Summary of Safety Through Feedback in Constrained Rl, by Shashank Reddy Chirra et al.
Safety through feedback in Constrained RL
by Shashank Reddy Chirra, Pradeep Varakantham, Praveen Paruchuri
First submitted to arxiv on: 28 Jun 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 A novel approach to designing cost functions for safety-critical reinforcement learning (RL) tasks is proposed, which scales to complex domains and extends beyond state-level feedback. The method learns a surrogate objective that transforms the problem into a state-level supervised classification task with noisy labels, allowing efficient solution. Additionally, novelty-based sampling is introduced to selectively involve human evaluators only when the agent encounters novel trajectories. Experimental results on benchmark Safety Gymnasium environments and realistic self-driving scenarios demonstrate the efficiency of the proposed method. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A new way to help computers make safe decisions in complex situations like self-driving cars is presented. The approach makes it easier to design a “cost function” that helps an agent avoid mistakes by learning from feedback collected offline. This approach can be used in more complicated environments and reduces the need for human evaluators to provide feedback on every decision made by the computer. |
Keywords
» Artificial intelligence » Classification » Reinforcement learning » Supervised
