Seminar: Glia-Augmented Artificial Neural Networks; Foundations and Applications

Zahra Sajedinia
M.Sc. Candidate
Supervisor: Dr. Todd Wareham

Glia-Augmented Artificial Neural Networks; Foundations and Applications

Department of Computer Science
Tuesday, March 24, 2015, 1:00p.m., Room EN 2022


Information processing in the human brain has always been considered as a source of inspiration in Artificial Intelligence; in particular, it has led researchers to develop different tools such as artificial neural networks. Recent findings in Neurophysiology provide evidence that not only neurons but also isolated and networks of astrocytes are responsible for processing information in the human brain. Artificial neural networks (ANNs) model neuron-neuron communications. Artificial neuron-glia networks (ANGN), in addition to neuron-neuron communications, model neuron-astrocyte connections. In this talk, in continuation of the research on ANGNs, first we propose, and evaluate a model of adaptive neuro fuzzy inference systems augmented with artificial astrocytes. Then, we propose a model of ANGNs that captures the communications of astrocytes in the brain; in this model, a network of artificial astrocytes are implemented on top of a typical neural network. The results of the implementation of both networks show that on certain combinations of parameter values specifying astrocytes and their connections, the new networks outperform typical neural networks. This research opens a range of possibilities for future work on designing more powerful architectures of artificial neural networks that are based on more realistic models of the human brain.