The intention is two fold: to gain a thorough understanding of neural network basics by building one from the ground up; also, gain a thorough grounding in F# 4.0 concepts and idioms At a later stage a C# version will be added to allow me to explore a comparision of OOP concepts vs FP concepts in this domain. The C# version will explore appropriate design patterns and SOLID principles.
At time of writing a simple perceptron with a parameterised vector function has been implemented. In neural network terms, this is a basic (and old fashioned) building block. In F# terms, passing in a function and simple pattern matching is as advanced as it gets.
Until the next version that is!