Event Details

MHI - Physics Joint Seminar Series - Andrew Vlasic, Friday, November 8th at 2pm in SSL 202

Fri, Nov 08, 2024
2:00 PM - 3:30 PM
Location: SSL 202
Speaker: Andrew Vlasic, PhD, Fundamental Research Lead Quantum Institute, Deloitte Consulting LLP

Talk Title: A Categorical Perspective of Encoding Real-World Data in Quantum Computers

Series: MHI Physics Joint Seminar Series

Abstract: The question of how to encode real-world data in quantum computer has a tremendous amount of importance in the quantum machine learning (QML) community. There are a few proposed metrics to quantify the efficacy of quantum feature maps with the most used criteria being 'expressibility' [1] and 'expressivity' [2]. However, as noted by the authors, there are shortcomings with these two techniques. Our empirical analysis of using the standard schemes of angle encoding, instantaneous quantum polynomial encoding (IQP), and amplitude encoding to perform machine learning tasks on different dataset reveals new insights into our metrics need to be considered when choosing a particular quantum encoding technique [3,4]. Using the perspective of category theory, we propose that quantum encoding techniques should preserve "structures" of classical data. Based on this insight, we proposed technique on comparing the entropy of a point-cloud against the analytic extension of von Neuman entropy applied to quantum operators [5], directly addressing one area of structure. [5] 

Biography: Andrew earned a PhD in mathematics from the University of Illinois at Urbana-Champaign and has extensive experience in fundamental and applied research in the academia, DoD, and industry. Andrew has been a postdoc at Queen's University in Ontario, an acting funding officer at the Army Research Office, a senior data scientist at Bank of America, and is currently the fundamental research lead in the Quantum Research Group at Deloitte Consulting. If interested, view Andrew's portfolio on Google Scholar.

Host: Quntao Zhuang, Eli Levinson-Falk, Jonathan Habif, Daniel Lidar, Kelly Luo, Todd Brun, Tony Levi, Stephan Haas