“Uncertainty Quantification for Trained Machine Learning Models”
Wednesday, Feb. 15 at 3:00pm
NEB 409

Abstract

This talk will present a novel RKHS based methodology to estimate the epistemic uncertainty of a trained deep learning model. We use as inspiration concepts from quantum mechanics where the Schrodinger’s equation is used to decompose the wave equation in moments that are automatically placed in the domain of the wave equation. Therefore, we select the higher order moments placed at the tails. We employed the same methodology implemented in a RKHS with data obtained directly from the parameters of the trained model. Our results are on par with Bayesian uncertainty techniques.

Biography

Dr. José C. Príncipe is a Distinguished Professor of Electrical and Computer Engineering and Biomedical Engineering at the University of Florida where he teaches statistical signal processing, machine learning, and brain-computer interfaces. He is the Eckis Endowed Professor and the Founder and Director of the University of Florida Computational NeuroEngineering Laboratory (CNEL) www.cnel.ufl.edu. His primary area of interest is time series analysis in functional spaces, information theoretic learning and AI cognitive architectures, applied to neurotechnology.

Dr. Principe is an IEEE, AAAS, IABME, AIMBE and NDA Fellow. He was awarded the IEEE Neural Network Pioneer Award from the Computational Intelligence Society, the IEEE Shannon Nyquist Technical Achievement Award from the Signal Processing Society, the IEEE EMBS Career Achievement Award from the Engineering Medicine and Biology Society, and the Teacher Scholar of the Year from the University of Florida. He was the past Chair of the Technical Committee on Neural Networks of the IEEE Signal Processing Society, Past-President of the International Neural Network Society, and Past-Editor in Chief of the IEEE Transactions on Biomedical Engineering. Dr. Principe has more than 900 publications and an H index of 94 (Google Scholar). He directed 106 Ph.D. dissertations and 65 Master theses. He wrote in 2000 an interactive electronic book entitled “Neural and Adaptive Systems” published by John Wiley and Sons and more recently co-authored several books on “Brain Machine Interface Engineering” Morgan and Claypool, “Information Theoretic Learning”, Springer, and “Kernel Adaptive Filtering”, Wiley.