In Seoul for #ICML2026 and excited to catch up with friends. With collaborators, presenting a few papers on knowledge-intensive, scientific, and mathematical reasoning, and also organizing a workshop AI4Physic. Details👇 Happy to chat, esp about agent post-training & adaptation, and efficient evals.
No Digg Deeper questions have been answered for this story yet.
📌 Demystifying Scientific Problem-Solving in LLMs by Probing Knowledge and Reasoning (Wed 14:30)
📌 VideoKR: Towards Knowledge- and Reasoning-Intensive Video Understanding (Tue 14:00)
📌 QEDBench: Quantifying the Alignment Gap in Automated Evaluation of University-Level Mathematical Proofs (Wed 10:30)
✨ AI4Physics Workshop (Sat, 11th)
In Seoul for #ICML2026 and excited to catch up with friends. With collaborators, presenting a few papers on knowledge-intensive, scientific, and mathematical reasoning, and also organizing a workshop AI4Physic. Details👇 Happy to chat, esp about agent post-training & adaptation, and efficient evals.