HiMaCon: Discovering Hierarchical Manipulation Concepts from Unlabeled Multi-Modal Data

NeuIPS2025 / Vision-Language-Action Model