Variational Inference is a powerful algorithm that turns the task of computing the posterior(p(z|x)) into an optimization problem. This project is about implementing two inference algorithms Mean Field ADVI and Full Rank ADVI based on ADVI paper in PyMC4. Mean Field ADVI posits a spherical Gaussian family and Full Rank ADVI posits a Multivariate Gaussian family to minimize KL divergence. The implementation will use tf and tfp libraries.

Organization

Student

Sayam Kumar

Mentors

  • Thomas Wiecki
  • Maxim Kochurov
close

2020