- This event has passed.
ESE Spring Seminar – “A Brief History of Feedback: Quantum Information Meets Machine Learning”
February 16 at 12:30 PM - 1:30 PM
A quest to harness quantum effects as a means to revolutionize information processing has spread across the globe. Public and private interests have coalesced making experimental progress remarkable. Despite such rapid experimental progress, a significant applications gap exists between theory and practice. Today’s quantum processors simply can not execute even mid-scale versions of the quantum algorithms and protocols found in textbooks. To circumvent hardware limitations at the quantum software level, a global research effort is developing a means to operate quantum processors through iterative quantum-to-classical feedback. In this talk, I will explain our discovery of several early limiting features of this approach. We discovered that problem instance structure can induce under parameterization in quantum approximate optimization, causing quantum approximate optimization to fail. Hence, we demonstrated that nascent findings were not statistically representative of typical problem instances. We also discovered avalanche effects in quantum circuit training, which provided the first counterexamples to the piecewise trainability conjecture. Several of our forward looking findings do cause optimism, including the first sufficient conditions for the problem instance independence of optimized circuit parameters—named parameter concentrations—and proving that the variational model is, in theory, a universal model of quantum computation. It is unclear how much further quantum hardware must progress to enable practical applications. What is clear is that quantum applications theory is changing dramatically with the recent introduction of feedback. At the end of the talk, I will mention some future prospects in this still unfolding “brief history of feedback.”
Jacob Biamonte
Jingdong Foundation Endowed Chair in Quantum Information and Computation, Yanqi Lake Beijing Institute of Mathematical Sciences and Applications
Throughout his career, Dr. Biamonte’s research has focused on theoretical studies of quantum information processing at the interface between quantum applications and physical hardware constraints.
Upon completing his undergraduate studies in electrical engineering at Portland State University, Dr. Biamonte worked as a quantum applications developer at D-Wave Systems Inc. in Vancouver, Canada and as a Fellow at Harvard University in the Aspuru-Guzik Group. Thereafter, he obtained a PhD from the University of Oxford in 2010. He then worked with John Carlos Baez as part of a joint Oxford/Singapore postdoctoral program and as a Lecturer at St Peter’s College Oxford before joining the Institute for Scientific Interchange (ISI Foundation) in Torino, Italy as their Quantum Science Research Group Leader (2012-2017). In 2017, Dr. Biamonte joined the MIT-founded Skolkovo Institute of Science and Technology as a Tenure Track Associate Professor. He was subsequently promoted to Head of the Laboratory for Quantum Information Processing and then to tenured Professor. Dr. Biamonte became the first American-born scientist to have successfully defended a higher Doctorate at the Moscow Institute of Physics and Technology. Dr. Biamonte’s time in Moscow was cut abruptly short due to the war against Ukraine.
Dr. Biamonte’s research vision is to develop a data-driven approach to quantify emergent and collective effects in quantum information processing tasks such as computing and sensing. A passionate educator, Dr. Biamonte teaches applied and engineering mathematics, quantum information theory and quantum computation.