Institution: Sungkyunkwan University
Advisor: Prof. Woocheol Choi
Period: Dec. 2023 – May 2025
Overview
This work analyzes the convergence of the Jacobi–Proximal ADMM, a parallelizable algorithm for large-scale multi-block optimization problems. We proved linear convergence under strongly convex and smooth objectives, and validated the results through numerical experiments.
Contributions
- Developed theoretical proofs establishing linear convergence of the Jacobi–Proximal ADMM
- Implemented numerical experiments to validate convergence results
Preprint
Preprint (arXiv:2503.18601), under revision at Computational Optimization and Applications
Link
Talks
- Contributed talk at KMS Spring Meeting 2025, KAIST, April 2025 | slide
- Seminar talk at SKKU, April 2025 | video