No, ML model is not evaluated empirically at all. Measures like accuracy and precision are not empirical. Neither is generalization as evaluated by using some m

No, ML model is not evaluated empirically at all. Measures like accuracy and precision are not empirical. Neither is generalization as evaluated by using some model data set. These measures are statistical predictions that may or may not be correct. Failure rate in the wild is empirical. Accuracy in the wild, the same. Since the designer does not have access to real data, they are actually not working empirically at all and as such you get common overfitting and methods that work well on model datasets but not on real data.