Nada is a python framework derived from
Dana framework. As Dana, Nada is dedicated to creation and simulation of neural networks defined at a mesoscopic level. Neural networks are composed of cortical maps and connections between these maps. Each map is defined by some matrices of activity fields and by evolution equations of these activity matrices. These evolution equations may use activities of incoming connections and all activities of the map. Thus, a map may be considered as a matrix of generic units, each unit uses common evolution equations of its activity fields. Each connection is defined by a pre synaptic activity field of a map and by a post synaptic map. Evolution equations for connection weights are also available, allowing learning in the neural network.
Current version of Nada is distributed
here under GNU licence.