CN114302236B - Video distribution method and system based on federal learning - Google Patents
Video distribution method and system based on federal learning Download PDFInfo
- Publication number
- CN114302236B CN114302236B CN202111578118.9A CN202111578118A CN114302236B CN 114302236 B CN114302236 B CN 114302236B CN 202111578118 A CN202111578118 A CN 202111578118A CN 114302236 B CN114302236 B CN 114302236B
- Authority
- CN
- China
- Prior art keywords
- video
- target
- time
- mobile edge
- profit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000004364 calculation method Methods 0.000 claims abstract description 42
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 40
- 238000012549 training Methods 0.000 claims abstract description 16
- 239000011159 matrix material Substances 0.000 claims abstract description 8
- 238000003491 array Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 3
- 238000012217 deletion Methods 0.000 claims 1
- 230000037430 deletion Effects 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 5
- 239000013598 vector Substances 0.000 description 4
- 238000013459 approach Methods 0.000 description 2
- 230000004931 aggregating effect Effects 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 230000003139 buffering effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
Abstract
The invention provides a video distribution method and a system based on federal learning, wherein the method comprises the steps of calculating the accepted degree of a target video in a mobile edge calculation server in target time; calculating profit generated by the target video in the target time; calculating average distribution profit of the target video in the target time; deleting the target video corresponding to the average distribution profit lower than the target profit; constructing a video interest matrix; acquiring user data; training a convolutional neural network model according to user data; updating convolutional neural network model parameters; uploading the updated convolutional neural network model parameters to a mobile edge computing server; calculating an average profit for each type of video; and deleting the video corresponding to the average profit attenuation degree of the video higher than the target attenuation degree. The invention solves the problems of indiscriminate video pushing to users, low pertinence, low click rate of the users on video application and unsatisfactory recall effect in the prior art.
Description
Technical Field
The invention belongs to the technical field of data processing, and particularly relates to a video distribution method and system based on federal learning.
Background
In the field of video pushing, a network architecture for caching video content in a mobile edge computing server (Mobile Edge Computing, MEC) is proposed that can reduce video content access latency and video retrieval times. Currently there are mainly two content distribution strategies: a passive content distribution strategy based on user demand; another is a policy where the video content provider caches a portion of the video content that the user is about to request or most desirous to view in advance in the cache of the mobile edge server before the user initiates a request for the video content.
At present, an indiscriminate video pushing mode is adopted for all users, so that videos are pushed to all users, and targeted pushing is difficult to achieve. The same video is pushed to all users by the same push approach, as in the same time period. However, in the prior art, video is pushed to a user indiscriminately, the pertinence is low, the click rate of the user on video application is low, and the recall effect is not ideal.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a video distribution method and system based on federal learning.
In a first aspect, the present invention provides a video distribution method based on federal learning, including:
acquiring the number of users of the target video watched in the mobile edge computing server in the target time;
Acquiring the payment times generated by a target video in a target time;
Calculating the acceptance degree of the target video in the mobile edge calculation server according to the number of users watched in the mobile edge calculation server and the generated payment times of the target video in the target time;
Calculating profit generated by the target video in the target time according to the accepted degree of the target video in the mobile edge calculation server in the target time;
Acquiring the times of caching the target video to a mobile edge computing server in the target time;
Calculating average distribution profit of the target video in the target time;
Deleting the target video corresponding to the average distribution profit lower than the target profit;
constructing a video interest matrix; the number of the video interest moment arrays is the number of video types, and the number of the lines is the number of users;
initializing a convolutional neural network model and parameters of the convolutional neural network model;
acquiring user data, wherein the user data comprises a watching video tag, a tag type, a payment mode, video watching entrance weight and user basic information;
Training a convolutional neural network model according to user data;
Updating convolutional neural network model parameters;
Uploading the updated convolutional neural network model parameters to a mobile edge computing server;
Distributing different types of videos in proportion, and calculating average profit of each type of video;
And deleting the video corresponding to the average profit attenuation degree of the video higher than the target attenuation degree.
Further, the calculating the acceptance degree of the target video in the target time according to the number of users watched in the mobile edge calculating server and the generated payment number, includes:
the extent to which the target video is accepted in the mobile edge computing server within the target time is calculated according to the following formula:
wherein, Calculating the accepted degree of the video j of the i type in the mobile edge server for the time t; /(I)Video j of type i for time t is watched by V users in the mobile edge computing server; /(I)Generating a B-time payment for video j of type i for time t; alpha and beta are respectively/>And/>Weight parameters of (c).
Further, the calculating profit generated by the target video in the target time according to the accepted degree of the target video in the mobile edge calculating server in the target time comprises the following steps:
Calculating profit generated by the target video in the target time according to the following formula:
wherein, Profit generated for video j of type i in time t in total; epsilon is the basic profit coefficient generated by each user for video j of type i in time t; /(I)The extent to which video j of type i is accepted in the mobile edge computing server for time t.
Further, deleting the video with the average profit attenuation degree higher than the target attenuation degree, includes:
the target attenuation is calculated according to the following formula:
Wherein I (t) is the target attenuation over time t; i (t 0) is the attenuation degree at the time t 0; e is a natural constant; gamma is the interest attenuation coefficient of the user to the video; l is the amount of time lapse.
In a second aspect, the present invention provides a federal learning-based video distribution system, comprising:
The first acquisition module is used for acquiring the number of users of which the target video is watched in the mobile edge computing server in the target time;
The second acquisition module is used for acquiring the payment times generated by the target video in the target time;
the first calculation module is used for calculating the accepted degree of the target video in the mobile edge calculation server according to the number of users watched in the mobile edge calculation server and the generated payment times of the target video in the target time;
the second calculation module is used for calculating profits generated by the target video in the target time according to the accepted degree of the target video in the mobile edge calculation server in the target time;
the third acquisition module is used for acquiring the times of caching the target video to the mobile edge computing server in the target time;
the third calculation module is used for calculating the average distribution profit of the target video in the target time;
The first deleting module is used for deleting the target video corresponding to the average distribution profit lower than the target profit;
The construction module is used for constructing a video interest matrix; the number of the video interest moment arrays is the number of video types, and the number of the lines is the number of users;
The initialization module is used for initializing the convolutional neural network model and parameters of the convolutional neural network model;
The fourth acquisition module is used for acquiring user data, wherein the user data comprises a watching video tag, a tag type, a payment mode, a video watching entrance weight and user basic information;
the training module is used for training the convolutional neural network model according to the user data;
the updating module is used for updating the convolutional neural network model parameters;
The uploading module is used for uploading the updated convolutional neural network model parameters to the mobile edge computing server;
The fourth calculation module is used for distributing different types of videos in proportion and calculating the average profit of each type of video;
and the second deleting module is used for deleting the video with the average profit attenuation degree higher than the target attenuation degree.
Further, the first computing module includes:
A first calculation unit for calculating the degree to which the target video is accepted in the mobile edge calculation server within the target time according to the following formula:
wherein, Calculating the accepted degree of the video j of the i type in the mobile edge server for the time t; /(I)Video j of type i for time t is watched by V users in the mobile edge computing server; /(I)Generating a B-time payment for video j of type i for time t; alpha and beta are respectively/>And/>Weight parameters of (c).
Further, the second computing module includes:
A second calculation unit for calculating profit generated by the target video in the target time according to the following formula:
wherein, Profit generated for video j of type i in time t in total; epsilon is the basic profit coefficient generated by each user for video j of type i in time t; /(I)The extent to which video j of type i is accepted in the mobile edge computing server for time t.
Further, the second deleting module includes:
A third calculation unit for calculating a target attenuation degree according to the following formula:
Wherein I (t) is the target attenuation over time t; i (t 0) is the attenuation degree at the time t 0; e is a natural constant; gamma is the interest attenuation coefficient of the user to the video; l is the amount of time lapse.
The invention provides a video distribution method and a system based on federal learning, wherein the method comprises the steps of obtaining the number of users watched by a target video in a mobile edge computing server in target time; acquiring the payment times generated by a target video in a target time; calculating the accepted degree of the target video in the mobile edge calculation server in the target time; calculating profit generated by the target video in the target time according to the accepted degree of the target video in the mobile edge calculation server in the target time; acquiring the times of caching the target video to a mobile edge computing server in the target time; calculating average distribution profit of the target video in the target time; deleting the target video corresponding to the average distribution profit lower than the target profit; constructing a video interest matrix; the number of the video interest moment arrays is the number of video types, and the number of the lines is the number of users; initializing a convolutional neural network model and parameters of the convolutional neural network model; acquiring user data, wherein the user data comprises a watching video tag, a tag type, a payment mode, video watching entrance weight and user basic information; training a convolutional neural network model according to user data; updating convolutional neural network model parameters; uploading the updated convolutional neural network model parameters to a mobile edge computing server; distributing different types of videos in proportion, and calculating average profit of each type of video; and deleting the video corresponding to the average profit attenuation degree of the video higher than the target attenuation degree. By adopting the scheme, the invention solves the problems of indiscriminate video pushing to users, low pertinence, low click rate of the users on video application and unsatisfactory recall effect in the prior art.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings that are needed in the embodiments will be briefly described below, and it will be obvious to those skilled in the art that other drawings can be obtained from these drawings without inventive effort.
FIG. 1 is a workflow diagram of a federal learning-based video distribution method according to an embodiment of the present invention;
Fig. 2 is a schematic structural diagram of a video distribution method based on federal learning according to an embodiment of the present invention;
Fig. 3 is a schematic structural diagram of a video distribution system based on federal learning according to an embodiment of the present invention;
fig. 4 is an application scene diagram of a video distribution method based on federal learning according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
As shown in fig. 4, the present invention selects a high configuration of a certain area, a 4K version of IPTV set-top box user group with strong computing capability, and selects a content distribution server of the area as an aggregation server in the federal learning framework, and uses IPTV video content as a material library.
As described in the background art, in the prior art, an indiscriminate video pushing manner is almost adopted for all users, so that video is pushed to each user, and targeted pushing is difficult to achieve. The same video is pushed to all users by the same push approach, as in the same time period. However, in the prior art, video is pushed to a user indiscriminately, the pertinence is low, the click rate of the user on video application is low, and the recall effect is not ideal.
Therefore, in order to solve the above-mentioned problems, an embodiment of the present invention provides a video distribution method based on federal learning, as shown in fig. 2, fig. 2 is a schematic structural diagram of the video distribution method based on federal learning.
Specifically, as shown in fig. 1, the video distribution method includes:
step S101, obtaining the number of users of the target video watched in the mobile edge computing server in the target time;
step S102, obtaining the payment times generated by a target video in a target time;
Step S103, calculating the accepted degree of the target video in the mobile edge computing server according to the number of users watched in the mobile edge computing server and the generated payment times of the target video in the target time;
in this step, the degree to which the target video is accepted in the mobile edge computing server in the target time is calculated according to the following formula:
wherein, Calculating the accepted degree of the video j of the i type in the mobile edge server for the time t; /(I)Video j of type i for time t is watched by V users in the mobile edge computing server; /(I)Generating a B-time payment for video j of type i for time t; alpha and beta are respectively/>And/>Weight parameters of (c).
Step S104, calculating profits generated by the target video in the target time according to the accepted degree of the target video in the mobile edge calculation server in the target time;
in this step, profit generated by the target video in the target time is calculated according to the following formula:
wherein, Profit generated for video j of type i in time t in total; epsilon is the basic profit coefficient generated by each user for video j of type i in time t; /(I)The extent to which video j of type i is accepted in the mobile edge computing server for time t.
Step S105, obtaining the times of caching the target video in the moving edge computing server in the target time;
Step S106, calculating average distribution profits of the target videos in the target time;
in this step, the larger the average distribution profit, the greater the user's interest in the target video, the more the target video needs to be retained, and otherwise, needs to be deleted.
Step S107, deleting the target video corresponding to the average distribution profit lower than the target profit;
Step S108, constructing a video interest matrix; the number of the video interest moment arrays is the number of video types, and the number of the lines is the number of users;
Step S109, initializing a convolutional neural network model and parameters of the convolutional neural network model;
Step S1010, obtaining user data, wherein the user data comprises a watching video tag, a tag type, a payment mode, a video watching entrance weight and user basic information;
step S1011, training a convolutional neural network model according to user data;
step S1012, updating convolutional neural network model parameters;
Step S1013, uploading the updated convolutional neural network model parameters to a mobile edge calculation server;
step S1014, distributing different types of videos in proportion, and calculating average profit of each type of video;
Step S1015, deleting the video corresponding to the average profit attenuation degree of the video being higher than the target attenuation degree.
In step S1010-S1015, a plurality of set top boxes are randomly selected to form a set Z train of all set top boxes participating in training, each set top box is traversed and the following operations are performed:
downloading model parameters from a mobile edge computing server for each set-top box involved in training Training convolutional neural network model more parameters/>, using current set-top box user dataUploading new model parameters/>Uploading the size |Data z | of each set-top box training Data to the mobile edge computing server:
The mobile edge computing server receives the model parameters, updates the overall training convolutional neural network model according to the data volume proportion,
And adding user interest labels and content labels in the newly introduced video of each set top box device, and adding the newly added content labels according to the prediction model.
And uploading the interest vector of each set top box to a mobile edge computing server.
Aggregating the interestingness vectors and the real video vectors uploaded by all the set top boxes to obtain user interest and income prediction results after federal learning;
distribution of content: and distributing different types of content according to different proportions according to the predicted user benefits and interest vectors in the steps, and calculating the average benefits of each video.
With the increase of time, deleting the video content with the highest watching times and faster profit attenuation from the cache, and if the newly added video j is not cached, obtaining a content tag ranking table { j 1,j2,...,jn } according to the overall profit ranking from high to low.
The mobile edge computing server memory space is |C| max, the video size |C j |, and the mobile edge computing server memory space is preceded according to the content tag ranking table { j 1,j2,...,jn }And adding the video into a cache and sequentially issuing.
As shown in fig. 3, the embodiment of the present invention further provides a video distribution system based on federal learning, including:
a first obtaining module 10, configured to obtain the number of users in the mobile edge computing server, the number of users having watched the target video in the target time;
A second obtaining module 20, configured to obtain the number of payouts generated by the target video in the target time;
A first calculating module 30, configured to calculate, according to the number of users that the target video is watched in the mobile edge calculating server in the target time and the generated payment number, the degree to which the target video is accepted in the mobile edge calculating server in the target time;
A second calculation module 40 for calculating profit generated by the target video in the target time according to the accepted degree of the target video in the mobile edge calculation server in the target time;
a third obtaining module 50, configured to obtain the number of times of buffering the target video in the mobile edge computing server in the target time;
A third calculation module 60 for calculating a target video average distribution profit for a target time;
A first deleting module 70, configured to delete a target video corresponding to an average distribution profit lower than a target profit;
A construction module 80, configured to construct a video interest matrix; the number of the video interest moment arrays is the number of video types, and the number of the lines is the number of users;
An initialization module 90 for initializing the convolutional neural network model and the convolutional neural network model parameters;
a fourth obtaining module 100, configured to obtain user data, where the user data includes a viewing video tag, a tag type, a payment method, a video viewing entry weight, and user basic information;
A training module 110 for training the convolutional neural network model based on the user data;
an updating module 120, configured to update the convolutional neural network model parameters;
The uploading module 130 is configured to upload the updated convolutional neural network model parameters to the mobile edge computing server;
a fourth calculation module 140 for distributing different types of videos in proportion, calculating an average profit for each type of video;
the second deleting module 150 is configured to delete the video whose average profit attenuation degree is higher than the target attenuation degree.
Optionally, the first computing module includes:
A first calculation unit for calculating the degree to which the target video is accepted in the mobile edge calculation server within the target time according to the following formula:
wherein, Calculating the accepted degree of the video j of the i type in the mobile edge server for the time t; /(I)Video j of type i for time t is watched by V users in the mobile edge computing server; /(I)Generating a B-time payment for video j of type i for time t; alpha and beta are respectively/>And/>Weight parameters of (c).
Optionally, the second computing module includes:
A second calculation unit for calculating profit generated by the target video in the target time according to the following formula:
wherein, Profit generated for video j of type i in time t in total; epsilon is the basic profit coefficient generated by each user for video j of type i in time t; /(I)The extent to which video j of type i is accepted in the mobile edge computing server for time t.
Optionally, the second deleting module includes:
A third calculation unit for calculating a target attenuation degree according to the following formula:
Wherein I (t) is the target attenuation over time t; i (t 0) is the attenuation degree at the time t 0; e is a natural constant; gamma is the interest attenuation coefficient of the user to the video; l is the amount of time lapse.
It will be apparent to those skilled in the art that the techniques of embodiments of the present invention may be implemented in software plus a necessary general purpose hardware platform. Based on such understanding, the technical solutions in the embodiments of the present invention may be embodied in essence or what contributes to the prior art in the form of a software product, which may be stored in a storage medium, such as a ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the embodiments or some parts of the embodiments of the present invention.
The same or similar parts between the various embodiments in this specification are referred to each other. In particular, for system embodiments, since they are substantially similar to method embodiments, the description is relatively simple, as far as reference is made to the description in the method embodiments.
The invention has been described in detail in connection with the specific embodiments and exemplary examples thereof, but such description is not to be construed as limiting the invention. It will be understood by those skilled in the art that various equivalent substitutions, modifications or improvements may be made to the technical solution of the present invention and its embodiments without departing from the spirit and scope of the present invention, and these fall within the scope of the present invention. The scope of the invention is defined by the appended claims.
Claims (8)
1. A federal learning-based video distribution method, comprising:
acquiring the number of users of the target video watched in the mobile edge computing server in the target time;
Acquiring the payment times generated by a target video in a target time;
Calculating the acceptance degree of the target video in the mobile edge calculation server according to the number of users watched in the mobile edge calculation server and the generated payment times of the target video in the target time;
Calculating profit generated by the target video in the target time according to the accepted degree of the target video in the mobile edge calculation server in the target time;
Acquiring the times of caching the target video to a mobile edge computing server in the target time;
Calculating average distribution profit of the target video in the target time;
Deleting the target video corresponding to the average distribution profit lower than the target profit;
constructing a video interest matrix; the number of the video interest moment arrays is the number of video types, and the number of the lines is the number of users;
initializing a convolutional neural network model and parameters of the convolutional neural network model;
acquiring user data, wherein the user data comprises a watching video tag, a tag type, a payment mode, video watching entrance weight and user basic information;
Training a convolutional neural network model according to user data;
Updating convolutional neural network model parameters;
Uploading the updated convolutional neural network model parameters to a mobile edge computing server;
Distributing different types of videos in proportion, and calculating average profit of each type of video;
And deleting the video corresponding to the average profit attenuation degree of the video higher than the target attenuation degree.
2. The federal learning-based video distribution method according to claim 1, wherein the calculating the degree to which the target video is accepted in the mobile edge computing server in the target time based on the number of users the target video has been watched in the mobile edge computing server in the target time and the number of payments generated, comprises:
the extent to which the target video is accepted in the mobile edge computing server within the target time is calculated according to the following formula:
wherein, Calculating the accepted degree of the video j of the i type in the mobile edge server for the time t; /(I)Video j of type i for time t is watched by V users in the mobile edge computing server; /(I)Generating a B-time payment for video j of type i for time t; alpha and beta are respectively/>And/>Weight parameters of (c).
3. The federal learning-based video distribution method according to claim 2, wherein calculating profit generated by the target video in the target time according to the degree to which the target video is accepted in the mobile edge calculation server in the target time comprises:
Calculating profit generated by the target video in the target time according to the following formula:
wherein, Profit generated for video j of type i in time t in total; epsilon is the basic profit coefficient generated by each user for video j of type i in time t; /(I)The extent to which video j of type i is accepted in the mobile edge computing server for time t.
4. The federal learning-based video distribution method according to claim 1, wherein deleting the video corresponding to the average profit attenuation degree higher than the target attenuation degree comprises:
the target attenuation is calculated according to the following formula:
Wherein I (t) is the target attenuation over time t; i (t 0) is the attenuation degree at the time t 0; e is a natural constant; gamma is the interest attenuation coefficient of the user to the video; l is the amount of time lapse.
5. A federal learning-based video distribution system, comprising:
The first acquisition module is used for acquiring the number of users of which the target video is watched in the mobile edge computing server in the target time;
The second acquisition module is used for acquiring the payment times generated by the target video in the target time;
the first calculation module is used for calculating the accepted degree of the target video in the mobile edge calculation server according to the number of users watched in the mobile edge calculation server and the generated payment times of the target video in the target time;
the second calculation module is used for calculating profits generated by the target video in the target time according to the accepted degree of the target video in the mobile edge calculation server in the target time;
the third acquisition module is used for acquiring the times of caching the target video to the mobile edge computing server in the target time;
the third calculation module is used for calculating the average distribution profit of the target video in the target time;
The first deleting module is used for deleting the target video corresponding to the average distribution profit lower than the target profit;
The construction module is used for constructing a video interest matrix; the number of the video interest moment arrays is the number of video types, and the number of the lines is the number of users;
The initialization module is used for initializing the convolutional neural network model and parameters of the convolutional neural network model;
The fourth acquisition module is used for acquiring user data, wherein the user data comprises a watching video tag, a tag type, a payment mode, a video watching entrance weight and user basic information;
the training module is used for training the convolutional neural network model according to the user data;
the updating module is used for updating the convolutional neural network model parameters;
The uploading module is used for uploading the updated convolutional neural network model parameters to the mobile edge computing server;
The fourth calculation module is used for distributing different types of videos in proportion and calculating the average profit of each type of video;
and the second deleting module is used for deleting the video with the average profit attenuation degree higher than the target attenuation degree.
6. The federal learning-based video distribution system according to claim 5, wherein the first computing module comprises:
A first calculation unit for calculating the degree to which the target video is accepted in the mobile edge calculation server within the target time according to the following formula:
wherein, Calculating the accepted degree of the video j of the i type in the mobile edge server for the time t; /(I)Video j of type i for time t is watched by V users in the mobile edge computing server; /(I)Generating a B-time payment for video j of type i for time t; alpha and beta are respectively/>And/>Weight parameters of (c).
7. The federal learning-based video distribution system according to claim 6, wherein the second computing module comprises:
A second calculation unit for calculating profit generated by the target video in the target time according to the following formula:
wherein, Profit generated for video j of type i in time t in total; epsilon is the basic profit coefficient generated by each user for video j of type i in time t; /(I)The extent to which video j of type i is accepted in the mobile edge computing server for time t.
8. The federal learning-based video distribution system according to claim 5, wherein the second deletion module comprises:
A third calculation unit for calculating a target attenuation degree according to the following formula:
Wherein I (t) is the target attenuation over time t; i (t 0) is the attenuation degree at the time t 0; e is a natural constant;
gamma is the interest attenuation coefficient of the user to the video; l is the amount of time lapse.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111578118.9A CN114302236B (en) | 2021-12-22 | 2021-12-22 | Video distribution method and system based on federal learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111578118.9A CN114302236B (en) | 2021-12-22 | 2021-12-22 | Video distribution method and system based on federal learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114302236A CN114302236A (en) | 2022-04-08 |
| CN114302236B true CN114302236B (en) | 2024-04-26 |
Family
ID=80968742
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111578118.9A Active CN114302236B (en) | 2021-12-22 | 2021-12-22 | Video distribution method and system based on federal learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114302236B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104284211A (en) * | 2013-07-05 | 2015-01-14 | 上海帕科软件科技有限公司 | Centralized video distributing and charging system |
| CN113315978A (en) * | 2021-05-13 | 2021-08-27 | 江南大学 | Collaborative online video edge caching method based on federal learning |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11533380B2 (en) * | 2018-06-25 | 2022-12-20 | Nokia Technologies Oy | Methods and apparatuses for internet content pushing service in a wireless mobile communication system |
-
2021
- 2021-12-22 CN CN202111578118.9A patent/CN114302236B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104284211A (en) * | 2013-07-05 | 2015-01-14 | 上海帕科软件科技有限公司 | Centralized video distributing and charging system |
| CN113315978A (en) * | 2021-05-13 | 2021-08-27 | 江南大学 | Collaborative online video edge caching method based on federal learning |
Non-Patent Citations (1)
| Title |
|---|
| 基于联邦学习的在线短视频内容分发策略;董文涛等;计算机应用;20210622;第41卷(第6期);第1551-1556页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114302236A (en) | 2022-04-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11303970B2 (en) | Systems and methods for internet video delivery | |
| US10397359B2 (en) | Streaming media cache for media streaming service | |
| US20090063496A1 (en) | Automated most popular media asset creation | |
| US8849095B2 (en) | Goal-based video delivery system | |
| JP7443486B2 (en) | Content item recommendations | |
| US8560293B2 (en) | Enhanced matching through explore/exploit schemes | |
| CN103974084B (en) | Streaming media data block caching method, file recommendation method and streaming media server | |
| US9498727B2 (en) | Pre-fetching items in a virtual universe based on avatar communications | |
| CN106649349A (en) | Method, device and system for data caching, applicable to game application | |
| CN105210352A (en) | Fingerprint-based, intelligent, content pre-fetching | |
| US20130325942A1 (en) | Recommender system for content delivery networks | |
| CN101141623A (en) | P2P technique based video on-demand program caching method | |
| WO2018011684A1 (en) | Method and system for recommending dynamic, adaptive and non- sequentially assembled videos | |
| KR20180053673A (en) | Clustering of viewing time for video search | |
| CN103150245B (en) | Determine method and the storage controller of the access characteristics of data entity | |
| CN106528851A (en) | Intelligent recommendation method and device | |
| CN111292001A (en) | Joint decision method and device based on reinforcement learning | |
| CN106649313A (en) | Method and equipment for processing cache data | |
| CN114302236B (en) | Video distribution method and system based on federal learning | |
| CN110083761B (en) | Data distribution method, system and storage medium based on content popularity | |
| JP4369525B1 (en) | Advertisement delivery control method and advertisement delivery control apparatus | |
| AT&T | CMMI10 | |
| Lim et al. | Neural Cloud Storage: Innovative Cloud Storage Solution for Cold Video | |
| CN113630632B (en) | Data processing method and device | |
| CN117633041A (en) | Information caching method, system, equipment and storage medium for road edge unit |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 17F, 501 Zhongshan South Road, Qinhuai District, Nanjing City, Jiangsu Province, 210006 Applicant after: Jiangsu Haobai Technology Co.,Ltd. Address before: 17F, 501 Zhongshan South Road, Qinhuai District, Nanjing City, Jiangsu Province, 210006 Applicant before: JIANGSU BEST TONE INFORMATION SERVICE CO.,LTD. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |