AI engineering glossary · Generative media multimodal
What is reparameterisation trick?
Reparameterisation trick: A method that rewrites sampling from a learned Gaussian as a deterministic function of its mean and standard deviation plus independent noise, so gradients can flow through the sampling step. It makes VAEs trainable with ordinary backpropagation.
Reparameterisation trick: an example
z = mu + sigma * eps with eps drawn from a standard normal lets the loss gradient reach mu and sigma.
Related terms
More Generative media multimodal terms
- Residual vector quantisation
- Reverse diffusion process
- RNN-Transducer
- Sample rate
- Score distillation sampling
- Score-based generative model
- Short-time Fourier transform
- SigLIP
Where you learn it
Reparameterisation trick is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.