Skip to content
conftrace_
2020 NeurIPS NeurIPS 2020

Optimal Algorithms for Stochastic Multi-Armed Bandits with Heavy Tailed Rewards

Abstract

In this paper, we consider stochastic multi-armed bandits (MABs) with heavy-tailed rewards, whose p-th moment is bounded by a constant nu_p for 1

🌉 Interdisciplinary Bridge - Artificial Intelligence and Mathematics & Optimization

Authors

Kyungjae Lee, Hongjun Yang, Sungbin Lim, Songhwai Oh

Topics

Machine Learning > Optimization & Theory > Stochastic Processes Mathematics & Optimization > Optimization > Online Algorithms

Keywords

stochastic optimization online learning multi-armed bandit regret bound heavy-tailed reward
Download PDF

More from NeurIPS 2020

All NeurIPS 2020 papers →
Estimation of Skill Distribution from a Tournament
Reparameterizing Mirror Descent as Gradient Descent
General Control Functions for Causal Effect Estimation from IVs
Certified Robustness of Graph Convolution Networks for Graph Classification under Topological Attacks
Zero-Resource Knowledge-Grounded Dialogue Generation
Targeted Adversarial Perturbations for Monocular Depth Prediction