Research map / Bandits and online learning / Online convex optimization and games
Offline Congestion Games: How Feedback Type Affects Data Coverage Requirement
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 facility-level feedback with one-unit deviation coverage and pessimism-style surrogate minimization with facility-wise UCB bonuses | offline congestion games · offline congestion games, facility-level feedback | approximate NE gap bound | Helped |
| Weaken facility-level coverage below one-unit deviation (allow one missing configuration) (ablation) | offline congestion games, facility-level feedback · hard instance class with one uncovered facility configuration | NE approximation achievable | Hurt |
| Switch from facility-level to agent-level (bandit) feedback while keeping one-unit deviation coverage (ablation) | offline congestion games, agent-level feedback · hard instance with two facilities and two players | approximate NE achievable | Hurt |
| Use linear bandit reformulation with ridge regression and weak covariance domination coverage under agent-level feedback | offline congestion games, agent-level feedback · agent-level feedback | approximate NE gap bound | Helped |
| Switch from agent-level to game-level (total reward) feedback under strong covariance domination (ablation) | offline congestion games, game-level feedback · hard instance with two facilities and two players | approximate NE achievable | Hurt |
| Use linear bandit with total-reward regression and strong covariance domination coverage under game-level feedback | offline congestion games, game-level feedback · game-level feedback | approximate NE gap bound | Helped |
Closest papers by meaning
- Learning in Congestion Games with Bandit Feedback (NeurIPS 2022 · similarity 0.89)
- Semi Bandit dynamics in Congestion Games: Convergence to Nash Equilibrium and No-Regret Guarantees. (ICML 2023 · similarity 0.89)
- Offline Learning in Markov Games with General Function Approximation (ICML 2023 · similarity 0.88)
- Learning Optimal Tax Design in Nonatomic Congestion Games (NeurIPS 2024 · similarity 0.86)
- The Harder Path: Last Iterate Convergence for Uncoupled Learning in Zero-Sum Games with Bandit Feedback (ICML 2025 · similarity 0.85)
- The Complexity of Optimizing Atomic Congestion (AAAI 2024 · 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.