Intro

📌 I'm currently on the 2025 job market, looking for a tenure-track faculty position in several countries. If you'd like to connect, please feel free to reach out!

I'm a Postdoctoral Research Associate in the MURGe-Lab (UNC-NLP Group) at the University of North Carolina, Chapel Hill, working with Prof. Mohit Bansal. Prior to joining UNC, I received my Ph.D. degree from Korea Advanced Institute of Science and Technology (KAIST) in 2023, advised by Prof. Sung Ju Hwang, and won the best Ph.D. Dissertation Award from the KAIST College of Engineering and School of Computing. During my Ph.D. study, I interned at Microsoft Research, mentored by Dr. Yue Cao, and worked as a short-term visiting student at the Weizmann Institute of Science, hosted by Prof. Yonina Eldar. I received my B.S. and M.S. degrees from Ulsan National Institute of Science and Technology (UNIST).

For my long-term research goal, I've been dedicated to developing lifelong adaptability in embodied AI systems and multimodal foundation models (e.g., VLM, MLLM, diffusion models) within dynamic real-world environments. These systems aim to be reliable, improvable, interactive, and compositional, addressing practical, real-world challenges to better understand human behavior and significantly impact our daily lives.

My recent research interest includes the following topics:

• LURE: Analyzing and Mitigating Object Hallucination in MLLMs (ICLR 2024)
• SELMA: Learning/Merging Skill-Specific T2I Experts with Auto-Generated Data (NeurIPS 2024)
• EnvGen: Generating and Adapting Environments via LLMs for Training Embodied Agents (COLM 2024)
• SAFREE: Training-Free and Adaptive Guard for Safe Text-to-Image And Video Generation (arXiv 24.10)

• Mementos: Comprehensive Benchmark for MLLM Reasoning over Image Sequences (ACL 2024)
• CREMA: Generalizable & Efficient Video-Language Reasoning via Any-Modal Modular Fusion (arXiv 24.02)
• VideoTree: Query-Adaptive Tree-based Long-Video Understanding (arXiv 24.05)
• RACCooN: Interactive Video Content Editing with Auto-Generated Descriptions & Planning (arXiv 24.05)

• OCS: Online Coreset Selection for Imbanalced and Noisy Continual Learning (ICLR 2022)
• GLAD: On the Study of Incremental Model Generalizers (ICML 2023)
• Carpe Diem: Evaluation of World Knowledge in Lifelong LLMs (NAACL 2024)
• STELLA: Continual Pre-training over Audio-Video Modalities (ICML 2024)
• Adapt-∞: Scalable Lifelong Multimodal Instruction Tuning via Dynamic Data Selection (arXiv 24.10)

• ProWD: Federated Learning with Bitwidth Heterogeneity across Participating Devices (ICML 2022)
• ECoFLaP: Coarse-to-Fine Layer-Wise Pruning for Multi-/Uni-modal Large Models (ICLR 2024)
• BECoTTA: Mixture-of-Domain-Experts for Continual Test-Time Adaptation (ICML 2024)
• EVEREST: Efficient Masked Video Autoencoder by Removing Redundant Tokens (ICML 2024)
• Glider: Global and Local Instruction-Driven Expert Router (arXiv 24.10)