1 paper · 1 filter
Shijie Wang, Qi Zhao, Minh Quan Do +3
What makes good representations for video understanding, such as anticipating future activities, or answering video-conditioned questions? While earlier approaches focus on end-to-…