Policy and Value Transfer for Lifelong RL

Code for experimenting with transfer approaches to lifelong RL, attached to our 2018 ICML paper "Policy and Value Transfer for Lifelong Reinforcement Learning" (pdf coming soon).

Experiments require simple_rl, which can be installed with the usual:

pip install simple_rl

To reproduce all of our plots, use run_all_learning_experiments.py (Figure 3), or run_all_policy_experiments.py (Figure 2).

Authors: David Abel, Yuu Jinnai, George Konidaris, Yue Guo, Michael L. Littman.

Contact Dave or Yuu with any questions.