RANDOM TESTING ACROSS THE QUANTUM COMPUTING STACK
Abstract
Quantum computing is notoriously hard. From both a hardware and software angle, it s difficult to write and execute a correct quantum program. Past efforts in formal verification of both quantum programs and quantum compilers have verified a small set of programs and optimization techniques, but this work has been slow, painstaking, and fundamentally limited- It has focused on a single abstraction layer, the quantum circuit.Actual quantum programs feature an interplay of classical and quantum computation. They can run on a variety of devices, frequently using transmon qubits, neutral atoms, or trapped ions. While these architectures can approximate the quantum circuit model on common gate sets, the circuit model is not native to it. Many promising optimizations focus instead on lower level microwave pulses, or even device-specific optimizations. Modeling these optimizations mathematically and verifying their behavior is near infeasible in a reasonable timespan. Lighter weight techniques like random testing, on the other hand, offer far more promise. We intend to use random testing to establish confidence in the correctness of programs and opĂ‚Âtimizations that bridge the quantum stack. We will take advantage of both simulators and quantum hardware to expose programs to a range of tests that are carefully calibrated to uncover potential bugs. This will allow us to build faster and more reliable quantum programs, and grant researchers a powerful new tool to explore future algorithms and optimizations.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 06, 2024
- Source ID
- FA95502310406
Entities
People
- Leonidas Lampropoulos
Organizations
- Air Force Office of Scientific Research
- United States Air Force
- University of Maryland