1 paper
Xilin Jiang, Qiaolin Wang, Junkai Wu +30
Internet audio-visual clips convey meaning through time-varying sound and motion, which extend beyond what text alone can represent. To examine whether AI models can understand suc…