Interactive and Balanced Multimodal Learning via Cross Attention and Gradient Modulation for Compressed Video Action Recognition.

Xinqi Li, Shaojie Li, Ming Ma

Browse the full ICASSP paper archive.