1 paper · 1 filter
Weilin Ruan, Bohan Zhou, Jinghao Li +12
Embodied spatial intelligence requires more than visual recognition: models must localize entities, reason about their relations, and express this knowledge in action-relevant form…