Minseon Gwak

/minsʌn/

prof_pic_sq.jpg

I am a postdoctoral researcher at Lawrence Berkeley National Laboratory (LBNL) and International Computer Science Institute (ICSI), working with Dr. Benjamin N. Erichson. My research bridges deep learning with system and control theories, focusing on Generative AI, World Models, and AI for Science.

Before joining LBNL and ICSI, I received my B.S.., M.S., and Ph.D. in Electrical Engineering from Pohang University of Science and Technology (POSTECH), where I was advised by Prof. PooGyeon Park. During my doctoral studies, my research centered on optimizing deep state space models for efficient sequence modeling.

✉️ Always happy to connect. mgwak [at] lbl.gov

Selected publications

  1. An Interpretable and Adaptable Framework Based on Generalized State Space Networks for Bearing Fault Diagnosis Under Sampling Rate Shift
    Minseon Gwak*, KyungSoo Kim*, Yelim Kim, Jeongmin Park, and PooGyeon Park†
    IEEE Transactions on Instrumentation and Measurement, Aug 2026
  2. Recovering Selectivity with LTI State Space Operators for Portable Long-Context Inference
    Minseon Gwak, N Benjamin Erichson, and PooGyeon Park†
    In International Conference on Machine Learning (ICML) Workshop on AdaptFM: Resource-Adaptive Foundation Model Inference, Jul 2026
  3. ConvT3: Structured State Kernels for Convolutional State Space Models
    Jaeyoung Hong*, YunYoung Choi*, Joohwan Ko, and Minseon Gwak†
    In International Conference on Learning Representations (ICLR), Apr 2026
  4. Layer-Adaptive State Pruning for Deep State Space Models
    Minseon Gwak, Seongrok Moon, Joohwan Ko, and PooGyeon Park†
    In Neural Information Processing Systems (NeurIPS), Dec 2024
  5. Scalable Robust Multi-Agent Reinforcement Learning for Model Uncertainty
    Younkyung Jwa*, Minseon Gwak*, Jiin Kwak*, Chang Wook Ahn, and PooGyeon Park†
    In 2023 62nd IEEE Conference on Decision and Control (CDC), Dec 2023
  6. Robust and explainable fault diagnosis with power-perturbation-based decision boundary analysis of deep learning models
    Minseon Gwak*, Min Su Kim*, Jong Pil Yun, and PooGyeon Park†
    IEEE Transactions on Industrial Informatics, May 2023

Honors and Awards

  • Best Graduate Research Award, POSTECH EE (2024, 2025)
  • Next-Generation Engineering Researcher Award, IPESK (2025)
  • Gold Prize, 3rd Koh Young AI Competition (2025)
  • Bronze Prize, 2nd Koh Young AI Competition (2024)
  • POSTECHIAN Innovation Fellowship, POSTECH (2024)
  • Excellent Paper Award, KIEE (2020)
  • Korea Electric Power Corporation Scholar (2020)
  • Best Senior Capstone Project Award, POSTECH EE (2019)

Academic Services

  • Reviewer
    • NeurIPS (2025, 2026)
    • NeurIPS Workshop (PriGM 2026)
    • ICML (Gold Reviewer, 2026)
    • ICML Workshop (AdaptFM 2026)
    • ICLR (2027)
    • ICLR Workshop (FM4Science 2026)