The Theory of Learning in Games

By Drew Fudenberg, David K. Levine. 1998

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

Deep Reinforcement Learning from Self-Play in Imperfect-Information Games (2016, background): “For a more detailed exposition we refer the reader to (Sutton and Barto, 1998), (Myerson, 1991), (Fudenberg, 1998) and (Heinrich et al., 2015).”
Multiagent learning using a variable learning rate (2002, background): “The framework of stochastic games helps provide a model for this learning problem and also illucidates the difﬁculties of attempting to learn an essentially moving target.”
Emergent Communication through Negotiation (2018, background): “More sophisticated players will often have an incentive to take a short-term loss in reward to teach or punish another player for long-term gain in the repeated setting (Fudenberg & Levine, 1998).”
The Evolutionary Foundations of Preferences (2011, background): “3See, for example, Fudenberg and Levine [49], Hofbauer and Sigmund [65], Mailath [80], Samuelson [119], van Damme [139, Chapter 9], Vega Redondo [141], Weibull [143] and Young [150].”
Dispersion games: general definitions and some specific learning results (2002, methodology): “The learning rules we evaluated above are an initial exploration, and clearly many other learning techniques also deserve consideration.”
The Expected Number of Nash Equilibria of a Normal Form Game (2005, background): “(11) This manifold has the following description.”
Survival of dominated strategies under evolutionary dynamics (2011, methodology): “Hofbauer and Swinkels (1996) and Hart and Mas-Colell (2003) show that no reasonable evolutionary dynamic converges to Nash equilibrium in all games: there are some games in which cycling or more complicated limit behavior far from any Nash equilibrium is the only plausible long run prediction.”
Joint Strategy Fictitious Play with Inertia for Potential Games (2005, methodology): “See the monographs [1], [11], [12], [13], [14] and survey article [15] for a more comprehensive review.”
Invariant Risk Minimization Games (2020, methodology): “This brings us to the question: how do we solve the game? We use classic best response dynamics [4], which has a very simple implementation.”
Generalised weakened fictitious play (2006, background): “…β i τ π −i = argmax π i ∈∆ i r i π i , π −i + τ v i π i where τ > 0 is a temperature parameter, and the function v i is a smooth, strictly differentiably concave function such that as π i approaches the boundary of ∆ i the slope of v i becomes infinite (Fudenberg and Levine, 1998, Chapter 4).”
Robust incentive techniques for peer-to-peer networks (2004, methodology): “In particular, we use a prisoners’ dilemma model to capture the essential tension between individual and social utility, asymmetric payoff matrices to allow asymmetric transactions between peers, and a learning-based [14] population dynamic model to specify the behavior of individual peers, which ca”
Learning to compete, coordinate, and cooperate in repeated games using reinforcement learning (2011, background): “Player i satisﬁes this property if it is ( G 1 , Ω , 1 − ε )-efﬁcient, where ε > 0 is small.”
Dissociable neural representations of reinforcement and belief prediction errors underlie strategic learning (2012, methodology): “D in the presence of competitive intelligent agents is fundamental for social and economic behavior (1, 2).”