Research map / Optimization / Nonconvex and smooth optimization
Task Embedded Coordinate Update: A Realizable Framework for Multivariate Non-Convex Optimization
Yiyang Wang, Risheng Liu, Long Ma, Xiaoliang Song
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 ADMM for the dictionary D subproblem rather than PITH for the sparse-coding W subproblem (ablation) | TECU for ℓ0DL · ℓ0DL synthetic, n=64 | computation time (s): 2.22s vs nearly 600s | Improved |
| Embed ADMM for the D subproblem | TECU for ℓ0DL · ℓ0DL synthetic | computation time (s) | Improved |
| Embed PITH for the W subproblem instead of using prox-linear updates (ablation) | TECU for ℓ0DL · ℓ0DL synthetic | computation time (s): 2.22s vs nearly 600s | Hurt |
| Embed a residual-type CNN for illumination-layer updates | TECU for low-light image enhancement · Low-light image enhancement example | PSNR | Improved |
| Embed a residual-type CNN in the coordinate-descent scheme rather than use the classical CD scheme alone | TECU for low-light image enhancement · Low-light image enhancement example | PSNR | Improved |
| Embed a residual-type CNN for illumination-layer updates | TECU for low-light image enhancement · NASA | NIQE: 3.43 -> 3.41 | Improved |
| Embed a residual-type CNN for illumination-layer updates | TECU for low-light image enhancement · Non-uniform | NIQE: 2.94 -> 2.91 | Improved |
| Use TECU rather than the strongest listed baseline, BCU | ℓ0DL · ℓ0DL synthetic, n=64 | computation time (s): 3.87 -> 2.22 | Improved |
| Use TECU rather than the strongest listed baseline, BCU | ℓ0DL · ℓ0DL synthetic, n=144 | computation time (s): 14.45 -> 11.96 | Improved |
| Use TECU rather than the strongest listed baseline, BCU | ℓ0DL · ℓ0DL synthetic, n=256 | computation time (s): 100.19 -> 75.19 | Improved |
Closest papers by meaning
- Cyclic Block Coordinate Descent With Variance Reduction for Composite Nonconvex Optimization (ICML 2023 · similarity 0.85)
- BPQP: A Differentiable Convex Optimization Framework for Efficient End-to-End Learning (NeurIPS 2024 · similarity 0.85)
- Differentiable Convex Optimization Layers (NeurIPS 2019 · similarity 0.85)
- LPGD: A General Framework for Backpropagation through Embedded Optimization Layers (ICML 2024 · similarity 0.84)
- Delve into the Applicability of Advanced Optimizers for Multi-Task Learning (NeurIPS 2026 · similarity 0.84)
- General framework for online-to-nonconvex conversion: Schedule-free SGD is also effective for nonconvex optimization (ICML 2025 · 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.
