Summary of Model-based Offline Reinforcement Learning with Lower Expectile Q-learning, by Kwanyoung Park and Youngwoon Lee
Model-based Offline Reinforcement Learning with Lower Expectile Q-Learning
by Kwanyoung Park, Youngwoon Lee
First submitted to arxiv on: 30 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
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 a novel model-based offline reinforcement learning (RL) method called Lower Expectile Q-learning (LEQ), which addresses the challenge of inaccurate value estimation from model rollouts. LEQ uses lower expectile regression to estimate values, outperforming previous model-based offline RL methods on long-horizon tasks like D4RL AntMaze. It matches or surpasses performance of model-free approaches and sequence modeling approaches. The method also works robustly across diverse domains, including state-based and pixel-based tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about a new way to learn from limited data using models. Instead of trying to get the exact right answer, it finds a good-enough solution by making imaginary moves with the model. This approach does better than previous methods on big problems like AntMaze. It’s also as good as other ways to solve these problems. |
Keywords
* Artificial intelligence * Regression * Reinforcement learning
