Page 1 of 1 [ 3 posts ] 

kevinjh
Veteran
Veteran

Joined: 6 Nov 2011
Gender: Male
Posts: 8,456
Location: .

20 May 2012, 1:49 am

Given the following, what might constitute a good introductory program with neural networks?
* Some experience with Python
* Fair knowledge of the concept of a neuron and the conditions by which it fires
* A decent amount of experience with Boolean algebra



echinopsis
Snowy Owl
Snowy Owl

User avatar

Joined: 3 Oct 2011
Age: 36
Gender: Female
Posts: 132
Location: Germany

20 May 2012, 10:26 am

given a fair amount of neurobiological and mathematical background id recommend the book "spikes" by bialek, rieke and warland. imo its vital to keep basic concepts like simple arithmetic and logic operations with neurons, hodgin-huxley, p/d/d2 type resonses, shunting inhibition, leaky integration, the time constant regarding integrator neurons and coincidence detector neurons, convolutions, correlations, filters, lateral inhibition, center/surround antagonism, ltp and hebbian learning in general and all such things in mind in order to get to a clear understanding, but that also depends on what you're interested in.



kevinjh
Veteran
Veteran

Joined: 6 Nov 2011
Gender: Male
Posts: 8,456
Location: .

21 May 2012, 5:28 pm

Thank you for your suggestions! :D

I shall begin finding ways to realize them.