the sum of the first time and the preset interval duration as a new first time, and according to the new first time, executing the step of generating the playing state parameter set according to the acquired playing amount index and the played amount of the at least one multimedia resource in each playing medium at the first time.

8. A multimedia resource allocation method, characterized in that a multimedia resource allocation model provided by the method according to any of claims 1-7 is used, comprising:

acquiring a target multimedia resource playing adjustment mode according to the multimedia resource distribution model and the playing state parameter set;

adjusting the playing amount index of the at least one multimedia resource in a plurality of playing media according to the playing adjustment mode of the target multimedia resource;

9. An apparatus for training a multimedia resource allocation model, comprising:

a first determining unit, configured to determine a corresponding multimedia resource playing adjustment manner according to a multimedia resource allocation model and the playing state parameter set, where the multimedia resource playing adjustment manner is used to adjust playing amount indexes of at least two playing media;

the pre-estimation unit is used for determining the pre-estimated playing amount after the multimedia resource playing adjustment mode is executed according to a multimedia resource distribution model and the playing state parameter set;

and the adjusting unit is used for adjusting parameters of the multimedia resource distribution model according to the play amount loss.

10. A multimedia resource allocation apparatus, characterized in that a multimedia resource allocation model provided by the method according to any one of claims 1 to 7 is adopted, comprising:

an obtaining unit, configured to obtain a target multimedia resource playing adjustment mode according to the multimedia resource allocation model and the playing state parameter set;

a first adjusting unit, configured to adjust a play amount indicator of the at least one multimedia resource in multiple play media according to the play adjusting manner of the target multimedia resource;