Hello. the teacher provided a nice description about how machine learning is different from traditional rule-based programming where we hard-coded the know relation. but here in the ML paradigm we ask a general algorithm like NN to do that for us.

here my question is about the cases where we know the general form of the function/relation between input and output parameter. what is unknown is only the coefficients of this known function. in this case, we use same logic the solve the problem. i mean some training and testing finally. is this also under ML?