EVA
Dataset Analysis
Proposes EVA, an embodied world model using vision-language and video generation models for multi-step video prediction and OOD handling.
Provenance
Collected from papers.
Derived from paper: EVA: An Embodied World Model for Future Video Anticipation