EVA: An Embodied World Model for Future Video Anticipation
For ICLR2025 Rebuttal