2026-09-19T01:19:03Zhttps://dspace.mit.edu/server/oai/requestoai:dspace.mit.edu:1721.1/1589342025-04-08T04:08:56Zcom_1721.1_7582com_1721.1_7581col_1721.1_131023
Daskalakis, Constantinos
Rakhlin, Alexander
Chen, Fan
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
2025-03-27T16:58:58Z
2025-03-27T16:58:58Z
2025-02
2025-03-04T17:27:24.828Z
https://hdl.handle.net/1721.1/158934
We study computational and statistical aspects of learning Latent Markov Decision Processes (LMDPs). In this model, the learner interacts with an MDP drawn at the beginning of each epoch from an unknown mixture of MDPs. To sidestep known impossibility results, we consider several notions of δ-separation of the constituent MDPs. The main thrust of this paper is in establishing a nearly-sharp statistical threshold for the horizon length necessary for efficient learning. On the computational side, we show that under a weaker assumption of separability under the optimal policy, there is a quasi-polynomial algorithm with time complexity scaling in terms of the statistical threshold. We further show a near-matching time complexity lower bound under the exponential time hypothesis.
S.M.
Massachusetts Institute of Technology
In Copyright - Educational Use Permitted
Copyright retained by author(s)
https://rightsstatements.org/page/InC-EDU/1.0/
Near-Optimal Learning and Planning in Separated Latent MDPs
Thesis
application/pdf
Master
Master of Science in Electrical Engineering and Computer Science
Publication
unknown
unknown
<Publication xmlns="https://www.openaire.eu/cerif-profile/1.1/" id="584e734d-67e1-4754-b93b-1ad5522b76e3">
<Type xmlns="https://www.openaire.eu/cerif-profile/vocab/COAR_Publication_Types">http://purl.org/coar/resource_type/c_1843</Type>
<Title>Near-Optimal Learning and Planning in Separated Latent MDPs</Title>
<PublishedIn>
<Publication>
</Publication>
</PublishedIn>
<PublicationDate>2025-02</PublicationDate>
<Authors>
<Author>
<DisplayName>Chen, Fan</DisplayName>
<Affiliation>
<OrgUnit>
</OrgUnit>
</Affiliation>
</Author>
</Authors>
<Editors>
</Editors>
<Publishers>
<Publisher>
<DisplayName>Massachusetts Institute of Technology</DisplayName>
<OrgUnit />
</Publisher>
</Publishers>
<License>https://rightsstatements.org/page/InC-EDU/1.0/</License>
<Abstract>We study computational and statistical aspects of learning Latent Markov Decision Processes (LMDPs). In this model, the learner interacts with an MDP drawn at the beginning of each epoch from an unknown mixture of MDPs. To sidestep known impossibility results, we consider several notions of δ-separation of the constituent MDPs. The main thrust of this paper is in establishing a nearly-sharp statistical threshold for the horizon length necessary for efficient learning. On the computational side, we show that under a weaker assumption of separability under the optimal policy, there is a quasi-polynomial algorithm with time complexity scaling in terms of the statistical threshold. We further show a near-matching time complexity lower bound under the exponential time hypothesis.</Abstract>
<Access xmlns="http://purl.org/coar/access_right"
>
</Access>
</Publication>