A Reinforcement Procedure Leading to Correlated Equilibrium

By Sergiu Hart, Andreu Mas-Colell. 2001

178 citations recorded by the source corpus.

Subfields: Economics and Econometrics, Management Science and Operations Research, Safety Research

View Paper: https://www.semanticscholar.org/paper/610511bf02cfbe219053f3296240a5d56a56f9f1

How researchers see this paper

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

In Economics and Econometrics

In Management Science and Operations Research

In Safety Research

Citations in the corpus

Improved second-order bounds for prediction with expert advice (2005, background): “Technically, this is important since it is common to divide by the probabilities (see [3, 6, 7, 10, 12]).”
Game Theory and Multi-agent Reinforcement Learning (2012, background): “Under some assumptions regret based learning can been shown to converge to some form of equilibrium play (Foster and Young, 2003; Hart and Mas-Colell, 2001).”
Game Theory in Wireless and Communication Networks (2008, methodology): “To overcome this limitation, the modified regret-tracking algorithm [197] estimates the decisions which have not been made, so the secondary user can make an approximately optimal decision.”
Stochastic Approximations and Differential Inclusions (2005, methodology): “Both classical fictitious play (Brown [10], Robinson [ 26 ]) and stochastic fictitious play (Bena¨ im and Hirsch [6], Fudenberg and Levine [15], Hofbauer and Sandholm [20]) assume that the strategy of player i, Qi = {Qi x}, can be written as”
R-MAX - A General Polynomial Time Algorithm for Near-Optimal Reinforcement Learning (2001, methodology): “A more recent algorithm by Hart and Mas-Colell (2000) features an algorithm that is much simpler than the algorithms by Banos and by Megiddo.”
A Unified Game-Theoretic Approach to Multiagent Reinforcement Learning (2017, unknown): “We try three: decoupled regret-matching [33], Exp3 (decoupled Hedge) [3], and decoupled PRD.”
Prediction, learning, and games (2006, methodology): “5 appears in Hart and Mas-Colell [147].”
Comparative evaluation of MAL algorithms in a diverse set of ad hoc team problems (2012, background): “First, we note that most MAL algorithms (e.g. [ 1,3,5,6,8,14,17,23,32]) were primarily evaluated in homogeneous groups of agents, meaning that all agents in the group were identical, and all agents were a priori aware of this fact.”
Generalised weakened fictitious play (2006, background): “Furthermore, they either prove that these averages converge only to correlated equilibria (Hart and Mas-Colell, 2001), or prove convergence only in a very weak sense, essentially due to the fact that if strategies vary randomly then at some point they must get close to a Nash equilibrium (Foster…”
Multi-Agent Reinforcement Learning: A Selective Overview of Theories and Algorithms (2019, methodology): “No-regret algorithms with asymptotically zero average regret guarantee the convergence to the equilibria of certain games (Hart and Mas-Colell, 2001; Bowling, 2005; Zinkevich et al., 2008), which essentially guarantee that the agent is not exploited by others.”
From External to Internal Regret (2005, unknown): “(The work of Hart and Mas-Colell (2001) also gives a multiarmed bandit algorithm whose internal regret is sublinear in T , but does not derive explicit bounds.)”
Autonomous agents modelling other agents: A comprehensive survey and open problems (2017, background): “However, MAL may also involve other types of learning, such as learning to coordinate without constructing models of other agents (e.g. Albrecht and Ramamoorthy, 2012; Bowling and Veloso, 2002; Hart and Mas-Colell, 2001) and learning based on communication.”
Efficient learning equilibrium (2004, background): “, [15] and in [10] (where special attention is given to the issue of efficiency).”