If I understand correctly, statistical learning theory is just one approach to machine learning,
What machine learning isn't statistical?
Based on my very limited understanding, I thought that things like PAC-learning or Empirical Risk minimization pretty much cover everything. Isn't statistics involved in all of this?
Is there any good source that clearly explains this?
EDIT
my comment below shows the answers I was looking for.
In the probabilistic setting, we use Bayes rule for inference and need to keep track of all possible hypotheses (computationally expensive, explains use of MCMC or Gaussian distribution).
These two sometimes overlap (e. g. least squares viewed as MLE) but are fundamentally different.
– user3629892 Apr 25 '21 at 09:12