Let us finally consider one rather trivial method that enables a number of agents to work on a learning task. All agents are involved in collecting data, but only one system is involved in learning. That is, all the data is transferred to the learning agent that incrementally updates its theory. As the purpose of this paper is not to discuss incremental learning methods, but rather systems with more complex interactions between agents, we shall let the interested reader consult appropriate literature (see e.g. Schlimmer and Fisher, 1988; Utgoff, 1988; Janikow, 1989).