/SIMPLE-NEURAL-NETWORK

A scratch written, regressive/progressive propagation therein gradient descent aligned artificial neural network framework program of mine, that reasonably models human neuronal phenomena. (written in java; by Jordan Micah Bennett)

Primary LanguageJava

OUTLINE

A scratch written, regressive/progressive propagation therein gradient descent aligned artificial neural network framework program of mine, that reasonably models human neuronal phenomena.

OTHER NEURAL NETWORKS WRITTEN BY MYSELF

https://github.com/JordanMicahBennett/NEURAL_NETWORK_PRACTICE/

DEMONSTRATION

The source sample includes demonstrations, configured via ( value * value * value ) neuron topology (this is extensible see ...USES); such that such:

i.Perceives handwritten numerals ( 1024x1024x10 neurons ). Conclusively, said neural network aptly perceives priorly unseen input numerals. (including terribly or incompletely written input numerals)

Alt text

ii.Perceives xor input vectors ( 2x2x1 neurons ).

Alt text

USES

This neural network utilizes a simplistic topological mechanism, therein encompassing the construction of sequentially hierarchically horizontal patterns of knowledge genera/perception.

Alt text

Alt text

INSTRUCTIONS

See INSTRUCTIONS.md.

WHY IS IT IMPORTANT TO HAVE AN INTUITIVE GRASP OF THE BASIC ARTIFICIAL NEURAL NETWORK?

https://www.quora.com/What-is-the-most-intuitive-explanation-of-artificial-neural-networks/answer/Jordan-Bennett-9

AUTHOR PORTFOLIO

http://folioverse.appspot.com/