Junhyeok Kang
AI Research Scientist @ LG AI Research | Ph.D @ KAIST
Contact: e-mail: junhyeok.kang@lgresearch.ai / Google Scholar: link / LinkedIn: link
AI Research Scientist @ LG AI Research | Ph.D @ KAIST
Contact: e-mail: junhyeok.kang@lgresearch.ai / Google Scholar: link / LinkedIn: link
I am an AI Research Scientist at LG AI Research, specializing in AI for time-series data. I received my Ph.D. degree from KAIST in August 2024, focusing on developing innovative methodologies for sequential data. I also bring extensive teaching expertise, having delivered over 300 hours of AI training for major industry and government institutions, including Samsung Electronics, the National Tax Service, and Statistics Korea.
Please feel free to contact me for research discussions or potential collaborations!
AI Research Scientist @ LG AI Research | Full-time | Nov. 2024 – Present
Research on time-series forecasting.
PhD Intern @ LG AI Research | Internship | Aug. 2024 – Nov. 2024
Research on time-series forecasting.
Graduate Student Researcher @ KAIST | Full-time | Feb. 2018 – Aug. 2024
Research on time-series analysis and large-language models.
Lab representative, Data Mining Lab (Apr 2022 - Jun 2024).
AI Instructor @ Elice | Freelance | Jul. 2019 – Sep. 2023
Delivered 250+ hours of AI instruction at Samsung Electronics, including executive training.
AI Researcher @ Korea Telecom(KT) | Contract | May. 2020 – Sep. 2020
Develop a predictive model for the self-diagnosis of infectious diseases using cellular trajectories [news]
Ph.D., School of Computing @ KAIST | Mar. 2020 – Aug. 2024
Data Mining Lab (Advisor: Jae-Gil Lee)
Dissertation: Time Series Forecasting via Time-Frequency Domain Analysis
M.S., Graduate School of Data Science @ KAIST | Mar. 2018 – Feb. 2020
Data Mining Lab (Advisor: Jae-Gil Lee)
B.S., School of AI Convergence* @ Handong Global University | Feb. 2012 – Feb. 2018
B.S. in ICT Convergence and B.A. in Management (Double Major)
Summa Cum Laude
*Formerly, School of Global Entrepreneurship and ICT
Channel-wise Retrieval for Multivariate Time Series Forecasting
Junhyeok Kang†, Jun Seo, Soyeon Park, Sangjun Han, Seohui Bae, Hyeokjun Choe, Soonyoung Lee
IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2026.
Large Language Models are Zero-Shot Point-of-Interest Recommenders [paper, code]
Joeun Kim, Youngjin Seo, Yeonsoo Kim, Junhyeok Kang, Jeeho Shin, Patara Trirat, Jae-Gil Lee
Data Mining and Knowledge Discovery, 2025.
Mitigating Source Label Dependency in Time-Series Domain Adaptation under Label Shifts [paper, code, news]
Jihye Na, Youngeun Nam, Junhyeok Kang, Jae-Gil Lee
The 31st ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2025.
RA-TTA: Retrieval-Augmented Test-Time Adaptation for Vision-Language Models [paper, code]
Youngjun Lee, Doyoung Kim, Junhyeok Kang, Jihwan Bang, Hwanjun Song, Jae-Gil Lee
The 30th International Conference on Learning Representations (ICLR), 2025.
VarDrop: Enhancing Training Efficiency by Reducing Variate Redundancy in Periodic Time Series Forecasting [paper, code, blog]
Junhyeok Kang, Yooju Shin, and Jae-Gil Lee
The 39th AAAI Conference on Artificial Intelligence (AAAI), 2025.
Multi-View POI-level Cellular Trajectory Reconstruction for Digital Contact Tracing of Infectious Diseases [paper]
Dongmin Park, Junhyeok Kang, Hwanjun Song, Susik Yoon, and Jae-Gil Lee
The 22nd IEEE International Conference on Data Mining (ICDM), 2022.
Hi-COVIDNet: Deep Learning Approach to Predict Inbound COVID-19 Patients and Case Study in South Korea [paper, news]
Minseok Kim, Junhyeok Kang, Doyoung Kim, Hwanjun Song, Hyangsuk Min, Youngeun Nam, Dongmin Park, Jae-Gil Lee
The 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2020.
Robust Sequential Variational Autoencoder for Multivariate Time Series Anomaly Detection
Yao Yu, Junhyeok Kang, Jae-Gil Lee, Jonghwa Kim, Kyungdeok Seo
International Joint Conferences on Artificial Intelligence (IJCAI) Workshop on AI for Internet of Things (AI4IoT), 2020.
Preprint / Under Review
Universal Time-Series Representation Learning: A Survey [paper]
Patara Trirat, Yooju Shin, Junhyeok Kang, Youngeun Nam, Jihye Na, Minyoung Bae, Joeun Kim, Byunghyun Kim, Jae-Gil Lee
arXiv:2401.03717, 2024.
Under review at ACM Computing Surveys
Adaptive Information Routing for Multimodal Time Series Forecasting [paper]
Jun Seo, Hyeokjun Choe, Seohui Bae, Soyeon Park, Wonbin Ahn, Taeyoon Lim, Junhyeok Kang, Sangjun Han, Jaehoon Lee, Dongwan Kang, Minjae Kim, Sungdong Yoo, Soonyoung Lee
arXiv:2512.10229, 2025.
Domestic Conference & Journal
Linear Probing without Pre-Training for Time Series Forecasting [paper]
Junhyeok Kang, Minyoung Bae, Yeonsu Kim, Jae-Gil Lee
Korea Computer Congress, 2024.
Reinforcement Learning for Time Series Forecasting [paper]
Minyoung Bae, Jaehyun Park, Youngjin Seo, Junhyeok Kang, Jae-Gil Lee
Korea Computer Congress, 2024.
An Empirical Investigation of Deep Learning Models for Defect Classification in Solar Cells Electroluminescence [paper]
Haenara Shin, Junhyeok Kang, Irfan Akbar, Seola Choi, Youngeun Nam, Jae-Gil Lee
Korea Computer Congress, 608-610, 2023.
Group Periodic Pattern Mining Considering Concurrency from Spatio-temporal Trajectories [paper]
Jihye Na, Junhyeok Kang, Jae-Gil Lee, Wooshin Lee, Soyeon Jin, Sangmin Kim
Database Research, Vol 38, No. 2, pp.3-17, 2022.
Contrastive Learning with Segment Supervision for Semi-supervised Time Series Classification [paper]
Yooju Shin, Dongmin Park, Junhyeok Kang, Sejin Kim
Korea Computer Congress, 842-843, 2022.
ActiveBoostThief: Model Extraction Attack Using Reliable Active Learning [paper]
Youngeun Nam, Junhyeok Kang, Jae-Gil Lee
Korea Computer Congress, pp. 594-596, 2021.
Deep Grid Embedding using a Road Network Topology for Route Prediction [paper]
Junhyeok Kang, Jae-Gil Lee, Byeongjin Kim, Suwon Lee
Database Research, Vol 36, No. 1, pp.57-73, 2020.
Anomalous Trajectory Detection Based on Seq2Seq Auto-Encoder [paper]
Hyunsoo Kim, Junhyeok Kang, Howon Moon, Jae-Gil Lee
Journal of Korean Society for Geospatial Information Science, Vol. 28, No. 1, pp. 35-40, 2020.
Latent Representation Learning for Autoencoder-based Top-K Recommender System [paper]
Dongmin Park, Junhyeok Kang, Jae-Gil Lee
Journal of KIISE, Vol. 47, No. 2, pp. 207-215, 2020.
Analysis of Trajectory Encoding Methodology Analysis Based on Experimental Evaluation for Deep Learning [paper]
Junhyeok Kang, Minseok Kim, Jae-Gil Lee
KIISE Transactions on Computing Practices, Vol. 25, No. 8, pp. 402-406, 2019.
A Survey on Deep Learning-based Anomaly Detection Models for Time Series Data [paper]
Yao Yu, Junhyeok Kang, Jae-Gil Lee
Korea Computer Congress, pp. 919-921, 2019.
Experimental Comparison of Trajectory Encoding Methods for Deep Learning [paper]
Junhyeok Kang, Jae-Gil Lee, Yoomi Park
Korea Software Congress, pp. 204-206, 2018.
Method and Apparatus of Variate Tokenization for Time Series Forecasting
Lee, J., Kang, J., and Shin, Y.
Korean Patent Application No: 10-2024-0152425, Oct. 31, 2024.
Method for Generating Pattern Information and Electronic Apparatus Supporting Thereof
Lee, W., Kim, S., Na, J., Kang, J., and Lee, J.
Korean Patent Publication No: 10-2024-0057816, May. 03, 2024.
Trajectories Embedding Method for Deep Learning and Route Prediction Method Using the Same
Lee, J., Kang, J., Kim, M., and Lee, J.
Korean Patent No:10-2647224-0000, Mar. 08, 2024.
Method and Apparatus for Predicting Imported Infectious Disease Information Based on Deep Neural Networks
Lee, J, Kim, M., Kang, J., Kim, D., Song, H., Min, H., Nam, Y., and Park, D.
US Patent Registration No: US11557401B2, Jan. 17, 2023.
Method and Apparatus for Predicting Confirmed Patients of Infectious Disease Based on Deep Neural Networks
Lee, J, Kim, M., Kang, J., Kim, D., Song, H., Min, H., Nam, Y., and Park, D.
Korean Patent No: 10-2349270-0000, Jan. 05, 2022.
AI, From Research to Product @ CJ OliveNetworks | 2 hours | Jul. 2025
Introduction to Time-series Forecasting @ AI Camp at Handong Global University | 8 hours | Jul. 2025
Digital Transformation Training for Experienced Employees @ Samsung Electronics Co., Ltd. | 84 hours | Jan. 2023 - Apr. 2023
Digital Agility Training for Executives @ Samsung Electronics Co., Ltd. | 16 hours | Nov. 2022
Digital Transformation Training for New Employees @ Samsung Electronics Co., Ltd. | 178 hours | Oct. 2022 - Sep. 2023
Machine Learning @ Statistics Korea | 7 hours | Jul. 2022
Regression Analysis @ National Tax Service | 12 hours | Jun. 2022
Cluster Analysis @ National Tax Service | 10 hours | May. 2022
Building Chatbots in Python @ Samsung Software Academy for Youth (SAFFY) | 40 hours | Jul. 2019
Teaching Assistant of KAIST AIB (AI Business Transformation Program for CEOs; 2nd and 3rd period) @ KAIST | 2021-2022 [news]
Python Programming @ KAIST IT academy | Summer 2022
Teaching Assistant of KSE525: Data Mining and Knowledge Discovery @ KAIST ISysE | Spring 2020
Teaching Assistant of Programming II @ Handong Global University | 2017
Teaching Assistant of Programming I @ Handong Global University | 2016-2017
1st Prize, Biomedical Object Detection Contest (USD 1,500) @ AIFactory | 2020
Best Presentation Paper Award @ Korea Software Congress | 2018
Summa Cum Laude @ Handong Global University | 2018