dc.contributor.author | Oberst, Michael | |
dc.contributor.author | Sontag, David Alexander | |
dc.date.accessioned | 2021-04-09T20:45:12Z | |
dc.date.available | 2021-04-09T20:45:12Z | |
dc.date.issued | 2019-06 | |
dc.identifier.uri | https://hdl.handle.net/1721.1/130437 | |
dc.description.abstract | We introduce an off-policy evaluation procedure for highlighting episodes where applying a reinforcement learned (RL) policy is likely to have produced a substantially different outcome than the observed policy. In particular, we introduce a class of structural causal models (SCMs) for generating counterfactual trajectories in finite partially observable Markov Decision Processes (POMDPs). We see this as a useful procedure for off-policy "debugging" in high-risk settings (e.g., healthcare); by decomposing the expected difference in reward between the RL and observed policy into specific episodes, we can identify episodes where the counterfactual difference in reward is most dramatic. This in turn can be used to facilitate review of specific episodes by domain experts. We demonstrate the utility of this procedure with a synthetic environment of sepsis management. | en_US |
dc.language.iso | en | |
dc.publisher | MLResearch Press | en_US |
dc.relation.isversionof | http://proceedings.mlr.press/v97/ | en_US |
dc.rights | Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use. | en_US |
dc.source | Proceedings of Machine Learning Research | en_US |
dc.title | Counterfactual off-policy evaluation with gumbel-max structural causal models | en_US |
dc.type | Article | en_US |
dc.identifier.citation | Oberst, Michael and David Sontag. "Counterfactual off-policy evaluation with gumbel-max structural causal models." Proceedings of the 36th International Conference on Machine Learning, June 2019, Long Beach, California, MLResearch Press, 2019. | en_US |
dc.contributor.department | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory | en_US |
dc.relation.journal | Proceedings of the 36th International Conference on Machine Learning | en_US |
dc.eprint.version | Final published version | en_US |
dc.type.uri | http://purl.org/eprint/type/ConferencePaper | en_US |
eprint.status | http://purl.org/eprint/status/NonPeerReviewed | en_US |
dc.date.updated | 2021-04-06T18:37:22Z | |
dspace.orderedauthors | Oberst, M; Sontag, D | en_US |
dspace.date.submission | 2021-04-06T18:37:23Z | |
mit.license | PUBLISHER_POLICY | |
mit.metadata.status | Authority Work and Publication Information Needed | |