Language of the library: Java

Execution environment: Windows 10

Additional information: The architecture of Katie-Ann is based on the Multilayer Perceptron (MLP) which is a large class of feedforward neural networks with neurons arranged in layers. In this structure, all neurons in adjacent layers are connected to through uni-directional links called synaptic weights. Learning of the MLP consist of the adaptation of all synaptic weights in such a way that the discrepancy between the actual output signals and the desired signals, averaged over all learning examples, is as small as possible. The standard back propagation algorithm which uses the “Steepest-Descent Gradient Approach” to minimize the mean-squared error function is used during the training phase.