News
Latest research updates, publications, and academic milestones.
2026
-
Our paper Breaking Failure Cascades was accepted to EMNLP 2026 Main! 🎉
October 2026
Our paper Breaking Failure Cascades was accepted to EMNLP 2026 Main, introducing step-aware reinforcement learning to improve medical multimodal reasoning.
-
New preprint! Reason Through the Latent! is now available on arXiv!
September 2026
Released Reason Through the Latent!, exploring how to make latent visual reasoning essential for multimodal predictions.
-
We presented CLAG at ACL 2026 Findings in San Diego, USA! 🎉
July 2026
Presented CLAG: Adaptive Memory Organization via Agent-Driven Clustering for Small Language Model Agents at ACL 2026 Findings in San Diego, USA.
-
New preprint! Do All Visual Tokens Matter Equally? is now available on arXiv!
July 2026
Released Do All Visual Tokens Matter Equally?, introducing object-aware token merging to preserve fine-grained visual evidence in vision-language retrieval.
-
New preprint! BUZZY is now available on arXiv!
March 2026
Released BUZZY, introducing contrastive scoring to mitigate text-induced bias in multimodal multiple-choice question answering.
2025
-
Started my M.S. in Computer Science at Korea University
September 2025
Joined the DMIS Lab at Korea University as an M.S. student, under the supervision of Prof. Jaewoo Kang, to pursue research in multimodal reasoning.