Total variation distance is a distance measure for two probability distributions. Formally, for distributions
where
Thus, we can also interpret this distance as
where
Finally, by Pinskerโs inequality, we have a relationship with โ๏ธ KL Divergence: