Neural Networks in Python without using any readymade libraries...i.e., from first principles..help!

Question

I am trying to learn programming in python and am also working against a deadline for setting up a neural network which looks like it's going to feature multidirectional associative memory and recurrent connections among other things. While the mathematics for all these things can be accessed from various texts and sources (and is accessible, so to speak), as a newbie to python (and programming as a profession) I am kinda floating in space looking for the firmament as I try to 'implement' things!! Information on any good online tutorials on constructing neural networks ab initio will be greatly appreciated :)

In the meantime I am moonlighting as a MatLab user to nurse the wounds caused by Python :)

Answer 1

I don't know anything about neural networks, but PyNeurGen seems like a good resource.

Answer 2

If you're familiar with Matlab, check out the excellent Python libraries numpy, scipy, and matplotlib. Together, they provide the most commonly used subset of Matlab functions.