Google has published the YouTube-8M Segments dataset, a collection of 237,000 five-second video segments annotated with time-localized labels to help train AI systems to predict video content. Google had previously published the full YouTube-8M dataset of videos which led to advancements in video classification algorithms, however the video clips lacked enough annotations for an AI system to predict what would happen next in a video. The Segments dataset consists of a portion of the full videos from the full YouTube-8M dataset with the addition of human-created labels indicating a video’s content at five-second intervals to enable AI systems to better understand and predict video sequences.
4.7K
Joshua New
Joshua New was a senior policy analyst at the Center for Data Innovation. He has a background in government affairs, policy, and communication. Prior to joining the Center for Data Innovation, Joshua graduated from American University with degrees in C.L.E.G. (Communication, Legal Institutions, Economics, and Government) and Public Communication. His research focuses on methods of promoting innovative and emerging technologies as a means of improving the economy and quality of life.
previous post
