Transferring Instances for Model-Based Reinforcement Learning

By Matthew E. Taylor, Nicholas K. Jong, Peter Stone. 2008

150 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.

In Artificial Intelligence

Citations in the corpus

Learning Invariant Feature Spaces to Transfer Skills with Reinforcement Learning (2017, methodology): “For instance, Taylor et al. (2008) ﬁnd a mapping between state spaces by searching through all possible pairings.”
Behavioral Cloning from Observation (2018, neutral): “Second, the learned models can be transferred across tasks [ Taylor et al. , 2008 ] .”
Transfer in Reinforcement Learning: A Framework and a Survey (2012, background): “For instance, in (Taylor et al, 2008b) target samples are used to identify the best mapping between source and target state-action variables and thus to transform the source policy into a target policy used to initialize the learning process in the target task.”
Transfer of samples in batch reinforcement learning (2008, methodology): “Finally, a method for mapping samples from a source to a target task with different state-action space is proposed in (Taylor et al., 2008).”
An automated measure of MDP similarity for transfer in reinforcement learning (2014, background): “Recent work on autonomous transfer (Taylor, Jong, and Stone 2008; Ammar et al. 2012; 2013) has focused on the latter two items, leaving the automated selection of relevant source tasks (item a) as a largely unsolved problem, yet a critical one for successful transfer learning.”
Combining manual feedback with subsequent MDP reward signals for reinforcement learning (2010, methodology): “Such mappings have been used in the target task to initialize the Q function [18], as we did in Method 3, and to use recorded (s, a, r, s′) experience tuples from the source task to estimate the transition and reward functions for model-based RL [16].”
Deep Transfer in Reinforcement Learning by Language Grounding (2017, unknown): “There have been a few attempts at model-based transfer for RL (Taylor et al., 2008; Nguyen et al., 2012; Gaˇsic et al., 2013; Wang et al., 2015) but these methods either rely on hand-coded inter-task mappings for state and actions spaces or require signiﬁcant interactions in the target task to…”
Transfer Learning for Reinforcement Learning Domains: A Survey (2009, background): “…10Years Later,” (Silver et al., 2005) had few RL-related transfer papers, the 2006 ICML workshop, “Structu al Knowledge Transfer for Machine Learning,” (Banerjee et al., 2006) had many, and the 2008 AAAI workshop, “Transfer Learning for Complex Tasks,” (Taylor et al., 2008a) focused on RL.”
Transfer learning for reinforcement learning on a physical robot (2010, methodology): “[17] using a model-based approach to reduce the samples needed, and they transfer observed (s, a, r, s′) instances, which allows the source and target agents to have different representations for the task.”