English

Efficient Competitions and Online Learning with Strategic Forecasters

Machine Learning 2021-06-14 v2 Computer Science and Game Theory

Abstract

Winner-take-all competitions in forecasting and machine-learning suffer from distorted incentives. Witkowski et al. 2018 identified this problem and proposed ELF, a truthful mechanism to select a winner. We show that, from a pool of nn forecasters, ELF requires Θ(nlogn)\Theta(n\log n) events or test data points to select a near-optimal forecaster with high probability. We then show that standard online learning algorithms select an ϵ\epsilon-optimal forecaster using only O(log(n)/ϵ2)O(\log(n) / \epsilon^2) events, by way of a strong approximate-truthfulness guarantee. This bound matches the best possible even in the nonstrategic setting. We then apply these mechanisms to obtain the first no-regret guarantee for non-myopic strategic experts.

Keywords

Cite

@article{arxiv.2102.08358,
  title  = {Efficient Competitions and Online Learning with Strategic Forecasters},
  author = {Rafael Frongillo and Robert Gomez and Anish Thilagar and Bo Waggoner},
  journal= {arXiv preprint arXiv:2102.08358},
  year   = {2021}
}

Comments

This paper will be presented at The Twenty-Second ACM Conference on Economics and Computation (EC '21), July 18-23, 2021, Budapest, Hungary

R2 v1 2026-06-23T23:13:24.244Z