Summary of Robust Deep Reinforcement Learning Through Adversarial Attacks and Training : a Survey, by Lucas Schott et al.
Robust Deep Reinforcement Learning Through Adversarial Attacks and Training : A Survey
by Lucas Schott, Josephine Delas, Hatem Hajri, Elies Gherbi, Reda Yaich, Nora Boulahia-Cuppens, Frederic Cuppens, Sylvain Lamprier
First submitted to arxiv on: 1 Mar 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A novel approach to deep reinforcement learning (DRL) is proposed to enhance the trustworthiness and robustness of autonomous agents. Current DRL methods excel in controlled environments but struggle with minor variations, casting doubt on their reliability for real-world applications. To address this issue, Adversarial Training is employed, where an agent is trained against well-suited attacks on observations and environmental dynamics. Our work provides a comprehensive analysis of contemporary adversarial attack and training methodologies, categorizing and comparing objectives and operational mechanisms. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Deep Reinforcement Learning (DRL) helps robots make smart decisions. But what happens when things change in the environment? Right now, DRL can struggle with small changes. To fix this, we’re going to train our agents against fake attacks on what they see and how they interact with their surroundings. This will help them be more reliable in real-life situations. |
Keywords
* Artificial intelligence * Reinforcement learning
