The evolution of communication in adaptive agents

This paper explores how communication can be understood as an adaptation by agents to their environment. We model agents as recurrent neural networks. After arguing against systems which use discrete symbols to evolve communication, we supply our agents with a number of continuous communications channels. The agents use these channels to initiate real-valued signals which propagate through the environment, decaying over distance, perhaps being perturbed by environmental noise. Initially, the agents ’ signals appear random; over time, a structure emerges as the agents learn to communicate task-specific information about their environment. We demonstrate how different communication schemes can evolve for a task, and then discover a commonality between the schemes in terms of information passed between agents. From this we discuss what it means to communicate, and describe how a semantics emerges in the agents ’ signals relative to their task domain. Key Words: communication; evolutionary algorithms; autonomous agents; neural network