> Today we are exploring a completely different Paradigm of Programming: Machine Learning. Or as I personally like to call it Automatic Mathematical Modeling. In this Paradigm you are not writing the code of the Model directly. You code what you expect from the Model. And the Learning process finds the configuration of the Model that fits your description the best.
In 3 hours he goes step by step, starting from a blank file, building a small neural network that can learn the XOR logic gate.
The interesting part is how he builds it up from first principles, no handwaving, no libraries or frameworks, each step justified and he even discovers experimentally why it is useful to have bias in neurons.
> Today we are exploring a completely different Paradigm of Programming: Machine Learning. Or as I personally like to call it Automatic Mathematical Modeling. In this Paradigm you are not writing the code of the Model directly. You code what you expect from the Model. And the Learning process finds the configuration of the Model that fits your description the best.
In 3 hours he goes step by step, starting from a blank file, building a small neural network that can learn the XOR logic gate.
The interesting part is how he builds it up from first principles, no handwaving, no libraries or frameworks, each step justified and he even discovers experimentally why it is useful to have bias in neurons.