Generalization Properties and Implicit Regularization for Multiple Passes SGM


Junhong Lin, Raffaello Camoriano, Lorenzo Rosasco ;
Proceedings of The 33rd International Conference on Machine Learning, PMLR 48:2340-2348, 2016.


We study the generalization properties of stochastic gradient methods for learning with convex loss functions and linearly parameterized functions. We show that, in the absence of penalizations or constraints, the stability and approximation properties of the algorithm can be controlled by tuning either the step-size or the number of passes over the data. In this view, these parameters can be seen to control a form of implicit regularization. Numerical results complement the theoretical findings.

Related Material