STELLA: Continual Audio-Video Pre-training with SpatioTemporal Localized Alignment.

Jaewoo Lee, Jaehong Yoon, Wonjae Kim, Yunji Kim, Sung Ju Hwang

Browse the full ICML paper archive.