LMLS Lab.

Large-Scale Machine Learning Systems Lab.

@ Inha University

대형 머신러닝 시스템 연구실

Our Mission

The Large-Scale Machine Learning Systems Lab. studies critical problems in Machine Learning, Optimization, and High-Performance Computing. We currently pursue the following three research directions: 1) System-efficient large-scale neural network training strategies, 2) Federated Learning on heterogeneous systems, and 3) Applied Machine Learning for Physics.

Lab. News

[November 2025] Sungmin, our external researcher, and Jisoo's paper, "GEM: A Scale-Aware and Distribution-Sensitive Sparse Fine-Tuning Framework for Effective Downstream Adaptation" has been accepted to AAAI, 2026. Congratulations!
[November 2025] Sunwoo Lee will attend Korean Artificial Intelligence Fall Conference (한국인공지능학회 추계학술대회).
[October 2025] Junhyuk and Jihyun's paper, "Asynchronous-Parallel Sharpness-Aware Minimization for Efficient Deep Learning" has been accepted to Journal of KIISE (정보과학회논문지), 2025. Congratulations!
[September 2025] Jisoo, Seungjun, and Taekyun won 3rd place (Encouragement Award) in Samsung AI Challenge 2025. Congratulations!
[September 2025] Jisoo's paper, "Layer-wise Update Aggregation with Recycling for Communication-Efficient Federated Learning" has been accepted to Annual Conference on Neural Information Processing Systems (NeurIPS) (BK-21 IF:4) 2025. Congratulations!
[August 2025] Jihyun and Taekyun won 'Excellence Award' in DACON AI Challenge at Inha. Congratulations!
[July 2025] Jihyun's paper, "Enabling Weak Client Participation via On-Device Knowledge Distillation in Heterogeneous Federated Learning" has been accepted to European Conference on Artificial Intelligence (ECAI) (BK-21 IF:1) 2025. Congratulations!
[June 2025] Sunwoo Lee has been recognized as an OUTSTANDING REVIEWER (top 10%) in KDD 2025.
[April 2025] Jihyun's paper, "Highly Stable Two-level Current Fluctuation in Complex Oxide Heterostructures" has been accepted to Nature Communications (JCR 6.6%, IF: 14.7). Congratulations!
[April 2025] Jihyun's paper, "Machine Learning Approach to Characterize Ferromagnetic La0.7Sr0.3MnO3 Thin Films via Featurization of Surface Morphology" has been accepted to Advanced Science (JCR 6.5%, IF: 14.3)! Congratulations!
[March 2025] Hyuntak will join LMLS-lab as a graduate student, Welcome!
[March 2025] Sunwoo Lee will serve as a Department Vice Chair from March 2025.
[February 2025] Sunwoo Lee will give a talk at Hanyang University about system-efficient federated learning methods on 2/26.
[February 2025] Sunwoo Lee gave a lecture (Introduction to Federated Learning) in BK-21 AI Winter School at Inha University.