Summary of Equivariant Offline Reinforcement Learning, by Arsh Tangri et al.
Equivariant Offline Reinforcement Learning
by Arsh Tangri, Ondrej Biza, Dian Wang, David Klee, Owen Howell, Robert Platt
First submitted to arxiv on: 20 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Robotics (cs.RO)
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 method utilizes SO(2)-equivariant neural networks for offline reinforcement learning (RL) in robotic manipulation tasks with limited demonstrations. By leveraging rotation-symmetry, the approach enhances Conservative Q-Learning (CQL) and Implicit Q-Learning (IQL) algorithms to excel in low-data regimes. Experimental results show equivariant versions outperforming non-equivariant counterparts. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Offline reinforcement learning for robotic manipulation can be challenging due to limited expert demonstrations and online RL difficulties. The proposed method uses SO(2)-equivariant neural networks to improve policy learning from a small number of demonstrations. Results show that using equivariant CQL and IQL algorithms leads to better performance in low-data situations. |
Keywords
* Artificial intelligence * Reinforcement learning
