• Venue: Large Lecture Theatre and Social Area, Department of Statistics, Oxford
• Venue: 7 May 2026, time TBC (tentatively 5pm)
• Title: From Classical to Quantum Learning Theory
• Abstract: Learning theory—the foundation of much of modern machine learning and artificial intelligence—studies which learning problems can be solved with efficient amounts of data and computation. Traditionally, this field has been developed in the context of classical computers accessing data stored in classical memory. Motivated by rapid advances in quantum algorithms and quantum hardware, the still young field of quantum learning theory seeks to extend learning problems to settings where data and computation can be quantum. This tutorial, which assumes no prior background in quantum computing, will give an introduction to two key directions in this vibrant research area. After a brief crash course in quantum computation, I will first discuss how quantum data and quantum computational resources can affect the sample complexity and the computational complexity in computational learning theory—highlighting what quantum computation has to offer to learning theorists. I will then turn to the converse perspective and present examples of how ideas from learning theory can inform quantum computing, including how to efficiently predict properties of quantum systems, the role of memory in processing quantum data, and how to learn physical interactions from data collected in quantum experiments.
