Reparameterization Trick

Concept

A key innovation that allows gradients to be computed through stochastic layers in VAEs by expressing the sampling process deterministically using an auxiliary variable, effectively separating the stochastic and deterministic parts.

Mentioned in 1 video

Videos Mentioning Reparameterization Trick

Foundations of Unsupervised Deep Learning (Ruslan Salakhutdinov, CMU)

Foundations of Unsupervised Deep Learning (Ruslan Salakhutdinov, CMU)

Lex Fridman

A key innovation that allows gradients to be computed through stochastic layers in VAEs by expressing the sampling process deterministically using an auxiliary variable, effectively separating the stochastic and deterministic parts.