Implementation of Gaussian Process Latent Variable models in pytorch/gpytorch

Key References:

1. Gaussian process latent variable models for visualisation of high dimensional data
2. Bayesian GPLVM
3. Local distance preservation in the GPLVM
4. Stochastic Variational Inference for back-constrained GPLVM
5. Gaussian Processes for Big data

Models

1. GPLVM
2. B-GPLVM
3. Back-constrained GPLVM
4. SVI GPLVM (with and without back-constraints)

Inference

1. ML-II Optimisation
2. Variational Inference with the collapsed bound
3. Stochastic Variational Inference with the uncollapsed bound

Code Layout

data/* has all the data loading utilities
models/* model classes

Usage

See demo.py