Summary of Adaptive Experimentation When You Can’t Experiment, by Yao Zhao et al.
Adaptive Experimentation When You Can’t Experiment
by Yao Zhao, Kwang-Sung Jun, Tanner Fiez, Lalit Jain
First submitted to arxiv on: 15 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Methodology (stat.ME)
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 introduces the concept of the confounded pure exploration transductive linear bandit (CPET-LB) problem, which arises when online services cannot directly assign users to specific control or treatment experiences due to business or practical reasons. To address this issue, the authors propose an adaptive experimental design approach that learns the best-performing treatment for encouragement designs in a linear structural equation model setting. The methodology employs elimination-style algorithms and a novel finite-time confidence interval on an instrumental variable style estimator, with sample complexity upper bounds nearly matching a minimax lower bound. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper explores how online services can learn the best-performing treatment without direct control or treatment assignments. The authors introduce a new problem called CPET-LB, where users are encouraged towards specific treatments. They develop an adaptive experimental design approach that works well in this setting and demonstrate its effectiveness through experiments. |
