How well can AI research agents explore new scientific ideas?
The latest paper from our lab, Heuresis, combines coding agents with search and quality-diversity algorithms. Algorithms that provide better exploration are the key to unlock new ideas.
Check out this thread👇
Key to realizing Auto Research Agents that can make novel discoveries in AI is understanding and improving their exploration capabilities.
To this end, we built Heuresis, a composable framework that combines coding agents with arbitrary search algorithms within a flexible loop.
Several of our students will be presenting at @icmlconf in Seoul this week! 🇰🇷 Work spanning RL, agents, automated discovery, spatial intelligence & robotics. Come say hi by the posters 👋
Congratulations to @_Chuhan_Li , @xwang_lk, and their collaborators for their SAW-Bench paper on receiving the Best Paper Award Runner-Up from the CVPR 2026 WMAS workshop! 🏆
Human perception is inherently situated – we understand the world relative to our own body, viewpoint, and motion.
To deploy multimodal foundation models in embodied settings, we ask:
“Can these models reason in the same observer-centric way?”
We study this through SAW-Bench:
