Saved in:
Bibliographic Details
Main Authors: Sun, Chung-En, Gao, Sicun, Weng, Tsui-Wei
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2406.18062
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929400497307648
author Sun, Chung-En
Gao, Sicun
Weng, Tsui-Wei
author_facet Sun, Chung-En
Gao, Sicun
Weng, Tsui-Wei
contents Robustness remains a paramount concern in deep reinforcement learning (DRL), with randomized smoothing emerging as a key technique for enhancing this attribute. However, a notable gap exists in the performance of current smoothed DRL agents, often characterized by significantly low clean rewards and weak robustness. In response to this challenge, our study introduces innovative algorithms aimed at training effective smoothed robust DRL agents. We propose S-DQN and S-PPO, novel approaches that demonstrate remarkable improvements in clean rewards, empirical robustness, and robustness guarantee across standard RL benchmarks. Notably, our S-DQN and S-PPO agents not only significantly outperform existing smoothed agents by an average factor of $2.16\times$ under the strongest attack, but also surpass previous robustly-trained agents by an average factor of $2.13\times$. This represents a significant leap forward in the field. Furthermore, we introduce Smoothed Attack, which is $1.89\times$ more effective in decreasing the rewards of smoothed agents than existing adversarial attacks.
format Preprint
id arxiv_https___arxiv_org_abs_2406_18062
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Breaking the Barrier: Enhanced Utility and Robustness in Smoothed DRL Agents
Sun, Chung-En
Gao, Sicun
Weng, Tsui-Wei
Machine Learning
Artificial Intelligence
Robustness remains a paramount concern in deep reinforcement learning (DRL), with randomized smoothing emerging as a key technique for enhancing this attribute. However, a notable gap exists in the performance of current smoothed DRL agents, often characterized by significantly low clean rewards and weak robustness. In response to this challenge, our study introduces innovative algorithms aimed at training effective smoothed robust DRL agents. We propose S-DQN and S-PPO, novel approaches that demonstrate remarkable improvements in clean rewards, empirical robustness, and robustness guarantee across standard RL benchmarks. Notably, our S-DQN and S-PPO agents not only significantly outperform existing smoothed agents by an average factor of $2.16\times$ under the strongest attack, but also surpass previous robustly-trained agents by an average factor of $2.13\times$. This represents a significant leap forward in the field. Furthermore, we introduce Smoothed Attack, which is $1.89\times$ more effective in decreasing the rewards of smoothed agents than existing adversarial attacks.
title Breaking the Barrier: Enhanced Utility and Robustness in Smoothed DRL Agents
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2406.18062