This respository contains a tutorial with code examples of the AMIDST toolbox. The slides of this tutorial can be found in the following link:
https://github.com/amidst/tutorial/raw/master/doc/slides/AMIDST_slides.pdf
This tutorial focuses on how to learn the predefined latent varaibles (i.e. a probabilistic graphical model with unobserved variables) models included in AMIDST. Two learning schemes are showed:
- Parellel multi-core learning based on Java 8 streams.
- Distributed learning functionalities based on Apache Flink.
In both cases, AMIDST offers Bayesian parameter learning by means of variational message passing.
We also show how new user-defined latent varaible models can be easily defined with this toolbox and, also, learnt from local or distributed data.
This is the list of latent variables models included in AMIDST.
