Youngwoo Shin

Ph.D. Candidate · Electrical Engineering, KAIST

I am a Ph.D. student at the Statistical Inference & Information Theory Laboratory (SIIT) at KAIST, advised by Prof. Junmo Kim.

I study how multimodal large language models represent and reason over time, and how to interpret the generation process to align better with intention.

I am currently looking for research internship opportunities. Feel free to reach out!

      Model Interpretability
      MLLMs
      Reasoning
      Generative Models
    

News

Mar 2026Started my Ph.D. at SIIT, KAIST.

Jan 2026One paper has been accepted to ICLR 2026!

Jul 2025One paper has been accepted to ACM Multimedia 2025!

Papers Under Review

Before It Fades: Reinforcing Temporal Representations at Inference Time in VideoLLMs Youngwoo Shin, Yusung Ro, Minseo Kim, Junmo Kim Reinforcing fading temporal signals in VideoLLMs at inference time, without additional training.

          Model InterpretabilityMLLMsReasoning
        

Publications

▼

SSG: Scaled Spatial Guidance for Multi-Scale Visual Autoregressive Generation

Youngwoo Shin*, Jiwan Hur*, Junmo Kim (* equal contribution)

ICLR 2026

Training-free guidance for visual autoregressive (VAR) models, derived from an information-theoretic view of next-scale generation.

            Model InterpretabilityGenerative Models
          

            arXivOpenReviewcode
          

B4DL: A Benchmark for 4D LiDAR LLM in Spatio-Temporal Understanding

Changho Choi*, Youngwoo Shin*, Gyojin Han, Dong-Jae Lee, Junmo Kim (* equal contribution)

ACM Multimedia 2025

The first benchmark and baseline MLLM for spatio-temporal understanding of 4D LiDAR.

            MLLMs
          

            arXivcode
          

Education

2026 – present

KAIST — Ph.D. in Electrical Engineering Statistical Inference & Information Theory Laboratory (SIIT)

2024 – 2026

KAIST — M.S. in Electrical Engineering Statistical Inference & Information Theory Laboratory (SIIT)

2016 – 2023

KAIST — B.S. in Electrical Engineering

Industry Experience

2023 – 2024

LG CNS — Backend Engineer

2021 – 2023

Vitasoft — Computer Vision Researcher

Skills & Languages

LanguagesPython, C, C++

FrameworksPyTorch, TensorFlow, ONNX, TensorRT

SpokenKorean (native), English (fluent)