"What's a neural network?"
Stop picturing a neural network as brain neurons.
Here's a bettermathematical definition:
"a neural network is a differentiable computational graph"
- a graph contains nodes and edges.
- each node represents a function (i.e. math operation)
- edges connect the nodes
Differentiable means the derivative exists at each point, so no holes or kinks.
During training, the neural network traces error backward, measures how values contributed to the result, and adjusts its learning parameters.
I plan to post YT trainings explaining deep learning theory & math... and I'm sure that it will get no views :D