Information Lattices for Interpretable Knowledge Discovery and Nearest-Neighbor Classification

Presenter: Lav R Varshney, Department of Electrical and Computer Engineering, University of Illinois

Tuesday, October 26, 2021

Slides: https://uofi.app.box.com/s/gxl8hoqi00xngvhiacastbxv2acekpqm

Video: https://uofi.app.box.com/s/x2vmrvza00xc6v99y25aef244knrzd10

Abstract

We discuss a novel, human-interpretable machine learning technique for knowledge discovery and nearest-neighbor classification. This non-neural network technique is called information lattice learning, and draws on mathematical techniques from information theory and group theory. We show how it achieves state-of-the-art performance in image classification in the regime of limited training data per class, and also how it can applied to geospatial data.

Biography

Lav R. Varshney is an associate professor in the Department of Electrical and Computer Engineering and the Coordinated Science Laboratory at the University of Illinois Urbana-Champaign, where his research interests include information theory, artificial intelligence, network science, and their applications, as well as AI ethics. He has further affiliations in computer science, neuroscience, industrial engineering, digital agriculture, and personalized nutrition, and with the Discovery Partners Institute of the University of Illinois System. He is chief scientist for AI for the IBM-ILLINOIS Center for Cognitive Computing Systems Research, as well as chief scientist for Ensaras, Inc., a startup company working at the intersection of wastewater treatment and AI. His research is supported by numerous governmental, industrial, and philanthropic sources.