A Simple and Optimal Policy Design for Online Learning with Safety against Heavy-tailed Risk.

David Simchi-Levi Zeyu Zheng Feng Zhu

Published in: NeurIPS (2022)

Keyphrases

online learning
optimal policy
heavy tailed
markov decision processes
infinite horizon
finite horizon
state dependent
develop a mathematical model
multistage
reinforcement learning
long run
dynamic programming
multiscale
state space