Dougal Maclaurin

TL;DR: In this paper, a convolutional neural network that operates directly on graphs is proposed to learn end-to-end learning of prediction pipelines whose inputs are graphs of arbitrary size and shape.

TL;DR: An integrated organic functional material design process that incorporates theoretical insight, quantum chemistry, cheminformatics, machine learning, industrial expertise, organic synthesis, molecular characterization, device fabrication and optoelectronic testing is reported.

TL;DR: In this article, a coexpression vector, Optopatch, enabled cross-talk-free genetically targeted all-optical electrophysiology without the use of conventional electrodes.

TL;DR: The experiments reported here demonstrate the viability of controlled single spin probes for nanomagnetometry in biological systems, opening up a host of new possibilities for quantum-based imaging in the life sciences.

TL;DR: In this article, the authors compute exact gradients of cross-validation performance with respect to all hyperparameters by chaining derivatives backwards through the entire training procedure, allowing them to optimize thousands of hyperparameter, including step-size and momentum schedules, weight initialization distributions, richly parameterized regularization schemes, and neural network architectures.