CAIML #18 - Uncertainty Quantification in Computer Vision and Robotics

CAIML #18 will be an online meetup and it will happen on May 18, 2021, 18:30. We will have a talk by Matias Valdenegro-Toro, a Q&A and discussion with Matias after the talk, and a networking session.

-----

Matias Valdenegro-Toro (Researcher at German Research Center for Artificial Intelligence (DFKI) , https://www.linkedin.com/in/matias-valdenegro-toro-29383717): Uncertainty Quantification in Computer Vision and Robotics

What if we train a model to classify dogs and cats, but it is later tested with an image of a human? Generally the model will output either dog or cat, and has no ability to signal that the image has no class that it can recognize.

This is because classical neural networks do not contain ways to estimate their own uncertainty (so called epistemic uncertainty), and this has practical consequences for the use of these models, like safety when cooperating with humans, autonomous systems like robots, and computer vision systems. A possible solution is the bayesian neural network.

In this talk I will cover the basic concepts of bayesian neural networks, and how they can help us to produce safer robots, with respect to their visual perception and control. I will include examples from the literature and my own research.

-----

Agenda

18h30: Start of the meetup via MS Teams and short introduction
18h40: Matias' talk
19h10: Collect questions for Q&A with Matias via sli.do
19h15: 15 minutes Q&A
19h30: Networking Session via Wonder

-----

We will share the full agenda soon and the link to the online event will be shared shortly before the meetup itself.