Sample-Efficient Reinforcement Learning through Transfer and Architectural Priors

Recent work in deep reinforcement learning has allowed algorithms to learn complex tasks such as Atari 2600 games just from the reward provided by the game, but these algorithms presently require millions of training steps in order to learn, making them approximately five orders of magnitude slower than humans. One reason for this is that humans build robust shared representations that are applicable to collections of problems, making it much easier to assimilate new variants. This paper first introduces the idea of automatically-generated game sets to aid in transfer learning research, and then demonstrates the utility of shared representations by showing that models can substantially benefit from the incorporation of relevant architectural priors. This technique affords a remarkable 50x positive transfer on a toy problem-set.

[1]  Peter Dayan,et al.  Q-learning , 1992, Machine Learning.

[2]  Qiang Yang,et al.  A Survey on Transfer Learning , 2010, IEEE Transactions on Knowledge and Data Engineering.

[3]  Pieter Abbeel,et al.  Value Iteration Networks , 2016, NIPS.

[4]  Joshua B. Tenenbaum,et al.  Building machines that learn and think like people , 2016, Behavioral and Brain Sciences.

[5]  David Silver,et al.  Deep Reinforcement Learning with Double Q-Learning , 2015, AAAI.

[6]  Peter Stone,et al.  Transfer Learning for Reinforcement Learning Domains: A Survey , 2009, J. Mach. Learn. Res..

[7]  Alex Graves,et al.  Playing Atari with Deep Reinforcement Learning , 2013, ArXiv.

[8]  Jürgen Schmidhuber,et al.  Discovering Neural Nets with Low Kolmogorov Complexity and High Generalization Capability , 1997, Neural Networks.

[9]  Wojciech Zaremba,et al.  OpenAI Gym , 2016, ArXiv.

[10]  David H. Wolpert,et al.  No free lunch theorems for optimization , 1997, IEEE Trans. Evol. Comput..

[11]  Chrisantha Fernando,et al.  PathNet: Evolution Channels Gradient Descent in Super Neural Networks , 2017, ArXiv.

[12]  Matthew E. Taylor,et al.  Initial Progress in Transfer for Deep Reinforcement Learning Algorithms , 2016 .

[13]  Qiang Yang,et al.  Transfer Learning via Dimensionality Reduction , 2008, AAAI.

[14]  Jude Shavlik,et al.  Chapter 11 Transfer Learning , 2009 .

[15]  Ann L. Brown,et al.  Preschool children can learn to transfer: Learning to learn and learning from example , 1988, Cognitive Psychology.

[16]  S. Mahadevan,et al.  Proto-transfer Learning in Markov Decision Processes Using Spectral Methods , 2006 .