1 paper
Wilson Wu, Victor Lecomte, Michael Winer +3
By far the most common way to estimate an expected loss in machine learning is to draw samples, compute the loss on each one, and take the empirical average. However, sampling is n…