Research map / Bandits and online learning / Online convex optimization and games
No-Regret and Incentive-Compatible Online Learning
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 WSU instead of MWU | Full-information online learning with strategic experts · FiveThirtyEight NFL datasets | regret | No clear effect |
| Use WSU-Aggr instead of selecting a single expert (ablation) | WSU · FiveThirtyEight NFL datasets | regret | No clear effect |
| Use Hedge instead of WSU | Full-information online learning with strategic experts · FiveThirtyEight NFL datasets | regret | Helped |
| Use WSU-UX instead of EXP3 | Partial-information online learning · FiveThirtyEight NFL datasets | regret | Hurt |
| Use ELF-X instead of WSU | Full-information online learning · FiveThirtyEight NFL datasets | regret | Hurt |
| Use ELF-X-Aggr instead of WSU-Aggr (ablation) | Full-information online learning · FiveThirtyEight NFL datasets | regret | Hurt |
| Add uniform exploration to WSU-style updates | Partial-information online learning · FiveThirtyEight NFL datasets | regret | Helped |
| Use ELF-X instead of WSU during the later half after a shift in which experts are best | Full-information online learning · Monte Carlo simulation, second half of the horizon | regret | Mixed |
| Add randomized selection of event winners rather than using ELF's winner probabilities | ELF-X · Worst-case online learning | regret | Helped |
| Report truthfully at every round rather than strategically misreporting | WSU · Forward-looking experts | incentive compatibility: 0.5+0.1125η -0.00188325η^2 -> 0.5+0.112499944η-0.0018719238η^3 | Hurt |
| Replace WSU's update with the standard MWU update | Online learning with strategic experts · Full information | incentive compatibility | Hurt |
Closest papers by meaning
- Nash Incentive-compatible Online Mechanism Learning via Weakly Differentially Private Online Learning (ICML 2024 · similarity 0.90)
- Multi-Agent Imitation Learning: Value is Easy, Regret is Hard (NeurIPS 2024 · similarity 0.88)
- Regret Analysis of Repeated Delegated Choice (AAAI 2024 · similarity 0.87)
- Online Information Acquisition: Hiring Multiple Agents (ICLR 2024 · similarity 0.87)
- Online Learning in Betting Markets: Profit versus Prediction (ICML 2024 · similarity 0.87)
- Do LLM Agents Have Regret? A Case Study in Online Learning and Games (ICLR 2025 · 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.