Beginner @ coding / First "AI" project

Hello everyone!
I am a beginner programmer and have been trying to learn python on my own for a few months now and decided it was time to reach out to the python community for help and resources.
i have an AI project I’ve been working on and would love to get some feedback on it.

‘’’ Basic feedforward neural network based on the following youtube link https://www.youtube.com/watch?v=ILsA4nyG7I0
does not include a function for updating weights.
any constructive critique is much appreciated (i’m still new at coding)
comments/questions [email protected]

P.S program still has bugs.

Input map is in order of > 1 2
3 4

Black = -1
White = 1
‘’’

input 4 pixel map

n1 = 1
n2 = -1
n3 = -1
n4 = 1

#layer 1 weights
wegt1 = 1
wegt2 = 1
wegt3 = 1
wegt4 = 1
wegt5 = 1
wegt6 = -1
wegt7 = 1
wegt8 = -1

#layer 1 neurons
neuron1 = (n1 * wegt1) + (n4 * wegt2)
#ReLU function {I think}
if neuron1 > 0.1:
neu1 = neuron1
else:
neu1 = 0
neuron2 = (n2 * wegt3) + (n3 * wegt4)
if neuron2 > 0.1:
neu2 = neuron2
else:
neu2 = 0
neuron3 = (n1 * wegt5) + (n4 * wegt6)
if neuron3 > 0.1:
neu3 = neuron3
else:
neu3 = 0
neuron4 = (n2 * wegt7) + (n3 * wegt8)
if neuron4 > 0.1:
neu4 = neuron4
else:
neu4 = 0

repackage of variables

neu1 = neuron1
neu2 = neuron2
neu3 = neuron3
neu4 = neuron4

#layer 2 weights
wegt1a = 1
wegt2a = 1
wegt3a = -1
wegt4a = 1
wegt5a = 1
wegt6a = -1
wegt7a = 1
wegt8a = 1

#layer 2 neurons
neuron1a = (neu1 * wegt1a) + (neu2 * wegt2a)
if neuron1a > 0.1:
neu1a = neuron1a
else:
neu1a = 0

neuron2a = (neu1 * wegt3a) + (neu2 * wegt4a)
if neuron2a > 0.1:
neu2a = neuron2a
else:
neu2a = 0

neuron3a = (neu3 * wegt5a) + (neu4 * wegt6a)
if neuron3a > 0.1:
neu3a = neuron3a
else:
neu3a = 0

neuron4a = (neu3 * wegt7a) + (neu4 * wegt8a)
if neuron4a > 0.1:
neu4a = neuron4a
else:
neu4a = 0

layer 3 weights

wegt_a = 1
wegt_b = -1
wegt_c = 1
wegt_d = -1
wegt_e = 1
wegt_f = -1
wegt_g = 1
wegt_h = -1

Output neurons

op_neu_1 = neu1a * wegt_a
op_neu_2 = neu1a * wegt_b
op_neu_3 = neu2a * wegt_c
op_neu_4 = neu2a * wegt_d
op_neu_5 = neu3a * wegt_e
op_neu_6 = neu3a * wegt_f
op_neu_7 = neu4a * wegt_g
op_neu_8 = neu4a * wegt_h

print(“Possible patterns recognized”)
print("")
print(“White Solid Detected”)
print(op_neu_1)
print(“Black Solid Detected”)
print(op_neu_2)
print(“Left Black Vertical Detected”)
print(op_neu_3)
print(“Right Black Vertical Detected”)
print(op_neu_4)
print(“Forward Slash Detected”)
print(op_neu_5)
print(“Back Slash Detected”)
print(op_neu_6)
print(“Upper White Horizontal Detected”)
print(op_neu_7)
print(“Lower White Horizontal Detected”)
print(op_neu_8)
print("")
print("")
print("")
print(“First Layer Output”)
print(neuron1)
print(neuron2)
print(neuron3)
print(neuron4)
print(“Second Layer Output”)
print(neu1a)
print(neu2a)
print(neu3a)
print(neu4a)

                                                          Thank You for giving it a look.
1 Like

You need to revisit programming fundamentals
Loops and Arrays will clean this project up tremendously.

1 Like

It is really good to know that you dived in the AI project.
Here is a simple Python tutorial to implement feed forward neural network in Python that you might like. This is done using the NumPy library.
And here is another project that you might love:
A speech/voice command calculator using speech recognition and PyAudio.
Like if you say " seven plus two" the output will be 9.
It can do a lot more calculations.