Summary of Learning to Cooperate with Humans Using Generative Agents, by Yancheng Liang et al.
Learning to Cooperate with Humans using Generative Agents
by Yancheng Liang, Daphne Chen, Abhishek Gupta, Simon S. Du, Natasha Jaques
First submitted to arxiv on: 21 Nov 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
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 paper proposes a novel approach to multi-agent reinforcement learning (MARL) by training agents that can coordinate zero-shot with humans. The current algorithms focus on training simulated human partner policies, but these approaches often struggle to produce a Cooperator that can coordinate well with real humans. To address this issue, the authors introduce a generative model of human partners that learns a latent variable representation of the human’s unique strategy, intention, experience, or style. This model can be trained from any agent interaction data and used to produce different partners to train Cooperator agents. The authors evaluate their method, called GAMMA (Generative Agent Modeling for Multi-agent Adaptation), on the challenging cooperative cooking game Overcooked, using both simulated and human datasets. The results show that GAMMA consistently improves performance with real human teammates. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper is about teaching machines to work together with humans without needing any specific training. Right now, we’re stuck using fake humans to teach our machines how to cooperate, but this doesn’t always work well in the real world. The researchers came up with a new idea: they created a model that can learn what makes each human unique and use it to create different partners for their machine learning agents. They tested this method on a game called Overcooked and found that it worked much better than before. |
Keywords
* Artificial intelligence * Generative model * Machine learning * Reinforcement learning * Zero shot
