Haeju Lee (이해주)
AI Researcher
Currently
LG AI Research
Research interests
Natural Language Processing (NLP), Reinforcement Learning (RL)
Education
2021 - 2023
KAIST
- Graduate School of AI, MS (advisor: Kee-Eung Kim)
2014 - 2021
Yonsei University
- Department of Computer Science, BS
Employment
2020.10 - 2020.12
Pozalabs (Seoul, Korea)
- Research Intern
- • Conducted research on finding how to fit generative model (e.g. GPT) to compose MIDI scores
2020.03 - 2020.09
Naver Labs (Seongnam, Korea)
- Research Engineer Intern, Robot Dynamics & Control Team
- • Implemented simulator for Ambidex, an ambidexterous robot designed by Naver Labs, to promote efficient learning (RL, IL), based on Mujoco and Metaworld
- • Created Python binding of existing Ambidex controller code (C++), to plug Ambidex into RL training loop
- • Designed and experimented various reward functions for Ambidex by running standard RL algorithms (PPO, SAC) in Ambidex simulator
2019.07 - 2019.12
Bear Robotics (San Francisco Bay Area, US)
- Software Engineer Intern, Robotics Team
- • Implemented features for serving robot by customizing ROS C++ code
- • Implemented program for testing robot navigation stack in Python
Publications
Papers
2024
LG AI Research: Soyoung An, Kyunghoon Bae, Eunbi Choi, Stanley Jungkyu Choi, Yemuk Choi, Seokhee Hong, Yeonjung Hong, Junwon Hwang, Hyojin Jeon, Gerrard Jeongwon Jo, Hyunjik Jo, Jiyeon Jung, Yountae Jung, Euisoon Kim, Hyosang Kim, Joonkee Kim, Seonghwan Kim, Soyeon Kim, Sunkyoung Kim, Yireun Kim, Youchul Kim, Edward Hwayoung Lee, Haeju Lee, Honglak Lee, Jinsik Lee, Kyungmin Lee, Moontae Lee, Seungjun Lee, Woohyung Lim, Sangha Park, Sooyoun Park, Yongmin Park, Boseong Seo, Sihoon Yang, Heuiyeen Yeen, Kyungjae Yoo, Hyeongu Yun, “EXAONE 3.0 7.8B Instruction Tuned Language Model”, ArXiv, 2024
2023
Haeju Lee, Minchan Jeong, Se-Young Yun, Kee-Eung Kim, “Bayesian Multi-Task Transfer Learning for Soft Prompt Tuning”, EMNLP Findings (and also Masters Degree Thesis), 2023
2022
Haeju Lee, Oh Joon Kwon, Yunseon Choi, Minho Park, Ran Han, Yoonhyung Kim, Jinhyeon Kim, Youngjune Lee, Haebin Shin, Kangwook Lee, Kee-Eung Kim, “Learning to Embed Multi-Modal Contexts for Situated Conversational Agents”, NAACL Findings, 2022
Haeju Lee, Oh Joon Kwon, Yunseon Choi, Minho Park, Ran Han, Yoonhyung Kim, Jinhyeon Kim, Youngjune Lee, Haebin Shin, Kangwook Lee, Kee-Eung Kim, “Tackling Situated Multi-Modal Task-Oriented Dialogs with a Single Transformer Model”, AAAI DSTC10 Workshop, 2022
2021
Youngjune Lee, Oh Joon Kwon, Haeju Lee, Joonyoung Kim, Kangwook Lee, Kee-Eung Kim, “Augment & Valuate : A Data Enhancement Pipeline for Data-Centric AI”, NeurIPS Data-Centric AI workshop, 2021
Patents
2023
트랜스포머 모델을 이용한 농산물 가격예측 방법 및 시스템
(Method and System for Predicting Price of Agricultural Product based on Transformer Model)
Awards & Competitions
2021
1st place on DSTC 10 challenge (track 3: SIMMC2.0)
Data-Centric AI Competition, Honorable Mention
2018
Samsung AI Hackathon, Special award
-
- • Indoor cleaning robot navigation using deep learning model