General Game Learning Using Knowledge Transfer

By Bikramjit Banerjee, Peter Stone. 2007

133 citations recorded by the source corpus.

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

Multi-agent reinforcement learning as a rehearsal for decentralized planning (2016, neutral): “In particular, value function transfer [5] has been used in reinforcement learning to bootstrap an agent's value function in a related task, by initializing it with learned values from a previous task.”
Heterogeneous Domain Adaptation for Multiple Classes (2014, methodology): “have been successfully applied to various real-world applications, such as WiFi-based localization [23, 24], image classification [35], video concept detection [14], sentiment analysis [5, 28], coreference resolution [32], vehicle routing [18], game playing [4], etc.”
Simulation-Based Approach to General Game Playing (2008, methodology): “Another strong GGP agent using the traditional method is UTEXAS LARG (Kuhlmann, Dresner, & Stone 2006) and its authors, among others, research knowledge transfers between GGP games (Banerjee, Kuhlmann, & Stone 2006; Banerjee & Stone 2007).”
Unsupervised Cross-Domain Transfer in Policy Gradient Reinforcement Learning via Manifold Alignment (2015, background): “However, the majority of existing work assumes that a) the source task and target task are similar enough that no mapping is needed (Banerjee & Stone 2007; Konidaris & Barto 2007), or b) an inter-task mapping is provided to the agent (Taylor et al. 2007a; Torrey et al. 2008).”
Transfer in Reinforcement Learning via Shared Features (2012, background): “Banerjee and Stone (2007) consider transfer learning for the case of General Game Playing, where knowledge gained from playing one game (e.”
Transfer in Reinforcement Learning: A Framework and a Survey (2012, background): “Banerjee and Stone (2007) also consider the transfer of value functions in the context of general games where different games can be represented by a common abstract structure.”
Transfer Learning through Analogy in Games (2011, methodology): “For example, Peter Stone’s group has focused on reinforcement learning and transferring evaluation functions (Banerjee and Stone 2007), including using their own version of SME for near transfer on domain pairs of their own choosing (Liu and Stone 2006).”
Transfer Learning for Reinforcement Learning Domains: A Survey (2009, background): “• First, although two of the papers (Banerjee and Stone, 2007; Kuhlmann and Stone, 2007) in this survey have examined extensive games, none consider repeated normal form games or stochastic games (Shapley, 1953).”
Actor-Mimic: Deep Multitask and Transfer Reinforcement Learning (2015, background): “Banerjee & Stone (2007) constructed agent-space features by ﬁrst generating a ﬁxed-depth game tree of the current state, classifying each future state in the tree as either { win, lose, draw, nonterminal } and then coalescing all states which have the same class or subtree.”
Decoupling Dynamics and Reward for Transfer Learning (2018, methodology): “More speciﬁcally, we look at other model-based RL meth-ods (Finn & Levine, 2016; Watter et al., 2015) that try to transfer and plan (Banerjee & Stone, 2007; Xie et al., 2015; Christiano et al., 2016).”