Research map / Bandits and online learning / Online convex optimization and games
Online Prediction with Selfish Experts
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 the standard absolute-loss update with a spherical proper scoring rule | WM with selfish experts · Absolute loss, worst case | loss ratio to best expert: 4 -> 2+√2 | Helped |
| Use Brier updates with θ=0.382 in θ-RWM | Randomized weighted majority · Absolute loss, worst case | loss ratio to best expert: 2.62 | Helped |
| Use proper scoring-rule updates to elicit truthful expert beliefs | Online prediction with selfish experts · HMM | loss ratio to best expert: 1.14 -> about 1.003 | Helped |
| Use the standard absolute-loss update instead of an incentive-compatible update | WM with selfish experts · Absolute loss, worst case | loss ratio to best expert: 2+√2 -> 4 | Hurt |
| Use quadratic loss with the standard WM update | Standard WM · Quadratic loss | incentive compatibility | Helped |
| Switch among the tested IC scoring rules on the HMM data (ablation) | IC randomized weighted majority · HMM | loss ratio to best expert: around 1.02 | No clear effect |
| Choose a spherical scoring rule rather than Brier or the tested Beta rules (ablation) | IC weighted majority · Greedy lower-bound instance | loss ratio to best expert: 2.3708 -> 2.2071 | Helped |
Closest papers by meaning
- Nash Incentive-compatible Online Mechanism Learning via Weakly Differentially Private Online Learning (ICML 2024 · similarity 0.86)
- Private Online Learning via Lazy Algorithms (NeurIPS 2024 · similarity 0.85)
- Online Mixture of Experts: No-Regret Learning for Optimal Collective Decision-Making (NeurIPS 2025 · similarity 0.85)
- Tracking The Best Expert Privately (ICML 2025 · similarity 0.85)
- Federated Online Prediction from Experts with Differential Privacy: Separations and Regret Speed-ups (NeurIPS 2024 · similarity 0.85)
- Stochastically Dominant Peer Prediction (NeurIPS 2025 · similarity 0.84)
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.