Jensen-Shannon Information Based Characterization of the Generalization Error of Learning Algorithms

Generalization error bounds are critical to understanding the performance of\nmachine learning models. In this work, we propose a new information-theoretic\nbased generalization error upper bound applicable to supervised learning\nscenarios. We show that our general bound can specialize in various previous\nbounds. We also show that our general bound can be specialized under some\nconditions to a new bound involving the Jensen-Shannon information between a\nrandom variable modelling the set of training samples and another random\nvariable modelling the hypothesis. We also prove that our bound can be tighter\nthan mutual information-based bounds under some conditions.\n