An Attempt At Demystifying Bayesian Deep Learning

Our November meeting will feature a talk from Eric Ma, about Bayesian Deep Learning. Join us to learn about this interesting topic and share your story with fellow bayesians!

Title:

An Attempt At Demystifying Bayesian Deep Learning

Abstract:

In this talk, I aim to do two things: demystify deep learning as essentially matrix multiplications with weights learned by gradient descent, and demystify Bayesian deep learning as placing priors on weights. I will then provide PyMC3 and Theano code to illustrate how to construct Bayesian deep nets and visualize uncertainty in their results.

Speaker Bio:

Eric Ma just defended his doctoral thesis in the Department of Biological Engineering at MIT in April 2017, where he developed a scalable algorithm for finding shuffled influenza viruses. He recently was an Insight Health Data Fellow in the Boston Summer 2017 session, where he developed Flu Forecaster, a project aiming to forecast influenza sequence evolution using deep learning. Eric has delivered talks and tutorials at the PyCon, SciPy and PyData conferences, covering applied topics including graph theory and Bayesian inference.

Agenda:

6:30pm: Networking

7:00pm: Talk by Eric Ma + Q&A

8:00pm: Networking

8:45pm: End of event.

Note to all attendees:

Please use your full name on meetup.com in order to access this event.

Sponsor:

This event will be sponsored by QuantumBlack