Fetching the paper…

Attend What You Need: Motion-Appearance Synergistic Networks for Video Question Answering · Around