TorchQC -- A framework for efficiently integrating machine and deep learning methods in quantum dynamics and control

Abstract

Machine learning has been revolutionizing our world over the last few years and is also increasingly exploited in several areas of physics, including quantum dynamics and control.The need for a framework that brings together machine learning models and quantum simulation methods has been quite high within the quantum control field, with the ultimate goal of exploiting these powerful computational methods for the efficient implementation of modern quantum technologies. The existing frameworks for quantum system simulations, such as QuTip and QuantumOptics.jl, even though they are very successful in simulating quantum dynamics, cannot be easily incorporated into the platforms used for the development of machine learning models, like for example PyTorch. The TorchQC framework introduced in the present work comes exactly to fill this gap. It is a new library written entirely in Python and based on the PyTorch deep learning library. PyTorch and other deep learning frameworks are based on tensors, a structure that is also used in quantum mechanics. This is the common ground that TorchQC utilizes to combine quantum physics simulations and deep learning models.TorchQC exploits PyTorch and its tensor mechanism to represent quantum states and operators as tensors, while it also incorporates all the tools needed to simulate quantum system dynamics. All necessary operations are internal in the PyTorch library, thus TorchQC programs can be executed in GPUs, substantially reducing the simulation time. We believe that the proposed TorchQC library has the potential to accelerate the development of deep learning models directly incorporating quantum simulations, enabling the easier integration of these powerful techniques in modern quantum technologies.