Summary of Learning to Assist Humans Without Inferring Rewards, by Vivek Myers et al.
Learning to Assist Humans without Inferring Rewards
by Vivek Myers, Evan Ellis, Sergey Levine, Benjamin Eysenbach, Anca Dragan
First submitted to arxiv on: 4 Nov 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Computers and Society (cs.CY); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG)
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 This paper addresses the problem of assistive agents making humans’ lives easier through empowerment, which measures an agent’s influence on environmental outcomes. We build upon prior work that focuses on maximizing human control over task completion, but overcome scalability limitations by introducing contrastive successor representations. Our proposed method formally proves to estimate a similar notion of empowerment as prior work and empirically outperforms existing methods on synthetic benchmarks and the cooperative game setting of Overcooked. Theoretically, our work connects information theory, neuroscience, and reinforcement learning, highlighting the potential for representations to play a critical role in solving assistive problems. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research is about making robots or computer programs help people do things more easily. Instead of just following instructions, these machines learn how to make humans’ actions have a bigger impact on what’s happening around them. The paper finds a new way to measure this “helpfulness” and shows that it can be used in games like Overcooked, where players work together. This breakthrough could help robots or computers assist people in many different situations. |
Keywords
* Artificial intelligence * Reinforcement learning
