VITA-E: Natural Embodied Interaction with Concurrent Seeing, Hearing, Speaking, and Acting

Xiaoyu Liu, Chaoyou Fu, Chi Yan +15

Current Vision-Language-Action (VLA) models are often constrained by a rigid, static interaction paradigm, which lacks the ability to see, hear, speak, and act concurrently as well…