Summary of Dcmac: Demand-aware Customized Multi-agent Communication Via Upper Bound Training, by Dongkun Huo et al.
DCMAC: Demand-aware Customized Multi-Agent Communication via Upper Bound Training
by Dongkun Huo, Huateng Zhang, Yixue Hao, Yuanlin Ye, Long Hu, Rui Wang, Min Chen
First submitted to arxiv on: 11 Sep 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: 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 A collaborative multi-agent reinforcement learning framework, Efficient communication is crucial for enhancing performance. The conventional approach of sharing observations via full communication leads to significant overhead. Existing methods attempt to perceive global states by using local information-based teammate models, but they ignore the uncertainty generated by predictions, making training challenging. To address this issue, we propose a Demand-aware Customized Multi-Agent Communication (DCMAC) protocol that utilizes upper bound training to obtain an ideal policy. This method allows agents to interpret the gain of sending local messages to teammates and generate customized messages using cross-attention mechanisms. Additionally, our approach adapts to communication resources and accelerates training by appropriating an ideal policy trained with joint observations. Experimental results demonstrate that DCMAC outperforms baseline algorithms in both unconstrained and communication-constrained scenarios. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A team of artificial agents work together to solve a problem. They need to share information, but sharing too much slows them down. Some ways agents share info try to guess what’s going on globally, but this can lead to confusion. To fix this, we created a new way for agents to communicate, called Demand-aware Customized Multi-Agent Communication (DCMAC). This system helps agents decide when and how to send messages to teammates based on their needs. It also learns to prioritize which messages are most important. When tested, DCMAC did better than other methods in both normal and slow-communication situations. |
Keywords
» Artificial intelligence » Cross attention » Reinforcement learning
