What is probabilistic programming all about?
There are three major trends in the machine learning side of data science: big data, deep learning (https://en.wikipedia.org/wiki/Deep_learning) and probabilistic programming. There has been a sustained focus on the first two, but recent advances in the way models are evaluated has brought the attention again to probabilitstic programming.
At the core of probabilistic programming is the idea that statistical models are written in code, which is then evaluated in turn by MCMC sampling algorithms. New variational inference algorithms have emerged as a means to scale these algorithms to production level.
There are three major reasons to consider probabilistic programming. With the Bayesian paradigm to guide model creation.
Customization - We can create models that have built-in hypothesis tests
Propagation of uncertainty - There is a degree of belief associated prediction
EstimationIntuition - The models are essentially 'white-box' which provides insight into our data
**Make sure to RSVP for this free event on Eventbrite (https://www.eventbrite.com/e/business-decisions-with-probabilistic-programming-tickets-41051977570)**
Ultimately, we create models to guide the decision making process. It is a typical task in data science to build a recommendation system or make a prediction about a medical diagnosis, but these recommendations or diagnoses mean so much more when they are accompanied by an estimated level of uncertainty.
This short talk is designed to familiarize Python programmers the basics concepts of probabilistic programming. We will introduce the Python package PyMC3 with a tutorial example. Then we will use it to build a simple recommendation system and finally we will finish up with an implementation of a probabilistic neural network.
What to bring:
Computer, fully charged
Desire to Learn
Meet your Instructor
Adam Richards | Senior Data Science Instructor - Galvanize