Research map / Optimization / Nonconvex and smooth optimization
Non-Smooth Weakly-Convex Finite-sum Coupled Compositional Optimization
Quanqi Hu, Dixian Zhu, Tianbao Yang
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 |
|---|---|---|---|
| replace smoothed TPAUC surrogate (SOTAs) with exact non-smooth objective via SONX | SONX · moltox21 (t0) | TPAUC (0.5, 0.5): 0.23 -> 0.252 | Improved |
| replace SOTA-s with SONX | SONX · molmuv (t1) | TPAUC (0.5, 0.5): 0.526 -> 0.575 | Improved |
| replace SOTA-s with SONX | SONX · molpcba (t0) | TPAUC (0.5, 0.5): 0.314 -> 0.335 | Improved |
| replace SOTA-s with SONX at stricter TPAUC range | SONX · moltox21 (t0) | TPAUC (0.6, 0.4): 0.068 -> 0.07 | Improved |
| replace SOTA-s with SONX at stricter TPAUC range | SONX · molmuv (t1) | TPAUC (0.6, 0.4): 0.327 -> 0.347 | Improved |
| replace CE loss with SONX | SONX · molmuv (t1) | TPAUC (0.6, 0.4): 0.161 -> 0.347 | Improved |
| use SONT (att) with exact pooled objective instead of mini-batch pooling SOTAs (att) | SONT (att) · Fox | TPAUC (0.5, 0.5): 0.024 -> 0.12 | Improved |
| replace SOTAs (att) with SONT (att) | SONT (att) · Lung | TPAUC (0.5, 0.5): 0.539 -> 0.639 | Improved |
| replace SOTAs (att) with SONT (att) | SONT (att) · MUSK2 | TPAUC (0.5, 0.5): 0.6 -> 0.7 | Improved |
| replace SOTAs (att) with SONT (att) | SONT (att) · Colon | TPAUC (0.1, 0.9): 0.911 -> 0.916 | Improved |
| SONT (att) vs AUC-M (att) at loose TPAUC range | SONT (att) · MUSK2 | TPAUC (0.1, 0.9): 0.867 -> 0.867 | No clear effect |
| replace MIDAM (att) with SONT (att) | SONT (att) · Colon | TPAUC (0.5, 0.5): 0.548 -> 0.8 | Improved |
Closest papers by meaning
- Stochastic Momentum Methods for Non-smooth Non-Convex Finite-Sum Coupled Compositional Optimization (NeurIPS 2025 · similarity 0.96)
- A Near-Optimal Single-Loop Stochastic Algorithm for Convex Finite-Sum Coupled Compositional Optimization (ICML 2025 · similarity 0.90)
- Gradient-Free Methods for Nonconvex Nonsmooth Stochastic Compositional Optimization (NeurIPS 2024 · similarity 0.89)
- Projection-Free Variance Reduction Methods for Stochastic Constrained Multi-Level Compositional Optimization (ICML 2024 · similarity 0.87)
- Optimizing $(L_0, L_1)$-Smooth Functions by Gradient Methods (ICLR 2025 · similarity 0.86)
- Stability and Generalization of Stochastic Compositional Gradient Descent Algorithms (ICML 2024 · similarity 0.86)
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.
