Summary of Spqr: Controlling Q-ensemble Independence with Spiked Random Model For Reinforcement Learning, by Dohyeok Lee et al.
SPQR: Controlling Q-ensemble Independence with Spiked Random Model for Reinforcement Learning
by Dohyeok Lee, Seungyub Han, Taehyun Cho, Jungwoo Lee
First submitted to arxiv on: 6 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
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 proposed spiked Wishart Q-ensemble independence regularization (SPQR) aims to alleviate overestimation bias in deep reinforcement learning by ensuring guaranteed independence between multiple Q-functions. By modifying the hypothesis testing criterion into a tractable KL divergence, SPQR achieves better performance than baseline algorithms in both online and offline RL benchmarks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Deep reinforcement learning can struggle with complex tasks or out-of-distribution data due to overestimation bias. To overcome this challenge, ensemble methods for Q-learning have been explored, focusing on network initialization and heuristically designed diversity injection methods. However, ensuring guaranteed independence between Q-functions has not been approached from a theoretical perspective. SPQR bridges this gap by introducing a novel regularization loss based on random matrix theory. |
Keywords
* Artificial intelligence * Regularization * Reinforcement learning
