1 paper
Jianxin Liang, Tan Yue, Yuxuan Wang +4
The performance of Video Question Answering (VideoQA) models is fundamentally constrained by the nature of their supervision, which typically consists of isolated, factual question…