Posts

KL divergence

How Maximum Likelihood Estimate and Variational Inference relate to the KL divergence

The reparameterisation trick

In this post we see a well known approach for computing gradients of the Free Energy that uses a reparameterisation of the approximating density, leading to a computationally tractable method that can exploit auto differentiation.

Deriving the Free Energy functional aka the ELBO

Deriving the free energy functional starting from Bayes' theorem.

Motivating Approximate Bayesian Inference

Short rationale for why we care about approximate Bayesian inference.

No matching items