Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2406.19531 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915180837863424 |
|---|---|
| author | Hao, Meiling Su, Pingfan Hu, Liyuan Szabo, Zoltan Zhao, Qingyuan Shi, Chengchun |
| author_facet | Hao, Meiling Su, Pingfan Hu, Liyuan Szabo, Zoltan Zhao, Qingyuan Shi, Chengchun |
| contents | Off-policy evaluation (OPE) is crucial for assessing a target policy's impact offline before its deployment. However, achieving accurate OPE in large state spaces remains challenging. This paper studies state abstractions -- originally designed for policy learning -- in the context of OPE. Our contributions are three-fold: (i) We define a set of irrelevance conditions central to learning state abstractions for OPE, and derive a backward-model-irrelevance condition for achieving irrelevance in %sequential and (marginalized) importance sampling ratios by constructing a time-reversed Markov decision process (MDP). (ii) We propose a novel iterative procedure that sequentially projects the original state space into a smaller space, resulting in a deeply-abstracted state, which substantially simplifies the sample complexity of OPE arising from high cardinality. (iii) We prove the Fisher consistencies of various OPE estimators when applied to our proposed abstract state spaces. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_19531 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Off-policy Evaluation with Deeply-abstracted States Hao, Meiling Su, Pingfan Hu, Liyuan Szabo, Zoltan Zhao, Qingyuan Shi, Chengchun Machine Learning G.3; I.2.6; G.1.2 Off-policy evaluation (OPE) is crucial for assessing a target policy's impact offline before its deployment. However, achieving accurate OPE in large state spaces remains challenging. This paper studies state abstractions -- originally designed for policy learning -- in the context of OPE. Our contributions are three-fold: (i) We define a set of irrelevance conditions central to learning state abstractions for OPE, and derive a backward-model-irrelevance condition for achieving irrelevance in %sequential and (marginalized) importance sampling ratios by constructing a time-reversed Markov decision process (MDP). (ii) We propose a novel iterative procedure that sequentially projects the original state space into a smaller space, resulting in a deeply-abstracted state, which substantially simplifies the sample complexity of OPE arising from high cardinality. (iii) We prove the Fisher consistencies of various OPE estimators when applied to our proposed abstract state spaces. |
| title | Off-policy Evaluation with Deeply-abstracted States |
| topic | Machine Learning G.3; I.2.6; G.1.2 |
| url | https://arxiv.org/abs/2406.19531 |