Sign inSign in
ConceptAI & ML

Reparameterization trick

The reparameterization trick rewrites a random variable zz from a parameterized Gaussian distribution as a deterministic function of the parameters and an independent standard Gaussian noise variable, enabling low-variance gradient estimation through backpropagation.
as Sergey Levine defines itUC BerkeleyThe wording closest to all 3 definitions
Watch it explained

Know first

Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.

Opens up

Concepts that name this one as a prerequisite.

Lecturers

2

Each in their own terms, from the easiest level.

  • Sergey Levine

    UC Berkeley · Undergraduate

    A technique used in variational autoencoders to backpropagate gradients through stochastic nodes by expressing the latent variable as a deterministic function of parameters and independent noise.
    • 2:36Variational Inference, Part 4 (CS 285, Lecture 18)
  • Phillip Isola

    MIT · Undergraduate

    A technique to sample from a Gaussian distribution by transforming a sample from a unit Gaussian using the mean and standard deviation.
    • 25:03Generative Models: Representation Learning Meets Generative Modeling

Connections

9

The lectures that define it or assume it, and the concepts next to it.

2 lecturers. Definitions are written by AI from each lecture’s transcript. A time with a play mark starts the lecture at the moment the lecturer says it.