Abstract: Every field of science and engineering starts with measurements. To quote Siddhartha Mukherjee in The Emperor of All Maladies, ""Science begins with counting. To understand a phenomenon, a scientist must first describe it; to describe it objectively, he must first measure it.""
Unfortunately, when working on supervised machine learning problems, modern data science experts rely more on computation (let's throw more GPUs at the problem) and guesswork (let's see if we can modify Alexnet for our specific problem) than on any kind of engineering measurements. This is because, until now, no one has developed an effective framework for measuring the information density (learnability) of the data sets that underlie all machine learning models.
In this talk, we will discuss an entirely new and different approach to supervised machine learning - one that is rooted in measurements. We will explain how this new approach (which is actually as old as science itself) can be used to solve difficult machine learning problems, many of which have previously been out of reach. We will explain the fundamentals of how measurements-based machine learning works, and also explore how the approach can be applied to solve real-world problems in bioinformatics and other fields.
This measurements-based approach builds on the work of Claude Shannon, John Hopfield, and David MacKay. It incorporates techniques from statistical mechanics, information theory and computer science. The theory and practice of the approach is explained in this paper (https://arxiv.org/abs/1810.02328), and in this YouTube lecture (https://www.youtube.com/watch?v=UZ5vhqDKyrY), both of which would be helpful background material to absorb before joining the webinar.
Bio: Dr. Gerald Friedland is the CTO of Brainome, Inc and is also teaching as an adjunct professor in the Electrical Engineering and Computer Sciences department of UC Berkeley. Before that, Dr. Friedland was at Lawrence Livermore National Lab and the International Computer Science Institute in Berkeley. Dr. Friedland's work is primarily in the areas of signal processing and machine learning. Dr. Friedland has published more than 250 peer-reviewed articles in conferences, journals, and 3 books. Dr. Friedland received his doctorate (summa cum laude) in computer science from Freie Universitaet Berlin, Germany, in 2006.