Jaeho Kim 김재호

Bio 

Hello! I'm Jaeho Kim, a Postdoc at Korea University where I am currently working with Prof. Changhee Lee. I received my Ph.D at Ulsan National Institute of Science and Technology (UNIST), South Korea, under the guidance of Prof. Seulki Lee. 

As a researcher, I own the end-to-end lifecycle of time-series modeling, spanning data collection, feature engineering, algorithm design, embedded AI deployment, and explainability. As such, I have a holistic perspective on building next-generation time-series systems.

That said, my current research focuses on controllable time-series generation. I believe that many challenges in time-series learning stem from the lack of high-quality data that is editable, interpretable, and available across multiple temporal granularities. To address this, I am currently developing a code-based frameworks that treat time series as structured, programmable objects, enabling controllable generation, reasoning, and manipulation. Through this perspective, my goal is to build foundations for next-generation time-series models that can generate, understand, and interact with temporal data in a more flexible and scalable way.

Beyond technical research, I also view it as essential for AI researchers to engage with broader societal questions. I have contributed to discussions on the future of peer review through a position paper, which was recognized with the ICML 2025 Outstanding Position Paper Award. I am interested in bridging technical advances with their implications for how we design, evaluate, and govern AI systems.

Honors and Awards

• UNIST Best Researcher Award (Graduate Thesis Award), 2026 

• ICML 2025 Outstanding Position Paper Award, 2025

• K-DS Conference Creative Researcher Award, 2024

• Best Poster Presentation Award, UNIST AIGS, 2024

• IVI Edge Device Research Proposal Competition, 3rd Place, Hyundai Mobis, 2024

• SMILES Molecular Image Conversion Competition, 3rd Place, LG AI Research, 2020

Current Status and Timeline

• I am working as a Postdoc researcher at Korea University with Prof. Changhee Lee from Aug 2025 - August 2026

• Due to mandatory military service in South Korea, I will be available for international positions starting in September 2026.

Research Focus and Expertise

My doctoral research spanned a range of time-series learning problems, including explainable AI (XAI), dataset collection, self-supervised learning, and domain adaptation.  

Building on this foundation, my current work is centered around time-series generation, with ongoing projects in:

 (1) Language–time series reasoning via code-based representations

 (2) Tokenization for time-series foundation models 

 (3) Improving the quality of AI conference peer review

 (4) AI dependency and safety, particularly focusing on how reliance on AI systems influences human judgment and decision-making.

Collaboration Opportunities

I am open to research collaborations that could be mutually beneficial.

Personal Background

• MBTI: ENFJ - Outgoing and organized

• Languages: Korean (native),  English (native), Chinese (basic), Japanese (learning)

• Interests: Cross-fit (2 years), boxing (3 years), swimming, travel, socializing, I especially love diving into waters.

• International Experience: 

  • Indonesia (2 years), Vietnam (2 years), Malaysia (3 years)

  • Here is the list of countries that I've traveled before! More to come :) 

  • Canada, Japan, Singapore, China, Cambodia, Vietnam, Malaysia, Indonesia, Taiwan, Mongolia, Australia, Croatia, Slovenia, Bosnia, Spain, USA 

News

💊 Feb 2026:  Honored with the UNIST Best Research Award (UBRA), one of 4 recipients selected from all 2026 graduates.

💊 Jan 2026: " TimeSeg: An Information-Theoretic Segment-Wise Explainer for Time-Series Prediction" is accepted to ICLR 2026.

💊 Sep 2025: " AliO: Output Alignment Matters in Long-Term Time Series Forecasting" is accepted to NeurIPS 2025.

💊 Aug 2025: " Smart ECU: Scalable On-Vehicle Deployment of Drivetrain Fault Classification Systems for Commercial Electric Vehicles" is accepted to CIKM 2025.

💊 July 2025: "Position: The AI Conference Peer Review Crisis Demands Author Feedback and Reviewer Rewards" received the Outstanding Position Paper Award at ICML 2025

💊 May 2025: Our paper, "Position: The AI Conference Peer Review Crisis Demands Author Feedback and Reviewer Rewards",  is accepted to ICML 2025 (Oral; top 3%).  I will be presenting both works in Canada!

💊 May 2025: Our paper, "TransPL: VQ-Code Transition Matrices for Pseudo-Labeling of Time Series Unsupervised Domain Adaptation",  is accepted to ICML 2025.

💊 Jan 2025: Our paper, "PPT: Patch Order Do Matters in Time Series Pretext Task",  is accepted to ICLR 2025. I will be presenting my work in Singapore!

💊 December 2024: Two patents related to time series have been registered with the Korean Intellectual Property Office.

💊 September 2024: Our paper, "CAFO: Feature-Centric Explanation on Time Series Classification", was awarded the best poster presentation in the UNIST AI Tech Workshop 2024. 

💊 June 2024: Our paper, "Unpacking Instagram Use: The Impact of Upward Social Comparisons on Usage Patterns and Affective Experiences in the Wild",  is accepted to International Journal of Human-Computer Studies (IJHCS) Journal.

💊 June 2024: I presented "Tackling Real-World Challenges in Deep Multivariate Time Series Learning" at the A*STAR Institute, Singapore.

💊 May 2024: Our paper, "CAFO: Feature-Centric Explanation on Time Series Classification",  is accepted to KDD 2024.

💊 Apr 2023: Our paper, "Multitask Deep Learning for Human Activity, Speed, and Body Weight Estimation using Commercial Smart Insoles",  is accepted to the IEEE Internet of Things Journal.