Research map / Optimization / Nonconvex and smooth optimization
Homotopy Smoothing for Non-Smooth Problems with Lower Complexity than $O(1/\epsilon)$
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 homotopy smoothing (HOPS-F) instead of fixed-parameter APG-F | Nesterov smoothing with FISTA-style APG · Linear classification w1a eps=1e-5 | running time (s): 7.69 -> 1.69 | Helped |
| Use HOPS-F instead of APG-F | APG-F · Image denoising eps=1e-4 | running time (s): 272.45 -> 16.52 | Helped |
| Use HOPS-D instead of APG-D | APG-D · Image denoising eps=1e-4 | iterations: 1726043 -> 4501 | Helped |
| Use HOPS-F instead of APG-F | APG-F · Matrix decomposition eps=1e-4 | running time (s): 9.16 -> 1.23 | Helped |
| Use HOPS-F instead of PD | PD primal-dual · Image denoising eps=1e-4 | running time (s): 94.26 -> 16.52 | Helped |
| Use HOPS-D instead of APG-D at loose accuracy | APG-D · Linear classification w1a eps=1e-4 | running time (s): 2.44 -> 0.44 | Helped |
| Use PD-HOPS (automatic per-epoch iterations via duality gap) instead of HOPS-F with manual tuning | HOPS-F · Matrix decomposition eps=1e-3 | running time (s): 0.91 -> 0.45 | Helped |
| Use PD-HOPS instead of HOPS-F | HOPS-F · Image denoising eps=1e-3 | running time (s): 6.99 -> 7.97 | Mixed |
| Use PD-HOPS instead of PD | PD · Linear classification w1a eps=1e-5 | running time (s): 4.33 -> 1.27 | Helped |
| Use HOPS-F instead of PD at loose accuracy | PD · Linear classification w1a eps=1e-4 | running time (s): 1.58 -> 0.46 | No clear effect |
| Use HOPS-D instead of HOPS-F | HOPS · Image denoising eps=1e-3 | running time (s): 6.99 -> 13.77 | Hurt |
| Tighten accuracy from eps=1e-3 to 1e-4 (APG-D) | APG-D · Image denoising | iterations: 179204 -> 1726043 | Hurt |
Closest papers by meaning
- Efficiently Escaping Saddle Points under Generalized Smoothness via Self-Bounding Regularity (NeurIPS 2025 · similarity 0.87)
- Nonlinearly Preconditioned Gradient Methods under Generalized Smoothness (ICML 2025 · similarity 0.87)
- Methods for Convex $(L_0,L_1)$-Smooth Optimization: Clipping, Acceleration, and Adaptivity (ICLR 2025 · similarity 0.86)
- Decentralized Gradient-Free Methods for Stochastic Non-smooth Non-convex Optimization (AAAI 2024 · similarity 0.86)
- Revisiting Frank-Wolfe for Structured Nonconvex Optimization (NeurIPS 2025 · similarity 0.85)
- Toward a Unified Theory of Gradient Descent under Generalized Smoothness (ICML 2025 · similarity 0.85)
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.