f-Divergence
A class of measures of the difference between two probability distributions, defined through a convex function f. Common forms include KL divergence, JS divergence and total variation distance, and they are widely used in distribution alignment and in the theoretical analysis of adversarial training.
Timeline (last 90 days)
DiSPaT is based on a theoretical framework of f-divergence minimization and derives a tight approximation, whose training objective is equivalent to narrowing the gap between the true normal data distribution and the distribution of samples generated by the model.
DiSPaT shares ideas with generative adversarial networks, but its theoretical foundation is built on f-divergence minimization rather than JS divergence.
f-divergence is a class of measures of the difference between two probability distributions, defined via a convex function f, with common forms including KL divergence, JS divergence, and total variation distance.
All Facts (3)
DiSPaT is based on a theoretical framework of f-divergence minimization and derives a tight approximation, whose training objective is equivalent to narrowing the gap between the true normal data distribution and the distribution of samples generated by the model.
50%UnverifiedDiSPaT shares ideas with generative adversarial networks, but its theoretical foundation is built on f-divergence minimization rather than JS divergence.
50%Unverifiedf-divergence is a class of measures of the difference between two probability distributions, defined via a convex function f, with common forms including KL divergence, JS divergence, and total variation distance.
50%