Jaewoo Ahn

Hi, I’m a Ph.D. candidate in the Department of Computer Science and Engineering at Seoul National University where I am advised by Prof. Gunhee Kim as a member of Vision & Learning Lab.

I am broadly interested in NLP and multimodal AI, particularly in human-like conversational agents. Ultimately, I aim to develop embodied conversational agents for individuals that can interact naturally in multimodal real-world environments, serving as digital avatars or personal assistants.

• In the pre-LLM era, I concentrated on dialogue systems incorporating individual personas and external knowledge in memory-based architectures.
• In the post-LLM era, I focused on persona-based role-playing LLM agents, with particular emphasis on evaluating and improving the factuality and reducing hallucinations in LLMs.
• My current research interest is in integrating multisensory perception into LLMs to support multimodal interactions. Specifically, I focus on evaluating (e.g., VLM-as-a-judge) and improving (e.g., via synthetic data generation) the factuality and reducing hallucinations in Multimodal LLMs.

You can find my CV here.