Adversarial Reinforcement Learning
نویسنده
چکیده
Reinforcement Learning has been used for a number of years in single agent environments. This article reports on our investigation of Reinforcement Learning techniques in a multi-agent and adversarial environment with continuous observable state information. We introduce a new framework, two-player hexagonal grid soccer, in which to evaluate algorithms. We then compare the performance of several single-agent Reinforcement Learning techniques in that environment. These are further compared to a previously developed adversarial Reinforcement Learning algorithm designed for Markov games. Building upon these efforts, we introduce new algorithms to handle the multi-agent, the adversarial, and the continuous-valued aspects of the domain. We introduce a technique for modelling the opponent in an adversarial game. We introduce an extension to Prioritized Sweeping that allows generalization of learnt knowledge over neighboring states in the domain; and we introduce an extension to the U Tree generalizing algorithm that allows the handling of continuous state spaces. Extensive empirical evaluation is conducted in the grid soccer domain.
منابع مشابه
Delving into adversarial attacks on deep policies
Adversarial examples have been shown to exist for a variety of deep learning architectures. Deep reinforcement learning has shown promising results on training agent policies directly on raw inputs such as image pixels. In this paper we present a novel study into adversarial attacks on deep reinforcement learning polices. We compare the effectiveness of the attacks using adversarial examples vs...
متن کاملGenerative Adversarial Imitation Learning
Consider learning a policy from example expert behavior, without interaction with the expert or access to reinforcement signal. One approach is to recover the expert’s cost function with inverse reinforcement learning, then extract a policy from that cost function with reinforcement learning. This approach is indirect and can be slow. We propose a new general framework for directly extracting a...
متن کاملVulnerability of Deep Reinforcement Learning to Policy Induction Attacks
Deep learning classifiers are known to be inherently vulnerable to manipulation by intentionally perturbed inputs, named adversarial examples. In this work, we establish that reinforcement learning techniques based on Deep Q-Networks (DQNs) are also vulnerable to adversarial input perturbations, and verify the transferability of adversarial examples across different DQN models. Furthermore, we ...
متن کاملGeneralizing Adversarial Reinforcement Learning
Reinforcement Learning has been used for a number of years in single agent environments. This article reports on our investigation of Reinforcement Learning techniques in a multi-agent and adversarial environment with continuous observable state information. Our framework for evaluating algorithms is two-player hexagonal grid soccer. We introduce an extension to Prioritized Sweeping that allows...
متن کاملDANCin SEQ2SEQ: Fooling Text Classifiers with Adversarial Text Example Generation
Machine learning models are powerful but fallible. Generating adversarial examples inputs deliberately crafted to cause model misclassification or other errors can yield important insight into model assumptions and vulnerabilities. Despite significant recent work on adversarial example generation targeting image classifiers, relatively little work exists exploring adversarial example generation...
متن کامل