Introduction to Probabilistic Machine Learning with PyMC3

Abstract:

Machine Learning has gone mainstream and now powers several real-world applications like autonomous vehicles at Uber & Tesla, recommendation engines on Amazon & Netflix, and much more. This meetup will introduce probabilistic machine learning and probabilistic programming with PyMC3 (http://docs.pymc.io/). We will discuss the basics of machine learning from a probabilistic/Bayesian perspective (http://mlg.eng.cam.ac.uk/zoubin/bayesian.html) and contrast it with traditional/algorithmic machine learning.

We will also discuss how to build probabilistic models in computer code using a new exciting programming paradigm called Probabilistic Programming (http://probabilistic-programming.org/wiki/Home) (PP). Particularly we shall use PyMC3 (http://docs.pymc.io/) a Python PP language, to build models ranging from simple generalized linear models to complex hierarchical models and nonparametric models for machine learning.

Pre-requisite: Please bring a mobile device with a modern web browser such as Google Chrome or Mozilla Firefox to explore the code on Binder (https://mybinder.org/).

Speaker Bio:

Daniel Emaasit (http://www.danielemaasit.com/) is a Data Scientist at Haystax Technology. His interests involve developing principled probabilistic models for problems where training data are scarce by leveraging knowledge from subject-matter experts and context information. In particular, he is interested in flexible Bayesian machine learning methods, such as Gaussian processes and Dirichlet processes, and data-efficient learning methods such as Bayesian optimization.