An Introduction to Intertask Transfer for Reinforcement Learning

By Matthew E. Taylor, Peter Stone. 2011

68 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Management Science and Operations Research

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 Management Science and Operations Research

Citations in the corpus

PathNet: Evolution Channels Gradient Descent in Super Neural Networks (2017, methodology): “The framework for AGI described above includes aspects of transfer learning [22], continual learning [16] and multitask learning [2].”
Progress & Compress: A scalable framework for continual learning (2018, background): “Slightly different than continual learning, different aspects of transfer learning for reinforcement learning are discussed and compared in Taylor & Stone (2011).”
Distral: Robust multitask reinforcement learning (2017, background): “e would expect that individual tasks should require less data or achieve a higher asymptotic performance. Indeed this intuition has long been pursued in the multitask and transfer-learning literature [3, 27, 30, 6]. Somewhat counter-intuitively, however, the above is often not the result encountered”
Deep reinforcement learning with successor features for navigation across similar environments (2016, background): “Historically, most existing work in this direction relied on simple task models and explicitly known relations between tasks or known dynamics [16], [17], [18].”
Progressive Neural Networks (2016, background): “2 Progressive Networks Continual learning is a long-standing goal of machine learning, where agents not only learn (and remember) a series of tasks experienced in sequence, but also have the ability to transfer knowledge from previous tasks to improve convergence speed [20].”
Continual and Multi-task Reinforcement Learning With Shared Episodic Memory (2019, background): “Weight sharing allows to generalize experience over tasks and facilitates the learning of individual tasks (Taylor & Stone, 2011).”
Measuring the Distance Between Finite Markov Decision Processes (2016, background): “If the source task is irrelevant to the target task, transfer may cause slower learning speed or even worse performance in the target task; this phenomenon has been recognized as negative transfer [18, 24, 27, 28].”
Probabilistically safe policy transfer (2017, background): “A number of recent papers have explored how to transfer policies from one domain to another [5], [6], [7], [11], [20], [24], [28], [29].”