Sangwoo Shin

SangwooShin.jpeg

Hello! I am a M.S. student at Sungkyunkwan University in South Korea. I am a member of the CSI-Agent Group and am being advised by Professor Honguk Woo.

My research focuses on language grounding for embodied control, particularly in developing efficient systems that can bridge the gap between large language models and physical robot actions. Specifically, I work on enabling robots to accomplish complex tasks in diverse environments and make LLMs reasoning align with physical environmental contexts.

Prior to my graduate studies, I earned a B.S degree in mathematics from Sungkyunkwan University.

news

Sep 23, 2024 Our CoREN paper got accepted to EMNLP Findings 2024!
May 16, 2024 Our SemGro paper got accepted to ACL Findings 2024!
Dec 9, 2023 Our SemTra paper got accepted to AAAI2024!

selected publications

  1. Conference
    LLM-Based Offline Learning for Embodied Agents via Consistency-Guided Reward Ensemble
    Yujeong Lee*, Sangwoo Shin*, Wei-Jin Park, and 1 more author
    In EMNLP Findings, 2024
  2. Conference
    Semantic Skill Grounding for Embodied Instruction-Following in Cross-Domain Environments
    Sangwoo Shin*, Seunghyun Kim*, Youngsoo Jang, and 2 more authors
    In ACL Findings, 2024
  3. Conference
    SemTra: A Semantic Skill Translator for Cross-Domain Zero-Shot Policy Adaptation
    Sangwoo Shin, Minjong Yoo, Jeongwoo Lee, and 1 more author
    In AAAI, 2024
  4. Conference
    One-shot Imitation in a Non-Stationary Environment via Multi-Modal Skill
    Sangwoo Shin, Daehee Lee, Minjong Yoo, and 2 more authors
    In ICML, 2023