Scholar
Doohyuk Jang
Google Scholar ID: lk0kPegAAAAJ
KAIST AI
Citations & Impact
Citations
43
H-index
2
i10-index
2
Publications
8
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- Publications: LANTERN: Accelerating Visual Autoregressive Models with Relaxed Speculative Decoding, ICLR 2025; PromptKD: Distilling Student-Friendly Knowledge for Generative Language Models via Prompt Tuning, Findings of EMNLP 2024; SeamsTalk: Seamless Talking Face Generation via Flow-Guided Inpainting, IEEE Access 2024; Meta-Awareness Enhances Reasoning Models: Self-Alignment Reinforcement Learning, Preprint; Reasoning Model is Stubborn: Diagnosing Instruction Overriding in Reasoning Models, Preprint; Med-PerSAM: One-Shot Visual Prompt Tuning for Personalized Segment Anything Model in Medical Domain, Preprint.
Research Experience
- Intern, Synopsys Korea, Gyeonggi, South Korea, Mar 2021 - Aug 2021; Developing a conversational language model for virtual doctors, AITRICS, Apr. 2024 - May. 2024.
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. Advisor: Prof. Eunho Yang.
Background
- Ph.D. student at KAIST Graduate School of AI, focusing on enhancing the efficiency of foundation models, particularly auto-regressive generative models. Aims to improve inference-time efficiency by optimizing memory usage and reducing latency, leveraging techniques such as speculative decoding and knowledge distillation. Also interested in advancing the efficiency and effectiveness of Large Reasoning Models (LRMs) on complex reasoning tasks.
Miscellany
- Teaching Experience: Teaching Assistant, Machine Learning for AI (AI501), KAIST; Teaching Assistant, Advanced Machine Learning for AI (AI601), KAIST; Academic Services: Workshop Reviewer, SCOPE@ICLR 2025.
Co-authors