Transfer Learning for Reinforcement Learning Domains: A Survey

By Matthew E. Taylor, Peter Stone. 2009

2,062 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Reinforcement Learning

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

In Reinforcement Learning

Citations in the corpus

Concrete Problems in AI Safety (2016, background): “n cases where it makes sense to retain one component but not the other. This would be similar to model-based RL approaches that attempt to transfer a learned dynamics model but not the value-function [155], the novelty being the isolation of side eects rather than state dynamics as the transferrable”
Revisiting the Arcade Learning Environment: Evaluation Protocols and Open Problems for General Agents (2017, background): “Taylor and Stone (2009) have shown one can face negative transfer depending on the similarity between the tasks being used.”
Actor-Mimic: Deep Multitask and Transfer Reinforcement Learning (2015, neutral): “A wide variety of methods have also been studied in the context of RL transfer learning (see Taylor & Stone (2009) for a more comprehensive review).”
Learning World Graphs to Accelerate Hierarchical Reinforcement Learning (2019, positive): “Finally, the goal-conditioned policy can be used to further promote transfer learning on downstream tasks [86].”
Deep Reinforcement Learning for Autonomous Driving: A Survey (2020, background): “A survey on transfer learning in RL is presented in [128].”
Reinforcement Learning in Healthcare: A Survey (2019, background): “rporated in the update procedure. This modiﬁcation has been shown to make the EPO dosing faster and more efﬁcient. While transfer learning has been extensively studied in the agent learning community [46], there is quite limited work on applying TRL techniques in healthcare settings. The learning pe”
Selective Experience Replay for Lifelong Learning (2018, background): “Training on a new task receives the jumpstart (Taylor and Stone 2009) benefit of prior knowledge and the network is resistant to forgetting.”
Disentangling Transfer in Continual Reinforcement Learning (2022, methodology): “[41, 50, 42, 39] present methods on how to ﬁnd and use mappings between different domains.”
Towards Continual Reinforcement Learning: A Review and Perspectives (2020, methodology): “We refer the reader to (Taylor and Stone, 2009) for an extensive review of work focused around transfer learning in RL.”
Safe Policy Improvement with Baseline Bootstrapping (2017, background): “This scenario is typically encountered when a policy is trained in a simulator and then run in its real environment, for instance in Transfer RL (Taylor & Stone, 2009); or when a system is designed with expert knowledge and then optimized, for example in Dialogue applications (Laroche et al., 2010).”
Accelerating Reinforcement Learning with Learned Skill Priors (2020, background): “The problem of inter-task transfer has been studied for a long time in the RL community [34].”
Deep Decentralized Multi-task Multi-Agent Reinforcement Learning under Partial Observability (2017, methodology): “In single-agent, fully-observable RL, each task is formalized as a distinct MDP (i.e., MDPs and tasks are synonymous) (Taylor & Stone, 2009).”
Multiagent Learning: Basics, Challenges, and Prospects (2012, background): “Two paradigms that we consider as particularly appropriate in this respect are transfer learning (for example, Taylor and Stone [2009, 2011]) and swarm intelligence (Dorigo and Gambardella 1997, Dorigo and Stützle 2004).”
Imagination-Augmented Agents for Deep Reinforcement Learning (2017, background): “The problem of making use of partially misspeciﬁed models has attracted attention in robotics, when transferring policies from simulation to real environments [29–31].”
Successor Features for Transfer in Reinforcement Learning (2016, background): “This process of leveraging knowledge acquired in one task to improve performance on other tasks is called transfer [25, 11].”
Schema Networks: Zero-shot Transfer with a Generative Causal Model of Intuitive Physics (2017, neutral): “Schema Networks transfer experience from one scenario to other similar scenarios that exhibit repeatable structure and sub-structure (Taylor & Stone, 2009).”
Learning modular neural network policies for multi-task and multi-robot transfer (2016, background): “Robotic skill learning via reinforcement learning has been studied extensively in recent years [5], [6], [7], [8], and transfer learning in particular has been recognized for some time as an important direction in robotic learning [9], [10], [11], [12], [13], due to its potential for reducing the bu”