Approximation theory, AI, and how they actually go together (talk at UNCC)
Updated: Jan 15, 2018
Here is a video of my talk at the University of North Carolina at Charlotte (01.12.2018). I show how at the beginning of AI, Deep Learning networks departed from the foundations of Approximation Theory. It includes multiresolution analysis, a sequence of nested spaces, optimal bases and two-scale relation.
What is interesting, now I can find many principles of Approximation Theory adopted in modern AI systems. Perhaps it makes sense to look at Approximation Theory for inspiration for future modernizing of AI.
Besides results from my publications, in this talk I also used following references (in order of appearance in my talk):
Shai Dekel https://www.shaidekel.com/
Konstantin Aslanidi http://www.opentradingsystem.com/
Andrew Ng http://www.andrewng.org/
Geoffrey Hinton http://www.cs.toronto.edu/~hinton/