Doohyuk Jang

Doohyuk Jang

Ph.D student at Graduate School of AI, Korea Advanced Institute of Science and Technology (KAIST)Daejeon, South Korea

Hello, I am a Ph.D student at Machine Learning and Intelligence Lab (MLILAB) in KAIST, advised by Prof. Eunho Yang.

My research centers on advancing the reasoning capabilities of large language models, mainly through reinforcement learning-based post-training. More recently, my interest has been moving toward physical AI, carrying that reasoning beyond text into embodied agents that perceive, act, and learn from continuous interaction with the physical world. I also continue to work on inference efficiency for foundation models, through techniques such as speculative decoding and knowledge distillation.

Publications

  • Verifying Meta-Awareness via Predictive Rewards in Reasoning Models
    Yoonjeon Kim*, Doohyuk Jang*, Eunho Yang
    ICML 2026 paper
  • Reasoning Model is Stubborn: Diagnosing and Mitigating Reasoning Rigidity in Large Language Models
    Doohyuk Jang*, Yoonjeon Kim*, Chanjae Park, Hyun Ryu, Eunho Yang
    Findings of EMNLP 2026 paper project page
  • LANTERN: Accelerating Visual Autoregressive Models with Relaxed Speculative Decoding
    Doohyuk Jang*, Sihwan Park*, June Yong Yang, Yeonsung Jung, Jihun Yun, Souvik Kundu, Sungyub Kim, Eunho Yang
    ICLR 2025 paper
  • Discounted Beta-Bernoulli Reward Estimation for Sample-Efficient Reinforcement Learning with Verifiable Rewards
    Haechan Kim, Soohyun Ryu, Gyouk Chu, Doohyuk Jang, Eunho Yang
    ICML 2026 paper
  • ReviewScore: Misinformed Peer Review Detection with Large Language Models
    Hyun Ryu, Doohyuk Jang, Hyemin S. Lee, Joonhyun Jeong, Gyeongman Kim, Donghyeon Cho, Gyouk Chu, Minyeong Hwang, Hyeongwon Jang, Changhun Kim, Haechan Kim, Jina Kim, Joowon Kim, Yoonjeon Kim, Kwanhyung Lee, Chanjae Park, Heecheol Yun, Gregor Betz, Eunho Yang
    Findings of EMNLP 2026 paper
  • PromptKD: Distilling Student-Friendly Knowledge for Generative Language Models via Prompt Tuning
    Gyeongman Kim, Doohyuk Jang, Eunho Yang
    Findings of EMNLP 2024 paper
  • SeamsTalk: Seamless Talking Face Generation via Flow-Guided Inpainting
    Yeongho Jeong, Gyeongman Kim, Doohyuk Jang, Jaeryong Hwang, Eunho Yang
    IEEE Access 2024 paper
  • Med-PerSAM: One-Shot Visual Prompt Tuning for Personalized Segment Anything Model in Medical Domain
    Hangyul Yoon, Doohyuk Jang, Jungeun Kim, Eunho Yang
    Preprint paper

Workshop Papers

  • LANTERN++: Enhancing Relaxed Speculative Decoding with Static Tree Drafting for Visual Auto-regressive Models
    Sihwan Park*, Doohyuk Jang*, Sungyub Kim, Souvik Kundu, Eunho Yang
    SCOPE @ ICLR 2025 Oral Presentation paper code

Education

  • Ph.D. in Graduate School of AI, Korea Advanced Institute of Science and Technology (KAIST)
    Sep. 2024 - Present
  • M.S. in Graduate School of AI, Korea Advanced Institute of Science and Technology (KAIST)
    Mar. 2023 - Aug. 2024
  • B.S. in Electrical Engineering, Computer Science, Korea Advanced Institute of Science and Technology (KAIST)
    Mar. 2018 - Feb. 2023

Work Experiences

  • Intern, Synopsys KoreaGyeonggi, South Korea
    Mar 2021 - Aug 2021
  • WBL Residency, NAVER CloudGyeonggi, South Korea
    Oct 2025 - Feb 2026

Projects

  • Developing a conversational language model for virtual doctors, AITRICS
    Apr. 2024 - May. 2024

Academic Services

  • Conference Reviewer
    • NeurIPS 2025
    • CVPR 2026
    • NeurIPS 2026
  • Workshop Reviewer
    • SCOPE@ICLR 2025

Teaching Experience

  • Teaching Assistant, Machine Learning for AI (AI501), KAIST
    Fall 2024
  • Teaching Assistant, Advanced Machine Learning for AI (AI601), KAIST
    Spring 2025, Fall 2025