Full Professor, Seoul National University
2 papers at NeurIPS 2025
We propose a novel RL-based MLLM post-training framework named RePIC for the personalized image captioning task. Our method significantly outperforms SFT-based baselines on multi-concept personalized image captioning.
We introduce the Generalized Induction Model (GIM), a retrieval-based in-context module that enhances interpretable next-token prediction in language modeling and fMRI response prediction.
