© Pint of Science, 2024. All rights reserved.
Join us in learning how computers see the World around us, from data analysis to the limitation of AI.
The importance of Why? And what huge AI models might still be failing on.
Elisabeth Ailer
(researcher)
Does sometimes ever more powerful algorithms make you feel bad? Now it is time to boost our self-esteem: Humans are still way smarter! We are able to reason. But those learning algorithms have extensive potential. However we need a reason for the algorithm’s decision, especially when applied in an ethically tricky context. Why does this MRI scan look like I have cancer? or Why am I not creditworthy? This talk wants to show the difference between correlation and causation and how algorithms try to be smart, trustworthy and most importantly: tell you why.
Why Mathematicians Confuse Coffee Cups With Doughnuts
Bastian Rieck
(researcher)
Nothing is dry and boring in the field of topology, the art of understanding shapes. With its visual proofs and mind-bending theorems, topology received a new life with the rise of ubiquitous compute power for everyone. In this talk, I will show the beauty of the nascent field of computational topology and how topology can help us make sense of complicated data sets. Along the way, we will discover why it can sometimes be helpful to assume the perspective of a little bug, what proof by contradiction is, and why doughnuts and coffee are critical ingredients of mathematical progress in general.
Map data © OpenStreetMap contributors.