This survey frames multi-agent communication via the Five Ws, tracing evolution from MARL's hand-designed protocols to emergent language and LLM-based systems. It highlights trade-offs in interpretability, scalability, and generalization across paradigms. Practical design patterns and open challenges are distilled for hybrid systems.
Key Points
- 1.Frames multi-agent communication using Five Ws framework
- 2.Traces evolution from MARL hand-designed protocols to emergent language and LLM systems
- 3.Highlights trade-offs in interpretability scalability and generalization across paradigms
Impact Analysis
Researchers developers and practitioners in multi-agent AI benefit from distilled design patterns and challenges for hybrid systems. It matters by clarifying paradigm evolution and trade-offs aiding informed design choices. Potential effects include accelerated progress toward scalable interpretable multi-agent applications.
Technical Details
Survey structures analysis via Five Ws categorizing who communicates what when where and why. Evolves from rigid MARL protocols through emergent languages to flexible LLM-based methods. Distills patterns for hybrid systems addressing open challenges like generalization.