Profile Photo

Cheongwoong Kang

Ph.D. Candidate
Statistical Artificial Intelligence Lab
KAIST
cw.kang@kaist.ac.kr
[CV] [Google Scholar] [GitHub] [LinkedIn]

I am a Ph.D. candidate at the Graduate School of AI, KAIST, affiliated with the SAIL (advisor: Jaesik Choi). My research focuses on explainable AI (XAI) and natural language processing (NLP), with a particular interest in how knowledge is encoded, processed and utilized for reasoning in language models, as well as how it can be effectively extracted. My work spans two key directions: (1) analyzing the internal knowledge representations and mechanisms of language models, and (2) enhancing models by controlling or augmenting knowledge.

Publications and Preprints

An up-to-date list is available on my Google Scholar.

2025
Neural ODE Transformers: Analyzing Internal Dynamics and Adaptive Fine-tuning
Anh Tong, Thanh Nguyen-Tang, Dongeun Lee, Duc Nguyen, Toan Tran, David Leo Wright Hall, Cheongwoong Kang and Jaesik Choi
ICLR 2025
2024
Balanced Domain Randomization for Safe Reinforcement Learning
Cheongwoong Kang, Wonjoon Chang and Jaesik Choi
Applied Sciences
2023
Impact of Co-occurrence on Factual Knowledge of Large Language Models
Cheongwoong Kang and Jaesik Choi
Findings of EMNLP 2023
2022
Why Do Neural Language Models Still Need Commonsense Knowledge to Handle Semantic Variations in Question Answering?
Sunjae Kwon, Cheongwoong Kang, Jiyeon Han and Jaesik Choi
Preprint
2021
Cooperative Multi-Robot Task Allocation with Reinforcement Learning
Bumjin Park, Cheongwoong Kang and Jaesik Choi
Applied Sciences
Scheduling PID Attitude and Position Control Frequencies for Time-Optimal Quadrotor Waypoint Tracking under Unknown External Disturbances
Cheongwoong Kang, Bumjin Park and Jaesik Choi
Sensors
2020
Why Do Masked Neural Language Models Still Need Common Sense Knowledge?
Sunjae Kwon, Cheongwoong Kang, Jiyeon Han and Jaesik Choi
Reasoning for Complex QA Workshop @ AAAI 2020
2018
Symbolizing Numbers to Improve Neural Machine Translation
Cheongwoong Kang, Youngheon Ro, Jisu Kim and Heeyoul Choi
Journal of Digital Contents Society