Diagrammatic Differentiation for Quantum Machine Learning

TL;DR

This paper introduces a diagrammatic approach to differentiation in quantum machine learning, enabling gradient calculations for quantum circuits using monoidal categories and providing practical tools for implementation.

Contribution

It generalizes diagrammatic differentiation to quantum circuits, extending existing methods to hybrid classical-quantum systems with open-source software support.

Findings

Diagrammatic differentiation computes quantum circuit gradients.

Extension of the parameter-shift rule to hybrid circuits.

Open-source tools facilitate practical quantum machine learning applications.

Abstract

We introduce diagrammatic differentiation for tensor calculus by generalising the dual number construction from rigs to monoidal categories. Applying this to ZX diagrams, we show how to calculate diagrammatically the gradient of a linear map with respect to a phase parameter. For diagrams of parametrised quantum circuits, we get the well-known parameter-shift rule at the basis of many variational quantum algorithms. We then extend our method to the automatic differentation of hybrid classical-quantum circuits, using diagrams with bubbles to encode arbitrary non-linear operators. Moreover, diagrammatic differentiation comes with an open-source implementation in DisCoPy, the Python library for monoidal categories. Diagrammatic gradients of classical-quantum circuits can then be simplified using the PyZX library and executed on quantum hardware via the tket compiler. This opens the door to many practical applications harnessing both the structure of string diagrams and the computational power of quantum machine learning.