Skip to content

Latest commit

 

History

History
17 lines (13 loc) · 576 Bytes

roadmap.md

File metadata and controls

17 lines (13 loc) · 576 Bytes

This document covers planned features for this library. More will be added later.

Modularity

  • Implementation of Keras like API for DGPs; for instance, initialising each layer on its own and then stacking the layers as a network.
  • Seperatin of model and inference procedure (similar to what GPflow does)

Multitask/Multi-kernel Support

  • Ability to create layers with multiple kernels
  • Ability to perform multi-task learning

Inference Procedures

  • Doubly Stochastic VI (Salimbeni and Diesenroth 2017)
  • Standard VI (Damianou and Lawrence 2013)
  • Some for of MCMC