Research map / Optimization / Nonconvex and smooth optimization
Optimal Algorithms for Stochastic Multi-Level Compositional Optimization
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 SMVR instead of the strongest reported baseline | Hierarchical TERM · Hiv-1 rare | classification accuracy (%): 76.7± 2.4 -> 79.0 ± 2.5 | Helped |
| Use SMVR instead of the strongest reported baseline | Hierarchical TERM · Hiv-1 overall | classification accuracy (%): 88.7± 1.2 -> 90.0 ± 1.2 | Helped |
| Use SMVR instead of the strongest reported baseline | Hierarchical TERM · Australian scale rare | classification accuracy (%): 79.4± 8.4 -> 82.9 ± 8.2 | Helped |
| Use SMVR instead of the strongest reported baseline | Hierarchical TERM · Australian scale overall | classification accuracy (%): 78.7± 5.8 -> 82.6 ± 4.6 | Helped |
| Use SMVR instead of the strongest reported baseline | Hierarchical TERM · Breast-cancer rare | classification accuracy (%): 61.8± 8.6 -> 74.0 ± 8.5 | Helped |
| Use SMVR instead of the strongest reported baseline | Hierarchical TERM · Breast-cancer overall | classification accuracy (%): 83.3± 6.6 -> 87.0 ± 4.3 | Helped |
| Use SMVR instead of the strongest reported baseline | Hierarchical TERM · Svmguide1 rare | classification accuracy (%): 76.5± 4.6 -> 80.9 ± 3.0 | Helped |
| Use SMVR instead of the strongest reported baseline | Hierarchical TERM · Svmguide1 overall | classification accuracy (%): 85.8± 2.1 -> 87.3 ± 2.7 | Helped |
| Add projection to Jacobian estimators (ablation) | SMVR · Risk-averse portfolio optimization | loss and gradient convergence | Hurt |
| Use a small projection bound rather than a large one (ablation) | SMVR · Risk-averse portfolio optimization | loss and gradient convergence | Hurt |
| Use adaptive learning rates (Adam-type) (ablation) | SMVR · 5-step MAML on Omniglot | accuracy | Helped |
Closest papers by meaning
- Projection-Free Variance Reduction Methods for Stochastic Constrained Multi-Level Compositional Optimization (ICML 2024 · similarity 0.92)
- Gradient-Free Methods for Nonconvex Nonsmooth Stochastic Compositional Optimization (NeurIPS 2024 · similarity 0.91)
- Stochastic Momentum Methods for Non-smooth Non-Convex Finite-Sum Coupled Compositional Optimization (NeurIPS 2025 · similarity 0.90)
- Faster Stochastic Variance Reduction Methods for Compositional MiniMax Optimization (AAAI 2024 · similarity 0.90)
- A Near-Optimal Single-Loop Stochastic Algorithm for Convex Finite-Sum Coupled Compositional Optimization (ICML 2025 · similarity 0.90)
- Adaptive Variance Reduction for Stochastic Optimization under Weaker Assumptions (NeurIPS 2024 · similarity 0.89)
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.