Please note: This seminar will take place in DC 1304.
Amanda Xu, PhD candidate
Computer Sciences Department, University of Wisconsin–Madison
Reducing the timeline for solving useful problems with quantum computing requires a flexible compilation stack that can adapt to emerging hardware. Optimizing compilers ensure that the programs we run on these error-prone devices obtain the intended output, while also informing resource estimates that guide the design of new quantum algorithms and hardware. But the diverse and rapidly evolving landscape of competing quantum architectures makes manual compiler development untenable. Each architecture is tuned to support a unique set of physical operations and connectivity constraints, which define a new compilation problem.
I will present my work on automatically synthesizing an architecture-dependent phase of a quantum compiler—a circuit optimizer—given a specification of the target device. I will show how in under two minutes, we can synthesize an optimizer with correctness guarantees that outperforms six state-of-the-art handcrafted alternatives at reducing overall error across five diverse architectures. Moving forward, I envision program synthesis and verification as critical components of building a software stack capable of unlocking the full potential of quantum computing.
Bio: Amanda Xu is a Ph.D. candidate at the University of Wisconsin-Madison advised by Aws Albarghouthi. Her research navigates the challenges in correctly and effectively compiling to rapidly evolving quantum hardware. Her work has appeared at top programming languages and architecture conferences and has been recognized with the ACM SIGPLAN John Vlissides Award, a Rising Stars in EECS Award, and a Cisco Systems Distinguished Graduate Fellowship. She received her B.S. in Computer Science from Cornell University in 2020 and has applied her expertise at Google Quantum AI and Amazon Web Services.