Regret Bounds for Risk-Sensitive Reinforcement Learning

被引：0

作者：

Bastani, Osbert ^{[1
]}

Ma, Yecheng Jason ^{[1
]}

Shen, Estelle ^{[1
]}

Xu, Wanqiao ^{[2
]}

机构：

[1] Univ Penn, Philadelphia, PA 19104 USA

[2] Stanford Univ, Stanford, CA USA

来源：

ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 35, NEURIPS 2022 | 2022年

关键词：

VALUE-AT-RISK;

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

In safety-critical applications of reinforcement learning such as healthcare and robotics, it is often desirable to optimize risk-sensitive objectives that account for tail outcomes rather than expected reward. We prove the first regret bounds for reinforcement learning under a general class of risk-sensitive objectives including the popular CVaR objective. Our theory is based on a novel characterization of the CVaR objective as well as a novel optimistic MDP construction.

引用

页数：11

共 50 条

[31] Good-Deal Bounds for Option Prices under Value-at-Risk and Expected Shortfall Constraints
Desmettre, Sascha
Laudage, Christian
Sass, Joern
[J]. RISKS, 2020, 8 (04) : 1 - 22
[32] Monitoring bank risk around the world using unsupervised learning
Mercadier, Mathieu
Tarazi, Amine
Armand, Paul
Lardy, Jean-Pierre
[J]. EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2025, 324 (02) : 590 - 615
[33] A scenario optimization approach to reliability-based and risk-based design: Soft-constrained modulation of failure probability bounds
Rocchetta, Roberto
Crespo, Luis G.
[J]. RELIABILITY ENGINEERING & SYSTEM SAFETY, 2021, 216
[34] Using financial risk measures for analyzing generalization performance of machine learning models
Takeda, Akiko
Kanamori, Takafumi
[J]. NEURAL NETWORKS, 2014, 57 : 29 - 38
[35] Learning Disturbances Online for Risk-Aware Control: Risk-Aware Flight with Less Than One Minute of Data
Akella, Prithvi
Wei, Skylar X.
Burdick, Joel W.
Ames, Aaron D.
[J]. LEARNING FOR DYNAMICS AND CONTROL CONFERENCE, VOL 211, 2023, 211
[36] How Banks' Value-at-Risk Disclosures Predict their Total and Priced Risk: Effects of Bank Technical Sophistication and Learning over Time
Chi-chun Liu
Stephen G. Ryan
Hung Tan
[J]. Review of Accounting Studies, 2004, 9 : 265 - 294
[37] How banks' value-at-risk disclosures predict their total and priced risk: Effects of bank technical sophistication and learning over time
Liu, CC
Ryan, SG
Tan, H
[J]. REVIEW OF ACCOUNTING STUDIES, 2004, 9 (2-3) : 265 - 294
[38] Financial Fraud Detection Using Value-at-Risk With Machine Learning in Skewed Data
Usman, Abdullahi Ubale
Abdullahi, Sunusi Bala
Yu, Liping
Alghofaily, Bayan
Almasoud, Ahmed S.
Rehman, Amjad
[J]. IEEE ACCESS, 2024, 12 : 64285 - 64299
[39] Value at Risk Measurement Method under Deep Learning in Analysing the Excessive Financialization of Enterprises
Shao, Bin-Tao
[J]. JOURNAL OF INFORMATION SCIENCE AND ENGINEERING, 2024, 40 (04) : 849 - 863
[40] Discussion of "how banks' value-at-risk disclosures predid their total and priced risk: Effects of bank technical sophistication and learning over time"
Ke, B
[J]. REVIEW OF ACCOUNTING STUDIES, 2004, 9 (2-3) : 295 - 299

← 1 2 3 4 5 →