Quantum neural ordinary and partial differential equations

Abstract: We present a unified framework called Quantum Neural Ordinary and Partial Differential Equations (QNODEs and QNPDEs) that brings the continuous-time formalism of classical neural ODEs/PDEs into quantum machine learning and quantum control. We define QNODEs as the evolution of finite-dimensional quantum systems, and QNPDEs as infinite-dimensional (continuous-variable) counterparts, governed by generalised Schrodinger-type Hamiltonian dynamics with unitary evolution, coupled with a corresponding loss function. Notably, this formalism permits gradient estimation using an adjoint-state method, facilitating efficient learning of quantum dynamics, and other dynamics that can be mapped (relatively easily) to quantum dynamics. Using this method, we present quantum algorithms for computing gradients with and without time discretisation, enabling efficient gradient computation that would otherwise be less efficient on classical devices. The formalism subsumes a wide array of applications, including quantum state preparation, Hamiltonian learning, learning dynamics in open systems, and the learning of both autonomous and non-autonomous classical ODEs and PDEs. In many of these applications, we consider scenarios where the Hamiltonian has relatively few tunable parameters, yet the corresponding classical simulation remains inefficient, making quantum approaches advantageous for gradient estimation. This continuous-time perspective can also serve as a blueprint for designing novel quantum neural network architectures, generalising discrete-layered models into continuous-depth models.