March 29 2017

5.4: Ctrl-Q Lecture on "Quantum Machine Learning" by V. Dunjko

Dr. Vedran Dunjko, University of Innsbruck
Wednesday April 5th at 11:00, in HS001 E1.3, Computer Science Department

Title: Machine learning and Quantum Information Processing: a perfect match

Abstract:

The nascent field of Quantum Machine Learning has been generating a substantial buzz in the last few years.
The broad theme of this field is the interplay between the disciplines of quantum information processing (QIP) and of machine learning (ML). The research is thus typically driven by one of two basic questions. The first question focuses on the ways in which QIP can help in ML problems. The complementary line of research studies the extent to which ML can be beneficially applied in QIP tasks. In this overview talk, I will present the basic ideas behind quantum computing and information processing. I will draw parallels between features of QIP and aspects of machine learning, which suggest that quantum effects may play an integral role in improved learning algorithms. While some of the features require full blown quantum computers, some can be addressed using near term devices.
This will be illustrated through a selection of recent results which probe the potential and limitations of quantum-enhanced learning, followed by a snapshot of fresh proposals addressing the complementary question of exploiting ML techniques in quantum experiments.
These results suggest not only that (Q)ML applications may be among the best reasons to build quantum computers in the first place (barring perhaps quantum simulations and cryptography), but also that ML may significantly help bringing about large-scale quantum computers. I will finish the talk with a perspective on the field we have developed in Innsbruck, which also touches the arguably broader topic of the interplay of artificial general intelligence and quantum mechanics.
--



back to overview of all news