Deviation bound for non-causal machine learning

Rémy Garnier, Raphaël Langhendries

Concentration inequalities are widely used for analyzing machine learning algorithms. However, current concentration inequalities cannot be applied to some of the most popular deep…