Research map / Bandits and online learning / RL theory and MDPs
Near-optimal Regret Bounds for Stochastic Shortest Path
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 |
|---|---|---|---|
| Replace Hoeffding-style confidence bounds with Bernstein-style confidence bounds (ablation) | SSP regret minimization · SSP with positive costs | regret upper bound: O(\sqrt{\frac{B_\star^3 |S|^2 |A| K}{\cmin}} L + \frac{B_\star^3 |S|^2 |A|}{\cmin^2} L^2) -> O(B_\star |S| \sqrt{|A| K} | Helped |
| Add a positive cost perturbation c_ε(s,a) = max{c(s,a), ε} (ablation) | Bernstein-type SSP algorithm · SSP with arbitrary costs | regret upper bound: O(B_\star^{3 / 2} |S| \sqrt{|A| K} \widetilde L + T_\star^{3 / 2} |S|^2 |A| \widetilde L^2) | Helped |
| Use Hoeffding-style rather than Bernstein-style confidence bounds (ablation) | SSP regret minimization · SSP with positive costs | regret upper bound: O(\sqrt{\frac{B_\star^3 |S|^2 |A| K}{\cmin}} L + \frac{B_\star^3 |S|^2 |A|}{\cmin^2} L^2) | Hurt |
| Remove the requirement that instantaneous costs are bounded away from zero by perturbing costs (ablation) | SSP regret minimization · SSP with arbitrary costs | regret upper bound: O(T_\star^3 |S|^{2/3} |A|^{1/3} K^{2/3} \widetilde L + T_\star^3 |S|^2 |A| \widetilde L^2); O(B_\star^{3 / 2} |S| \sqrt{ | Mixed |
Closest papers by meaning
- Minimax Regret for Stochastic Shortest Path (NeurIPS 2021 · similarity 0.95)
- Stochastic Shortest Path: Minimax, Parameter-Free and Towards Horizon-Free Regret (NeurIPS 2021 · similarity 0.95)
- Regret Lower Bounds for Decentralized Multi-Agent Stochastic Shortest Path Problems (NeurIPS 2025 · similarity 0.89)
- Minimax Optimal Regret Bound for Reinforcement Learning with Trajectory Feedback (ICML 2025 · similarity 0.88)
- Achieving Constant Regret in Linear Markov Decision Processes (NeurIPS 2024 · similarity 0.87)
- Horizon-Free Regret for Linear Markov Decision Processes (ICLR 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.