Systems and methods for predicting video quality based on objectives of video producer

Abstract

Systems, methods, and non-transitory computer-readable media can collect a set of training videos as training data, wherein the set of training videos are labeled with one or more labels based on one or more video quality metrics associated with an evaluation objective. A machine learning model is trained based on the training data. A video to be evaluated is received. The video is assigned to a first video quality category of a plurality of video quality categories based on the machine learning model.

What is claimed is: 
     
         1 . A computer-implemented method comprising:
 training, by a computing system, a machine learning model based on videos labeled with a plurality of video quality categories determined based on video quality metrics;   receiving, by the computing system, a first version of an unpublished video associated with a selected video quality category desired by a publisher for publication of video;   based on the machine learning model, assigning, by the computing system, the first version of the unpublished video to a first video quality category of lesser quality than the selected video quality category;   receiving, by the computing system, a second version of the unpublished video; and   based on the machine learning model, assigning, by the computing system, the second version of the unpublished video to the selected video quality category desired by the publisher for publication of video.   
     
     
         2 . The computer-implemented method of  claim 1 , wherein the video quality metrics pertain to an evaluation objective associated with at least one of viewer retention time, comments, or shares relating to video. 
     
     
         3 . The computer-implemented method of  claim 1 , wherein the first video quality category is associated with a first range of values and the selected video quality category is associated with a second range of values larger than the first range of values. 
     
     
         4 . The computer-implemented method of  claim 1 , wherein the machine learning model is a multi-stage model comprising a deep neural network and a sparse neural network. 
     
     
         5 . The computer-implemented method of  claim 4 , wherein the deep neural network is configured to receive image and sound data associated with a video and generate a vector representation of the video. 
     
     
         6 . The computer-implemented method of  claim 5 , wherein the sparse neural network is configured to receive metadata associated with the video and the vector representation of the video and generate respective likelihood scores corresponding to each video quality category of the plurality of video quality categories. 
     
     
         7 . The computer-implemented method of  claim 6 , wherein the unpublished video is assigned to the selected video quality category based on the selected video quality category having a highest likelihood score of the plurality of video quality categories. 
     
     
         8 . The computer-implemented method of  claim 1 , wherein a plurality of versions of the unpublished video were generated between the first version of the unpublished video and the second version of the unpublished video. 
     
     
         9 . The computer-implemented method of  claim 1 , wherein the training is based on a first set of videos from a first set of pages of a social networking system. 
     
     
         10 . The computer-implemented method of  claim 9 , wherein the training is further based on a second set of pages of the social networking system that are similar
 to the first set of pages, the second set of pages identified based on a second machine learning model.   
     
     
         11 . A system comprising:
 at least one processor; and   a memory storing instructions that, when executed by the at least one processor,   cause the system to perform operations comprising:   training a machine learning model based on videos labeled with a plurality of video quality categories determined based on video quality metrics;   receiving a first version of an unpublished video associated with a selected video quality category desired by a publisher for publication of video;   based on the machine learning model, assigning the first version of the unpublished video to a first video quality category of lesser quality than the selected video quality category;   receiving a second version of the unpublished video; and   based on the machine learning model, assigning the second version of the unpublished video to the selected video quality category desired by the publisher for publication of video.   
     
     
         12 . The system of  claim 11 , wherein the video quality metrics pertain to an evaluation objective associated with at least one of viewer retention time, comments, or shares relating to video. 
     
     
         13 . The system of  claim 11 , wherein the first video quality category is associated with a first range of values and the selected video quality category is associated with a second range of values larger than the first range of values. 
     
     
         14 . The system of  claim 11 , wherein the machine learning model is a multi-stage model comprising a deep neural network and a sparse neural network. 
     
     
         15 . The system of  claim 14 , wherein the deep neural network is configured to receive image and sound data associated with a video and generate a vector representation of the video. 
     
     
         16 . A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to perform operations comprising:
 training a machine learning model based on videos labeled with a plurality of video quality categories determined based on video quality metrics;   receiving a first version of an unpublished video associated with a selected video quality category desired by a publisher for publication of video;   based on the machine learning model, assigning the first version of the unpublished video to a first video quality category of lesser quality than the selected video quality category;   receiving a second version of the unpublished video; and   based on the machine learning model, assigning the second version of the unpublished video to the selected video quality category desired by the publisher for publication of video.   
     
     
         17 . The non-transitory computer-readable storage medium of  claim 16 , wherein the video quality metrics pertain to an evaluation objective associated with at least one of viewer retention time, comments, or shares relating to video. 
     
     
         18 . The non-transitory computer-readable storage medium of  claim 16 , wherein the first video quality category is associated with a first range of values and the selected video quality category is associated with a second range of values larger than the first range of values. 
     
     
         19 . The non-transitory computer-readable storage medium of  claim 16 , wherein the machine learning model is a multi-stage model comprising a deep neural network and a sparse neural network. 
     
     
         20 . The non-transitory computer-readable storage medium of  claim 19 , wherein the deep neural network is configured to receive image and sound data associated with a video and generate a vector representation of the video.