Petri nets for classification

Research in neural networks has reached an intense state of activity with growing interest in practical and commercial applications. Neural networks provide a solution for tasks which require a huge processing capacity, by using a large number of simple elements operating in parallel. One such task is pattern classification. This Letter discusses Petri nets for classification, which use the same basic principles as neural networks. Because Petri nets also have appealing simplicity in expressing existing interactions in concurrent systems the authors have found them suitable for modeling or emulating neural networks.