Distributed learning of information fusion: a multi-agent approach
Vladimir Gorodetsky, O. Karsaeyv, Vladimir Samoilov · 2003
An important task of information fusion scope is learning of decision making and decision combining. This task that falls under Distributed Learning scope is a subject of the paper. It is supposed that distributed learning is carried out by a component of information fusion system performing supervised of-line training and testing of its decision making component. The core problem of a distributed leaming component design does not concern particular data mining techniques. Instead of this, its core problem is development of an infrastructure and protocols supporting coherent collaborative operations of distributed software components (agents) responsible for distributed leaming. The paper is focused on architecture of multi-agent information firsion systems possessing learning capabilities, on a technology supported by a software tool and on protocols of sofiare tool agents' interaction, particularly, distributed data mining protocol. Solutions conceming the aforementioned aspects form the basis for the multi-agent information fusion system technology and respective software tool.