北京通用人工智能研究院BIGAI

科研成果

Ag2Manip: Learning Novel Manipulation Skills with Agent-Agnostic Visual and Action Representations

SlotLifter: Slot-guided Feature Lifting for Learning Object-centric Radiance Fields

F-HOI: Toward Fine-grained Semantic-Aligned 3D Human-Object Interactions

SceneVerse: Scaling 3D Vision-Language Learning for Grounded Scene Understanding

Unifying 3D Vision-Language Understanding via Promptable Queries

An Embodied Generalist Agent in 3D World