Meta-learning in Reinforcement Learning

  title={Meta-learning in Reinforcement Learning},
  author={Nicolas Schweighofer and Kenji Doya},
  journal={Neural networks : the official journal of the International Neural Network Society},
  volume={16 1},
Meta-parameters in reinforcement learning should be tuned to the environmental dynamics and the animal performance. Here, we propose a biologically plausible meta-reinforcement learning algorithm for tuning these meta-parameters in a dynamic, adaptive manner. We tested our algorithm in both a simulation of a Markov decision task and in a non-linear control task. Our results show that the algorithm robustly finds appropriate meta-parameter values, and controls the meta-parameter time course, in… CONTINUE READING


Publications citing this paper.
Showing 1-10 of 73 extracted citations


Publications referenced by this paper.
Showing 1-10 of 17 references

Functional MRI study of short-term and long-term prediction of reward

  • S. Tanaka, K. Doya, G. Okada, K. Ueda, Y. Okamoto, S. Yamawaki
  • Proceedings of the Eighth International…
  • 2002
1 Excerpt

On the complexity of solving Markov decision problems

  • M. L. Littman, Dean, L T.
  • Eleventh International Conference on Uncertainty…
  • 1995
1 Excerpt

Similar Papers

Loading similar papers…