Summary of End-to-end Neuro-symbolic Reinforcement Learning with Textual Explanations, by Lirui Luo et al.
End-to-End Neuro-Symbolic Reinforcement Learning with Textual Explanations
by Lirui Luo, Guoxi Zhang, Hongming Xu, Yaodong Yang, Cong Fang, Qing Li
First submitted to arxiv on: 19 Mar 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 proposed neuro-symbolic reinforcement learning framework refines structured states with rewards, enhancing explainability in decision-making. It integrates a perception module from vision foundation models and policy learning, addressing accessibility issues by reducing domain knowledge requirements. The approach is evaluated on nine Atari tasks, demonstrating efficacy. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A new way to make decisions is being developed, called neuro-symbolic reinforcement learning (NS-RL). This method helps us understand why computers are making certain choices. Currently, it’s hard for people without special knowledge to know what computers mean when they make decisions. The researchers created a system that combines two important parts: one that helps the computer see its environment and another that learns how to make good decisions. They tested this system on nine different games and found that it worked well. |
Keywords
* Artificial intelligence * Reinforcement learning
