Yunhun Nam

About

I am an M.S. student at the Trustworthy AI Lab, advised by Prof. Jongheon Jeong at Korea University.
My research interests focus on AI safety for generative models, including large language models (LLMs), multimodal large language models (MLLMs), and diffusion models, with particular emphasis on alignment and robustness against adversarial and out-of-distribution challenges.
Ultimately, my goal is to build AI solutions that people can genuinely trust and safely rely on in the real world.

Publications

(*: Equal contribution, C: Conferences, J: Journal, W: Workshop, P: Preprint).

[P2] Robust Finetuning from Generative Representations via Nuisance-extended Information Bottleneck

Minsun Lee, Gwangyeol Yu, Yunhun Nam, Jinwoo Shin, Jongheon Jeong

Paper

[P1] Vision-driven Preference Synthesis for Mitigating Hallucinations in VLMs

Yunhun Nam, Jongheon Jeong

Arxiv Code

[C2] Learning from the Undesirable: Robust Adaptation of Language Models without Forgetting

Yunhun Nam, Jaehyung Kim, Jongheon Jeong

AAAI, 2026 (Oral Presentation)

Arxiv Code

[C1] BlurGuard: A Simple Approach for Robustifying Image Protection Against AI-Powered Editing

Jinsu Kim*, Yunhun Nam*, Minseon Kim, Sangpil Kim, Jongheon Jeong

NeurIPS, 2025

Paper Arxiv Code

[J1] SBDD-Net: Cross-Source Robust 3D Structural Block and Double Descriptor for Point Place Recognition

Yunhun Nam, Soomok Lee

IEEE Access, vol. 12, pp. 147310–147319, 2024

Paper

Education

Korea University Mar. 2025 - Current

M.S. Student in Graduate School of Artificial Intelligence

Ajou University Mar. 2018 - Feb. 2025

B.S. in Mechanical Engineering & Computer Science