Summary of Off-policy Evaluation with Deeply-abstracted States, by Meiling Hao et al.
Off-policy Evaluation with Deeply-abstracted States
by Meiling Hao, Pingfan Su, Liyuan Hu, Zoltan Szabo, Qingyuan Zhao, Chengchun Shi
First submitted to arxiv on: 27 Jun 2024
Categories
- Main: Machine Learning (stat.ML)
- Secondary: Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper investigates off-policy evaluation (OPE) in large state spaces by utilizing state abstractions originally designed for policy learning. The contributions are three-fold: defining irrelevance conditions, proposing an iterative procedure to sequentially project the original state space into a smaller one, and proving the Fisher consistencies of various OPE estimators when applied to these abstracted state spaces. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Off-policy evaluation is important for assessing a target policy’s impact offline before deployment. The paper shows how using state abstractions can make it easier to evaluate policies in large state spaces. It does this by defining conditions that help learn the right abstraction and proposing a way to simplify the sample complexity of OPE. |
