Research map / Bandits and online learning / Online convex optimization and games
Online Learning with Feedback Graphs: The True Shape of Regret
Tomáš Kocák, Alexandra Carpentier
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 |
|---|---|---|---|
| Add explicit exploration distribution over dominating sets of informative actions (Exp3-EX) to standard exponential weights (Exp3-style) | Exp3-EX on feedback graphs · Theoretical minimax regret (adversarial, feedback graphs) | regret bound order | Improved |
| Hub-graph example: sqrt(alpha T) rate replaced by delta^{1/3} T^{2/3} regret rate when T is small | Graph with one hub connected to N-1 independent vertices · Theoretical regret, T < alpha^3 | regret order | Improved |
| Use problem complexity R* (instead of sqrt(alpha T)) as the minimax rate characterization for any T | Minimax regret of online learning with feedback graphs · Theoretical, all graphs and horizons | regret lower bound order | No clear effect |
| No experiments were run; all results are theoretical bounds | n/a · n/a | n/a | No clear effect |
Closest papers by meaning
- On the Minimax Regret for Online Learning with Feedback Graphs (NeurIPS 2023 · similarity 0.93)
- Stochastic contextual bandits with graph feedback: from independence number to MAS number (NeurIPS 2024 · similarity 0.91)
- Efficient Contextual Bandits with Uninformed Feedback Graphs (ICML 2024 · similarity 0.90)
- A Simple and Adaptive Learning Rate for FTRL in Online Learning with Minimax Regret of $\Theta(T^{2/3})$ and its Application to Best-of-Both-Worlds (NeurIPS 2024 · similarity 0.89)
- On Interpolating Experts and Multi-Armed Bandits (ICML 2024 · similarity 0.89)
- Online Learning with Sublinear Best-Action Queries (NeurIPS 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.
