Summary of Entropy-regularized Point-based Value Iteration, by Harrison Delecki et al.
Entropy-regularized Point-based Value Iteration
by Harrison Delecki, Marcell Vazquez-Chanlatte, Esen Yel, Kyle Wray, Tomer Arnon, Stefan Witwicki, Mykel J. Kochenderfer
First submitted to arxiv on: 14 Feb 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: None
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 proposes an entropy-regularized model-based planner for partially observable problems, addressing the challenges of model uncertainty during planning and goal uncertainty during objective inference. By promoting policy robustness through entropy regularization, the planner encourages policies to be no more committed to a single action than necessary. The authors evaluate the performance of entropy-regularized policies in three problem domains, showing that they outperform non-entropy-regularized baselines in terms of expected returns under modeling errors and objective inference accuracy. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper creates a new way for computers to plan and make decisions when there’s uncertainty about what might happen. This is helpful when we don’t know exactly how things will go, but want to be prepared for different outcomes. The method uses something called entropy regularization to help the computer not get too stuck on one idea and instead consider many possibilities. The authors tested this approach in three scenarios and found that it worked better than other methods at dealing with uncertainty. |
Keywords
» Artificial intelligence » Inference » Regularization
