This lesson explains how Neural Networks work starting from a simple example: how does Facebook recognise somebody in a picture? The lesson points to important but often overlooked theoretical works. These works explain the success behind deep learning.
Article on Quanta Magazine by Duke University’s Professor Ingrid Daubechies – https://www.quantamagazine.org/20151203-big-datas-mathematical-mysteries/
Links to papers:
+ Neural Networks as Universal Approximators – http://deeplearning.cs.cmu.edu/pdfs/Kornick_et_al.pdf
+ On the Power of Small-Depth Threshold Units – https://www.computer.org/csdl/proceedings/focs/1990/2082/00/089582.pdf
video on the Universality of Neural Networks https://youtu.be/7MIrrQ7pwUI
Speaker: El Mahdi El Mhamdi
Editor: Flora Bressieux
Photo credit:
Morgan Freeman, from the following Creative Commons Licenced video https://www.youtube.com/watch?v=1tPSH11mtE4
UCD_VizaraVT9yDy1egNlWCQ
source
