Search

A Hessian based optimal control method is presented in Liouville space to mitigate previously undesirable polynomial scaling of computation time. This new method, an improvement to the state-of-the-art Newton-Raphson GRAPE method, is derived with respect to two exact time-propagator derivative techniques: auxiliary matrix and ESCALADE methods. We observed that compared to the best current implementation of Newton-Raphson GRAPE method, for an ensemble of 2-level systems, with realistic conditions, the new auxiliary matrix and ESCALADE Hessians can be 4-200 and 70-600 times faster, respectively.

Conventional refocusing pulses are optimised for a single spin without considering any type of coupling. However, despite the fact that most couplings will result in undesired distortions, refocusing in delay-pulse-delay-type sequences with desired heteronuclear coherence transfer might be enhanced considerably by including coupling evolution into pulse design. We provide a proof of principle study for a Hydrogen-Carbon refocusing pulse sandwich with inherent J-evolution following the previously reported ICEBERG-principle with improved performance in terms of refocusing performance and/or overall effective coherence transfer time. Pulses are optimised using optimal control theory with a newly derived quality factor and z-controls as an efficient tool to speed up calculations. Pulses are characterised in detail and compared to conventional concurrent refocusing pulses, clearly showing an improvement for the J-evolving pulse sandwich. As a side-product, also efficient J-compensated refocusing pulse sandwiches -- termed BUBU pulses following the nomenclature of previous J-compensated BUBI and BEBE(tr) pulse sandwiches -- have been optimised., Comment: 17 pages, 5 figures

We have recently demonstrated supervised deep learning methods for rapid generation of radiofrequency pulses in magnetic resonance imaging (https://doi.org/10.1002/mrm.27740, https://doi.org/10.1002/mrm.28667). Unlike the previous iterative optimization approaches, deep learning methods generate a pulse using a fixed number of floating-point operations - this is important in MRI, where patient-specific pulses preferably must be produced in real time. However, deep learning requires vast training libraries, which must be generated using the traditional methods, e.g. iterative quantum optimal control methods. Those methods are usually variations of gradient descent, and the calculation of the fidelity gradient of the performance metric with respect to the pulse waveform can be the most numerically intensive step. In this communication, we explore various ways in which the calculation of fidelity gradients in quantum optimal control theory may be accelerated. Four optimization avenues are explored: truncated commutator series expansions at zeroth and first order, a novel midpoint truncation scheme at first order, and the exact complex-step method. For the spin systems relevant to MRI, the first-order truncation is found to be sufficiently accurate, but also up to five times faster than the machine precision gradient. This makes the generation of training databases for the machine learning methods considerably more realistic., Comment: The following article has been submitted to the Journal of Chemical Physics. After it is published, it will be found at https://aip.scitation.org/journal/jcp

Organisations are increasingly putting machine learning models into production at scale. The increasing popularity of serverless scale-to-zero paradigms presents an opportunity for deploying machine learning models to help mitigate infrastructure costs when many models may not be in continuous use. We will discuss the KFServing project which builds on the KNative serverless paradigm to provide a serverless machine learning inference solution that allows a consistent and simple interface for data scientists to deploy their models. We will show how it solves the challenges of autoscaling GPU based inference and discuss some of the lessons learnt from using it in production., Comment: 4 pages, 1 figure, presented at workshop on "Challenges in Deploying and Monitoring Machine Learning System" at ICML 2020

We show that the duration of broadband universal control pulses can be halved by choosing control targets with a quadratic function of phase dispersion. This class of control pulses perform a broadband universal rotation around an axis, in the Bloch sphere representation of two-level systems, given by this phase dispersion function. We present an effective optimal control method to avoid the problem of convergence to local extrema traps., Comment: 8 pages, 4 figures