Importance Weighted Transfer of Samples in Reinforcement Learning

Andrea Tirinzoni, Andrea Sessa, Matteo Pirotta, Marcello Restelli
Proceedings of the 35th International Conference on Machine Learning, PMLR 80:4936-4945, 2018.

Abstract

We consider the transfer of experience samples (i.e., tuples < s, a, s’, r >) in reinforcement learning (RL), collected from a set of source tasks to improve the learning process in a given target task. Most of the related approaches focus on selecting the most relevant source samples for solving the target task, but then all the transferred samples are used without considering anymore the discrepancies between the task models. In this paper, we propose a model-based technique that automatically estimates the relevance (importance weight) of each source sample for solving the target task. In the proposed approach, all the samples are transferred and used by a batch RL algorithm to solve the target task, but their contribution to the learning process is proportional to their importance weight. By extending the results for importance weighting provided in supervised learning literature, we develop a finite-sample analysis of the proposed batch RL algorithm. Furthermore, we empirically compare the proposed algorithm to state-of-the-art approaches, showing that it achieves better learning performance and is very robust to negative transfer, even when some source tasks are significantly different from the target task.

Cite this Paper


BibTeX
@InProceedings{pmlr-v80-tirinzoni18a, title = {Importance Weighted Transfer of Samples in Reinforcement Learning}, author = {Tirinzoni, Andrea and Sessa, Andrea and Pirotta, Matteo and Restelli, Marcello}, booktitle = {Proceedings of the 35th International Conference on Machine Learning}, pages = {4936--4945}, year = {2018}, editor = {Dy, Jennifer and Krause, Andreas}, volume = {80}, series = {Proceedings of Machine Learning Research}, month = {10--15 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v80/tirinzoni18a/tirinzoni18a.pdf}, url = {https://proceedings.mlr.press/v80/tirinzoni18a.html}, abstract = {We consider the transfer of experience samples (i.e., tuples < s, a, s’, r >) in reinforcement learning (RL), collected from a set of source tasks to improve the learning process in a given target task. Most of the related approaches focus on selecting the most relevant source samples for solving the target task, but then all the transferred samples are used without considering anymore the discrepancies between the task models. In this paper, we propose a model-based technique that automatically estimates the relevance (importance weight) of each source sample for solving the target task. In the proposed approach, all the samples are transferred and used by a batch RL algorithm to solve the target task, but their contribution to the learning process is proportional to their importance weight. By extending the results for importance weighting provided in supervised learning literature, we develop a finite-sample analysis of the proposed batch RL algorithm. Furthermore, we empirically compare the proposed algorithm to state-of-the-art approaches, showing that it achieves better learning performance and is very robust to negative transfer, even when some source tasks are significantly different from the target task.} }
Endnote
%0 Conference Paper %T Importance Weighted Transfer of Samples in Reinforcement Learning %A Andrea Tirinzoni %A Andrea Sessa %A Matteo Pirotta %A Marcello Restelli %B Proceedings of the 35th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2018 %E Jennifer Dy %E Andreas Krause %F pmlr-v80-tirinzoni18a %I PMLR %P 4936--4945 %U https://proceedings.mlr.press/v80/tirinzoni18a.html %V 80 %X We consider the transfer of experience samples (i.e., tuples < s, a, s’, r >) in reinforcement learning (RL), collected from a set of source tasks to improve the learning process in a given target task. Most of the related approaches focus on selecting the most relevant source samples for solving the target task, but then all the transferred samples are used without considering anymore the discrepancies between the task models. In this paper, we propose a model-based technique that automatically estimates the relevance (importance weight) of each source sample for solving the target task. In the proposed approach, all the samples are transferred and used by a batch RL algorithm to solve the target task, but their contribution to the learning process is proportional to their importance weight. By extending the results for importance weighting provided in supervised learning literature, we develop a finite-sample analysis of the proposed batch RL algorithm. Furthermore, we empirically compare the proposed algorithm to state-of-the-art approaches, showing that it achieves better learning performance and is very robust to negative transfer, even when some source tasks are significantly different from the target task.
APA
Tirinzoni, A., Sessa, A., Pirotta, M. & Restelli, M.. (2018). Importance Weighted Transfer of Samples in Reinforcement Learning. Proceedings of the 35th International Conference on Machine Learning, in Proceedings of Machine Learning Research 80:4936-4945 Available from https://proceedings.mlr.press/v80/tirinzoni18a.html.

Related Material