CVPR 2026: Deccan AI Lightning Talks - Research Preview
Registration
Approval Required
Your registration is subject to host approval.
About Event
Join the Deccan AI research team at Booth #849 for a series of lightning talks previewing our latest work in spatial and embodied reasoning.
Schedule & Session Topics
June 5, 2:00 PM & June 6, 1:00 PM
Do Vision Language Models Understand 3D Scenes or Just Catalogue Objects? — A new benchmark revealing how vision-language models struggle to simulate hidden, rearranged, and counterfactual 3D scenes.June 6, 3:30 PM & June 7, 1:00 PM
Can Embodied AI Systems Reason Through Actions, Goals, and Safe Rearrangements? — A unified benchmark and SFT dataset for evaluating safety-aware embodied agents on action understanding, long-horizon planning, and spatial rearrangement.
Speakers: Divyansh Sahu and Animesh Maheshwari, Deccan AI Research Team.