End-to-endMulti-modalVideoTemporalGrounding

To integrate the three modalities more effectively and enable inter-modal learning, we design a dynamic fusion scheme with transformers to model the interactions between modalities.