News
2026
- AGA Lab participates in KIAS Center for AI and Natural Sciences Workshop. Sangwoong gives an invited talk.
- Sangwoong gives a seminar at the Department of Electrical Engineering and Computer Science, DGIST, hosted by Professor Chang-Eop Kim.
- GDSD: Reinforcement Learning as Guided Denoiser Self-Distillation for Diffusion Language Models is accepted at NeurIPS 2026
- Rethinking Knowledge Distillation for Diffusion Language Models is accepted at NeurIPS 2026
- Sangwoong attends ICML 2026 in Seoul, Republic of Korea
- Sangwoong co-hosts Post-ICML@UNIST workshop.
- Sangwoong gives a talk at 2026 KSIAM Spring Conference at UNIST
- Multi-Task GRPO: Reliable LLM Reasoning Across Tasks is accepted at ICML 2026
- RSPO: Regularized Self-Play Alignment of Large Language Models is accepted at ICML 2026
- Awarded funding for supporting post-doctoral researchers. AGA Lab participates in the InnoCORE program as a member of the BIO-MAX research group
- Sangwoong attends ICLR 2026 in Rio de Janeiro
- Sangwoong gives a talk at Korea Institute for Advanced Study (KIAS)
- Awarded a 50,000 GPU-hour compute grant from the Swiss AI Initiative as part of a collaboration with University College London and the University of Basel
- Sangwoong gives a talk, Generative Modeling and Reinforcement Learning: Energy, Reward, and Value, at UNIST AIGS Seminar
- Sangwoong gives a tutorial session at 2026 AI Winter School (2026 인공지능 동계단기강좌) hosted by Korea Artificial Intelligence Association
- wd1: Weighted Policy Optimization for Reasoning in Diffusion Language Models is accepted at ICLR 2026
- Robust Multi-Objective Controlled Decoding of Large Language Models is accepted at ICLR 2026
- Value Gradient Sampler: Learning Invariant Value Functions for Equivariant Diffusion Sampling is accepted at AISTATS 2026