Adversarial Reinforcement Learning

By William T. B. Uther, Manuela M. Veloso. 1997

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Learning with Opponent-Learning Awareness (2017, methodology): “Detailed studies have analysed the dynamics of JALs in general sum settings: This includes work by Uther and Veloso [43] in zero-sum settings and by Claus and Boutilier [8] in cooperative settings.”
Factorized Q-learning for large-scale multi-agent systems (2018, background): “In the best-response learning paradigm, each agent just tries to learn a policy that is optimal with respect to the joint policy of the other players (Claus and Boutilier 1998; Uther and Veloso 1997).”
Multiagent learning using a variable learning rate (2002, methodology): “MG + MDP = Stochastic Game Solver Game Theory Reinforcement Learning LP TD(0) Shapley (9) Minimax-Q (14) LP TD(1) Pollatschek and Avi-Itzhak (11) – LP TD(λ) Van der Wal (11) – QP TD(0) – Hu and Wellman (15) FP TD(0) Fictitious Play (16; 11) Opponent-Modeling (17)/JALs (18)”
Opponent Modeling in Deep Reinforcement Learning (2016, methodology): “…where a model (e.g., decision trees, neural networks, Bayesian models) is built to directly predict parameters of the opponent, e.g., actions (Uther & Veloso, 2003; Ganzfried & Sandholm, 2011), private information (Billings et al., 1998b; Richards & Amir, 2007), or domain-speciﬁc strategies…”
Comparative evaluation of MAL algorithms in a diverse set of ad hoc team problems (2012, unknown): “Game theory provides a mathematically well deﬁned framework for the analysis of multiagent interactive decision mak-Appears in: Proceedings of the 11th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2012) , Conitzer, Winikoff, Padgham, and van der Hoek (eds.)”
Tree Based Discretization for Continuous State Space Reinforcement Learning (1998, background): “Uther & Veloso (1997) compare a number of reinforcement learning algorithms in the Hexcer domain .”
A Theoretical Framework for Robustness of (Deep) Classifiers against Adversarial Examples (2016, background): “…a classifier with best performance under the worst perturbation) (Dekel et al., 2010; Globerson & Roweis, 2006), exploring online and non-stationary learning (Dahlhaus, 1997; Cesa-Bianchi & Lugosi, 2006), and formalizing as an adversarial reinforcement learning problem (Uther & Veloso, 1997).”
Cooperative Multi-Agent Learning: The State of the Art (2005, background): “Benefits from modeling agents in other purely competitive settings are also reported in [45, 272].”
A Comprehensive Survey of Multiagent Reinforcement Learning (2008, unknown): “MARL can be regarded as a fusion of temporal-difference RL, game theory, and more general direct policy search techniques.”
A Deep Policy Inference Q-Network for Multi-Agent Systems (2017, methodology): “• Unlike the previous works [6, 15, 35] only focusing on competitive environments, our models are capable of handling”
Multi-agent Reinforcement Learning: An Overview (2010, background): “All agent-tracking algorithms and many agent-aware algorithms use some form of opponent modeling to keep track of the other agents’ policies [25,49,133].”