what are weights in a machine learning model

Question

what are weights in a machine learning model

In: 7

Answer 1

Machine learning reduces to taking input values, doing a bunch of multiplications and additions with parameters and getting output values. It’s usually implemented as matrix multiplications because that’s easy to do on GPU. The trick is in finding the correct parameters, there are processes for starting with random values and working towards ones that actually work – that’s the machine learning part. Anyway, a parameter you multiply with input value is a “weight”, a parameter you just add without multiplying it to anything is a “bias”