Olivier Pfister received the B.S. in Physics from Université Côte d’Azur in 1987, and the M.S. and the Ph.D. in Physics from Sorbonne Paris-Nord Université in 1989 and 1993. In 1994, he was a lecturer at Conservatoire National des Arts et Métiers. He was a research associate with John L. Hall at JILA, University of Colorado (1994-97) and with Daniel J. Gauthier at Duke University (1997-99). In 1999, he joined the faculty of the University of Virginia, where he is a professor of physics with a courtesy appointment in electrical and computer engineering. Olivier Pfister is a fellow of the American Physical Society and a member of Optica, IEEE, and SPIE. His general research area is quantum optics and quantum information, with a focus on quantum computing with continuous variables of light.
Presentation Title:
AI-driven, deterministic non-Gaussian resource generation for fully scalable photonic quantum computing
Presentation Abstract:
Photonic quantum computing (QC) over continuous variables (CV) has the most scalability promise of all known QC platforms (including qubit-based ones), leveraging deterministic Gaussian quantum state generation to produce record-scale CV cluster states which are a substrate for universal quantum computing. Reaching fault tolerance, however, will require the generation of non-Gaussian resources, e.g. Gottesman-Kitaev-Preskill (GKP) CV qubit encodings. While recent experimental progress has been made in generating GKP states, the process is still probabilistic and cannot scale without quantum memories. I will present recent theoretical results that show that deterministic non-Gaussian state generation is possible using deep reinforcement learning, which hints that hybrid AI-quantum circuits might lift this last scalability bottleneck.