Research map / Optimization / Nonconvex and smooth optimization
MADMM: A Generic Algorithm for Non-smooth Optimization on Manifolds
Artiom Kovnatsky, Klaus Glashoff, Michael M. Bronstein
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use MADMM instead of subgradient descent for robust Euclidean embedding | REE (L1 MDS) · REE, US cities n=500 | iterations to converge: 500 -> 50 | Improved |
| use MADMM instead of SDP for robust Euclidean embedding | REE (L1 MDS) · REE scalability, random data | scalability with problem size: SDP did not scale beyond n = 100 | Improved |
| use robust L1 embedding solved by MADMM instead of classical MDS | Euclidean embedding with outlier distances · US cities n=500, doubled distances | embedding quality (qualitative) | Improved |
| vary X-step solver (trust regions vs conjugate gradients) and number of X-step iterations (ablation) | MADMM · Compressed modes, n=500, k=10 | convergence behavior | No clear effect |
| change random initialization (ablation) | MADMM · Compressed modes, n=500, k=10 | convergence | No clear effect |
| use MADMM instead of splitting methods of Lai-Osher and Neumann et al. | compressed modes · Human mesh, n=8K | convergence speed | Improved |
| use MADMM with L2,1 norm instead of least-squares solution | functional correspondence · TOSCA, L=6 shapes, 16% outliers | correspondence accuracy | Improved |
| use MADMM instead of smoothing the L2,1 norm | functional correspondence · TOSCA, L=2 shapes | convergence | Improved |
| increase subgradient initial step size c (ablation) | subgradient REE solver · US cities n=500 | convergence | Mixed |
Later papers that compared with this method
1 later accepted paper reports MADMM as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Adaptive Riemannian ADMM for Nonsmooth Optimization: Optimal Complexity without Smoothing (NeurIPS 2025 · better on 4 of 4 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Adaptive Riemannian ADMM for Nonsmooth Optimization: Optimal Complexity without Smoothing (NeurIPS 2025 · similarity 0.85)
- Riemannian coordinate descent algorithms on matrix manifolds (ICML 2024 · similarity 0.85)
- Efficient Optimization with Orthogonality Constraint: a Randomized Riemannian Submanifold Method (ICML 2025 · similarity 0.85)
- Implicit Riemannian Optimism with Applications to Min-Max Problems (ICML 2025 · similarity 0.84)
- Nonconvex Federated Learning on Compact Smooth Submanifolds With Heterogeneous Data (NeurIPS 2024 · similarity 0.84)
- Optimizing ADMM and Over-Relaxed ADMM Parameters for Linear Quadratic Problems (AAAI 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
