Description
With the advancements in the domain of Deep learning, it has found application in various real-life problems such as self-driving cars and healthcare diagnosis. But with great powers comes greater responsibility, so the question arises, “Is our AI safe ?”
The critical part of any machine learning system is understanding what it does not know. Unfortunately, today’s deep learning algorithms are usually unable to understand their uncertainty.
This talk will provide an introduction to the resurging filed of Bayesian Deep Learning. I would be discussing various theoretical aspects and the current state-of-the-art in this domain. I would be taking you through the code for constructing Bayesian deep nets and visualizing their uncertainty in their results using libraries like PyMC3, Pytorch.
The prerequisites for this talk familiarity with basic probability, Intermediary python, and passion to learn something new :)