Quantum computing with differentiable quantum transforms

As quantum computing increases in scale and scope, the manner in which we write quantum algorithms is necessarily changing. While many situations still require a gate-level approach and knowledge of the finer details of both software and hardware, higher-level abstractions are being developed that enable programmers to redirect focus to algorithms and applications rather than what happens under the hood. This talk will focus on the concept of differentiable quantum transforms, which are metaprograms that manipulate quantum programs in a way that preserves their differentiability. We will explore how transforms can be used to think about quantum programs in a different way, and showcase their practical implementation for three applications: differentiable compilation, noise characterization, and error mitigation. We conclude by highlighting a recent application in condensed-matter physics, the computation of fidelity susceptibility, in which transforms are used to facilitate the study of the effects of error mitigation on physical quantities computed from quantum gradients.