Evidencing quantum effects in machine learning
Since the advent of quantum machine learning, a large focus has been placed on finding computational advantage, often in the form of speed ups or intractability. While this goal has clear motivation, there is usually little-to-no understanding of exactly what quantum effect produces said advantage, and how. In general, studying quantum effects is not straightforward, and it is particularly difficult in a machine learning setting, where things like data loading and non-convex optimization come into play. In this talk, we introduce a simple quantum machine learning model from which a particular criterion can be extracted and used to understand if the model actually needs anything quantum to learn. The purpose of this is to attempt to build a framework where one can directly study the contribution of quantum effects in machine learning tasks, in order to design more meaningful quantum machine learning algorithms.
Watch more Quantum Summer Symposium 2022 sessions → https://goo.gle/QSS22
Subscribe to Google Quantum AI → https://goo.gle/QuantumAI
#QSS22
