Changhun Kim

Changhun Kim

Machine Learning Researcher

AITRICS

I am a machine learning researcher at AITRICS, a healthcare AI startup in Korea, where I am fulfilling my alternative military service under the supervision of Professor Eunho Yang. Prior to joining AITRICS, I completed my Master’s degree in Artificial Intelligence at KAIST, also under Professor Yang’s guidance. Additionally, I hold a Bachelor’s degree in Computer Science and Mathematics from KAIST.

My research has focused on the essential dimensions of trustworthy machine learning, explored across various application domains including 3D point cloud classification, automatic speech recognition, tabular learning, and time series analysis:

  • Out-of-distribution generalization
  • Explainability
  • Uncertainty calibration

As I progress in my academic and research journey, I am particularly drawn to the development of trustworthy deep generative models and the theoretical frameworks that can demystify artificial intelligence systems, especially in terms of generalizaion error bounds. My long-term research vision aims to establish truly trustworthy machine learning systems through two complementary approaches:

  • Designing empirically well-motivated and theoretically grounded practical algorithms
  • Conducting rigorous theoretical analyses of empirical machine learning techniques

I am open to research collaborations on a global scale and intend to pursue doctoral studies beginning in Fall 2026. Remote collaboration opportunities are welcome. Should my research interests align with your work, I would be pleased to discuss potential collaborations.

Contact
  • AP Tower 13F, 218, Teheran-ro, Gangnam-gu, Seoul 06221, Korea
  • (+82) 10-3264-6509
  • changhun.a.kim@gmail.com
Education
  • M.S. in Artificial Intelligence, 2024

    Korea Advanced Institute of Science and Technology (KAIST)

  • B.S. in Computer Science and Mathematics, 2022

    Korea Advanced Institute of Science and Technology (KAIST)

News
Professional Experience

AITRICS
AITRICS, Seoul, Korea
Machine Learning Researcher, Nov 2023 – Present
KAIST Machine Learning and Intelligence Lab
KAIST Machine Learning and Intelligence Lab, Daejeon, Korea
Graduate Researcher, Mar 2022 – Feb 2024
Undergraduate Researcher, Jun 2021 – Feb 2022
DeepNatural
DeepNatural, Seoul, Korea
Machine Learning Engineer, Sep 2020 – Feb 2021
KAIST Vehicular Intelligence Lab
KAIST Vehicular Intelligence Lab, Daejeon, Korea
Undergraduate Researcher, Oct 2019 – Aug 2020
Netmarble
Netmarble, Seoul, Korea
Data Engineer, Jun 2019 – Aug 2019
Publications
Soft Equivariance Regularization for Invariant Self-Supervised Learning
Soft Equivariance Regularization for Invariant Self-Supervised Learning

Manuscript, 2025
Delta-XAI: A Unified Framework for Explaining Prediction Shifts in Online Time Series
Delta-XAI: A Unified Framework for Explaining Prediction Shifts in Online Time Series

Manuscript, 2025
SPAM: Sampling Pattern Meta-Learning for Domain Generalization on Irregular Time Series
SPAM: Sampling Pattern Meta-Learning for Domain Generalization on Irregular Time Series

Manuscript, 2025
DeltaSHAP: Explaining Prediction Evolutions in Online Patient Monitoring with Shapley Values
DeltaSHAP: Explaining Prediction Evolutions in Online Patient Monitoring with Shapley Values

Manuscript, 2025
AdapTable: Test-Time Adaptation for Tabular Data via Shift-Aware Uncertainty Calibrator and Label Distribution Handler
AdapTable: Test-Time Adaptation for Tabular Data via Shift-Aware Uncertainty Calibrator and Label Distribution Handler

NeurIPS Workshop on Table Representation Learning, 2024
Paper Code
TIMING: Temporality-Aware Integrated Gradients for Time Series Explanation
TIMING: Temporality-Aware Integrated Gradients for Time Series Explanation

International Conference on Machine Learning (ICML), 2025 (Spotlight, 313/12107=2.6%)
ICLR Workshop on XAI4Science: From Understanding Model Behavior to Discovering New Scientific Knowledge, 2025
Code
Stable-TTS: Stable Speaker-Adaptive Text-to-Speech Synthesis via Prosody Prompting
Stable-TTS: Stable Speaker-Adaptive Text-to-Speech Synthesis via Prosody Prompting

IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2025
Paper
CloudFixer: Test-Time Adaptation for 3D Point Clouds via Diffusion-Guided Geometric Transformation
CloudFixer: Test-Time Adaptation for 3D Point Clouds via Diffusion-Guided Geometric Transformation

European Conference on Computer Vision (ECCV), 2024
Paper Code
SGEM: Test-Time Adaptation for Automatic Speech Recognition via Sequential-Level Generalized Entropy Minimization
SGEM: Test-Time Adaptation for Automatic Speech Recognition via Sequential-Level Generalized Entropy Minimization

Conference of the International Speech Communication Association (INTERSPEECH), 2023 (Oral, 348/2293=15.2%)
Paper Code
Patents
Method for Providing Explanation for Patient State Prediction and Electronic Apparatus Therefor

KR Patent App. 10-2025-0009664, 2025
Test-Time Adaptation for Automatic Speech Recognition via Sequential-Level Generalized Entropy Minimization

US Patent App. 18/594,442, 2024
KR Patent App. 10-2024-0006413, 2024
KR Patent App. 10-2024-0023266, 2024
Awards and Honors

Dongwon Full Masters Scholarship
Dongwon Group

Magna Cum Laude
KAIST College of Engineering

Silver Prize
Korean Undergraduate Mathematics Competition, Korean Mathematics Society

Overseas Exchange Scholarship
Mirae Asset

Representative of Student Exchange Ambassador
KAIST

Honor Student
KAIST College of Engineering

KAIST Convergence AMP Scholarship
KAIST School of Computing

Winner
Science Quiz, KAIST-POSTECH Science War

National Full Undergraduate Scholarship
Korea Student Aid Foundation
Teaching Experience

Teaching Assistant
Tabular Learning, Hanwha Ocean Capstone Project

Teaching Assistant
AI Soccer Challenge, Bokja Girls’ High School AI Education Program
Academic Service

Journal Reviewer

Conference Reviewer