Which tasks can a simple perceptron do? Im learning about neural networks and i need something to code. I've done logic gates and pixel brightness, what else?
>>8972617
shut up
>>8972618
appreciate the bump tho
>>8972617
perceptron what
>>8972625
basic form of an artificial neural network. Inputs, neuron, activation function
>>8972617
recognize shapes
read letters
control the universe
sort items
>>8972644
>simple perceptron
one neuron, only simple things
Sum. Weighted sum. Calculate averages.
>>8972617
>Which tasks can a simple perceptron do
Why don't you read the chapter to learn that very question?
Single perceptrons can only learn linearly separable tasks.
>>8972617
Mapping string landscape.
Start with Calabi-Yau from reflexive polytopes. Predict how many generation each object has, whether Hodge numbers are high or low, and so on.
You can easily generate the dataset.
>>8972617
Code a protein structure predictor.
>>8972816
only if the perceptron is using a linear function, it doesn't have to
>>8974403
Not true. Doesn't matter what activation function you use, a single perceptron will always remain linear.
>>8974403
false
activation function describes how perceptron converges on solution
>>8974519
what if you square your inputs?
>>8974524
Still linear.
>>8974524
if i gave you the equation y = 3*f(x),
is y linear?