Research map / Bandits and online learning / Online convex optimization and games
Luckiness in Multiscale Online Learning
Wouter M Koolen, Muriel Felipe Pérez
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 |
|---|---|---|---|
| run MUSCADA on i.i.d. data satisfying Massart's condition (lambda_2 = -1/5) instead of gap-free data | MUSCADA (Tuning 1/2) · Synthetic experts, K=50, sigma_k=1/k | regret of best expert R_t,k* | Improved |
| hard stochastic data with no gap (all lambda_k = 0) | MUSCADA · Synthetic experts, K=50, sigma_k=1/k | regret growth | Hurt |
| add optimism (guess of next losses) to MUSCADA and play two instances against each other | Optimistic MUSCADA · Multiscale zero-sum game, sigma_max/100 relevant scale | iterate-average saddle-point gap | Improved |
| remove optimism from MUSCADA in multiscale games (ablation) | MUSCADA (non-optimistic) · Multiscale zero-sum game | saddle-point gap rate | Hurt |
| use multiscale-aware MUSCADA instead of single-scale algorithms in games with multiscale payoffs | Optimistic MUSCADA · Multiscale zero-sum game | convergence to saddle point | Improved |
Closest papers by meaning
- A Simple and Optimal Approach for Universal Online Learning with Gradient Variations (NeurIPS 2024 · similarity 0.87)
- Optimism Without Regularization: Constant Regret in Zero-Sum Games (NeurIPS 2025 · similarity 0.86)
- Gradient-Variation Online Learning under Generalized Smoothness (NeurIPS 2024 · similarity 0.86)
- Replicable Online Learning (NeurIPS 2025 · similarity 0.86)
- Improved Regret Bounds for Online Fair Division with Bandit Learning (AAAI 2025 · similarity 0.86)
- Online Learning in the Random-Order Model (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.
