Research map / Optimization / Optimizers
Non-Convex Bilevel Optimization with Time-Varying Objective Functions
Sen Lin, Daouda Sow, Kaiyi Ji, Yingbin Liang, Ness Shroff
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 |
|---|---|---|---|
| window-average hypergradients over previous steps (SOBOW) instead of current estimate only (OGD) (ablation) | SOBOW · Online HR (static and dynamic) | regret | Improved |
| SOBOW instead of OAGD (no re-evaluation of past functions) | online HR, K=50 · Online HR, 5000 steps | running time (s): 228 -> 11 | Improved |
| SOBOW instead of OAGD | online HR · Online HR | regret | No clear effect |
| SOBOW-4 instead of OAGD-4 | static online HO · 20 Newsgroup static | time (s): 2304 -> 899 | Improved |
| SOBOW-4 instead of OAGD-4 | static online HO · 20 Newsgroup static | accuracy (%): 65.96 -> 65.87 | No clear effect |
| SOBOW-50 instead of OAGD-50 | static online HO · 20 Newsgroup static | time (s): 20161 -> 1188 | Improved |
| SOBOW-50 instead of OAGD-50 | static online HO · 20 Newsgroup static | accuracy (%): 66.44 -> 66.32 | No clear effect |
| increase window size from 4 to 50 (ablation) | SOBOW · 20 Newsgroup static | accuracy (%): 65.87 -> 66.32 | Improved |
| increase window size from 4 to 50 (ablation) | OAGD · 20 Newsgroup static | accuracy (%): 65.96 -> 66.44 | Improved |
| SOBOW-4 instead of OAGD-4 | dynamic online HO · 20 Newsgroup dynamic, end of 20% stream | accuracy (%): 62.61 -> 58.39 | Hurt |
| SOBOW-4 instead of OAGD-4 | dynamic online HO · 20 Newsgroup dynamic, end of 30% stream | accuracy (%): 59.26 -> 59.70 | Improved |
| SOBOW-4 instead of OAGD-4 | dynamic online HO · 20 Newsgroup dynamic | time (s): 3072 -> 1198 | Improved |
Closest papers by meaning
- Stochastic Regret Guarantees for Online Zeroth- and First-Order Bilevel Optimization (NeurIPS 2025 · similarity 0.92)
- TSP: A Two-Sided Smoothed Primal-Dual Method for Nonconvex Bilevel Optimization (ICML 2025 · similarity 0.88)
- A Nearly Optimal Single Loop Algorithm for Stochastic Bilevel Optimization under Unbounded Smoothness (ICML 2024 · similarity 0.88)
- Tuning-Free Bilevel Optimization: New Algorithms and Convergence Analysis (ICLR 2025 · similarity 0.87)
- Bilevel Optimization under Unbounded Smoothness: A New Algorithm and Convergence Analysis (ICLR 2024 · similarity 0.87)
- An Accelerated Algorithm for Stochastic Bilevel Optimization under Unbounded Smoothness (NeurIPS 2024 · similarity 0.87)
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.
