A fully adaptive neural network modeling using colored Petri nets

What is it about?

Motivated by a long lasting interest in using the formalism of PNs either to emulate or to design neural network architectures, we revisit this research topic by resorting to the powerful modeling framework of hierarchical timed colored PNs (HTCPNs) to introduce a novel approach that builds a fully adaptive one-hidden-layered multilayer perceptron (MLP) model trained by the famed backpropagation algorithm. The resulting proposed model is called HTCPN-MLP and consists of a general structure capable of handling classification and regression tasks.

Featured Image

Why is it important?

It is important due the following prominent characteristics can be assigned to the proposed approach: 1. ability to handle any data set; 2. ability to generate random weights for training initialization; 3. use of time constraints to present data in epochs; 4. hierarchical structure and modeling of activation functions; 5. implementation based on CPNs while storing information in their tokens; 6. ability to propagate the error owing to the backpropagation algorithm; 7. all learning stages involve training and testing.

Perspectives