Construct basic neural networks in Common Lisp.
$ cat ./examples/xor.txt
0 0 | 0
0 1 | 1
1 0 | 1
1 1 | 0
(load-training-data-from-txt-file "./examples/xor.txt")
(setq *network-architecture* '(2 1)) ; list of neurons by layer { 2 x 1 }
(construct-network)
(learn-backprop 2000) ; e.g. 2000 iterations
(check-results)
$ ./main.fasl
...
------------------------------------
input : #(0 0)
expected output: #(0)
network output : #(0.10375877)
------------------------------------
input : #(0 1)
expected output: #(1)
network output : #(0.8785281)
------------------------------------
input : #(1 0)
expected output: #(1)
network output : #(0.8776233)
------------------------------------
input : #(1 1)
expected output: #(0)
network output : #(0.12952015)
------------------------------------
- Machine Learning in C - https://www.youtube.com/watch?v=PGSba51aRYU
- What is backpropagation really doing? - https://www.youtube.com/watch?v=Ilg3gGewQ5U
- Backpropagation - https://en.wikipedia.org/wiki/Backpropagation