Lee Hyun profile photo

I am an AI researcher at Samsung Advanced Institute of Technology (SAIT). I received my M.S. in Electrical Engineering from POSTECH, where I was advised by Prof. Tae-Hyun Oh.

My research focuses on generative AI and multimodal large language models, spanning laughter understanding, video reasoning, speech-to-3D talking head generation, and music generation. More recently, I have been particularly interested in LLM agents for scientific discovery, including post-training, multi-agent systems, and traceable reasoning for forward and inverse design, with a focus on reward design for RL.

News

2026 Our paper on multimodal laughter reasoning, SMILE-Next, is accepted to ACL 2026 Oral.
2025 Our paper on a rejection sampling for LLM-based materials discovery appears at AI for Science @ NeurIPS 2025.
2025 VLM's Eye Examination is published in TMLR 2025.
2025 Our paper on 3D face mesh video reconstruction is invited to ICLR 2025 from TMLR.
2024 I joined Samsung Advanced Institute of Technology as an AI Researcher.
2024 SMILE is accepted to NAACL 2024, and LaughTalk is accepted to WACV 2024.

Selected Publications

  1. SMILE-Next figure
    SMILE-Next: Teaching Large Language Models to Detect, Classify, and Reason about Laughter
    Lee Jung-Mok, Kim Sung-Bin, Joohyun Chang, Lee Hyun, and Tae-Hyun Oh
    Annual Meeting of the Association for Computational Linguistics (ACL), 2026 Oral
    Project Page
  2. Materials discovery figure
    Aligning Reasoning LLMs for Materials Discovery with Physics-aware Rejection Sampling
    Lee Hyun, Sohee Yoon, Jinwoo Park, and collaborators
    AI for Science Workshop at NeurIPS, 2025
    Paper
  3. VLM eye examination thumbnail
    VLM's Eye Examination: Instruct and Inspect Visual Competency of Vision Language Models
    Nam Hyeon-Woo, Moon Ye-Bin, Wonseok Choi, Lee Hyun, and Tae-Hyun Oh
    Transactions on Machine Learning Research (TMLR), 2025
    Paper
  4. SMILE dataset thumbnail
    SMILE: Multimodal Dataset for Understanding Laughter in Video with Language Models
    Lee Hyun*, Kim Sung-Bin*, Seungju Han, Youngjae Yu, and Tae-Hyun Oh (*equal contribution)
    Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL), 2024
    Paper Dataset
  5. LaughTalk thumbnail
    LaughTalk: Expressing 3D Talking Head Generation with Laughter
    Kim Sung-Bin, Lee Hyun, Dahye Hong, Suekyeong Nam, Janghoon Joo, and Tae-Hyun Oh
    IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2024
    Paper Project Page
  6. ComMU thumbnail
    ComMU: Dataset for Combinatorial Music Generation
    Lee Hyun, Teahyun Kim, Hyolim Kang, Minjoo Ki, Hyeonchan Hwang, and SeonJoo Kim
    Conference on Neural Information Processing Systems (NeurIPS), 2022
    Paper Project Page

Work Experience

2024 - Present AI Researcher, Samsung Advanced Institute of Technology (SAIT)
2022 - 2024 Graduate Researcher, POSTECH AMI Lab
2021 - 2022 Machine Learning Engineer, POZAlabs
2017 - 2018 Military Service, Republic of Korea Army, Vanguard Unit, discharged as sergeant

Teaching and Mentoring

2024 Instructor, Introduction to LLM, Fastcampus
2021 - 2025 Teaching Assistant, Naver Boostcamp AI Tech
2022 - 2023 Teaching Assistant, NLP and speech synthesis courses at Samsung Software Academy For Youth
2022 Teaching Assistant, Computer Vision (EECE 490) at POSTECH
Mentoring JungMok Lee, now a Ph.D. student at KAIST, collaborated on visual LLM reasoning work accepted to ACL 2026

Professional Services

Conference Reviewer: CVPR, ICLR, NeurIPS, ACL, ICML, ICCV, AAAI