Research map / Bandits and online learning / Online convex optimization and games
Tracking the Best Expert in Non-stationary Stochastic Environments
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 full-information GD-based algorithm (Algorithm 2) with learning rate η=sqrt(Γ/(Λ+KΓ)) for switching distributions | full-information switching distributions · switching distributions (Γ, Λ) | expected dynamic regret | Helped |
| Bandit regret lower bound for switching distributions even with zero variance (Λ=0) and constant gaps | bandit algorithms, switching distributions · switching distributions with Γ≥2 | worst-case expected regret | No clear effect |
| Use UCB-V-based restarted bandit algorithm (Rerun-UCB-V) with variance-aware confidence intervals in blocks of size B | bandit drifting distributions · drifting distributions (V, Λ) | expected dynamic regret | Helped |
| Add a constant-regret-at-interval-boundary fix: reset learning rate per interval but use cumulative past deviations ||ℓτ−ℓτ−1||² instead of a fresh reset | full-information drifting distributions algorithm · drifting distributions (V, Λ) | expected dynamic regret | Helped |
| Parameter-free Optimistic-Adapt-ML-Prod with sleeping experts reduction (no knowledge of Γ, V, Λ) | full-information switching and drifting distributions · switching (Γ,Λ) and drifting (V,Λ) distributions | expected dynamic regret | No clear effect |
Closest papers by meaning
- Non-stationary Bandit Convex Optimization: A Comprehensive Study (NeurIPS 2025 · similarity 0.90)
- Tracking Most Significant Shifts in Infinite-Armed Bandits (ICML 2025 · similarity 0.90)
- Tracking The Best Expert Privately (ICML 2025 · similarity 0.89)
- Non-Stationary Lipschitz Bandits (NeurIPS 2025 · similarity 0.88)
- Constrained Feedback Learning for Non-Stationary Multi-Armed Bandits (NeurIPS 2025 · similarity 0.88)
- Adaptive Regret for Bandits Made Possible: Two Queries Suffice (ICLR 2024 · similarity 0.88)
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.