Convolutional neural networks, inspired by biological visual cortex systems, are a powerful category of artificial neural networks that can extract the hierarchical features of raw data to provide greatly reduced parametric complexity and to enhance the accuracy of prediction. They are of great interest for machine learning tasks such as computer vision, speech recognition, playing board games and medical diagnosis. Optical neural networks offer the promise of dramatically accelerating computing speed using the broad optical bandwidths available. Here we demonstrate a universal optical vector convolutional accelerator operating at more than ten TOPS (trillions (1012) of operations per second, or tera-ops per second), generating convolutions of images with 250,000 pixels—sufficiently large for facial image recognition. We use the same hardware to sequentially form an optical convolutional neural network with ten output neurons, achieving successful recognition of handwritten digit images at 88 per cent accuracy. Our results are based on simultaneously interleaving temporal, wavelength and spatial dimensions enabled by an integrated microcomb source. This approach is scalable and trainable to much more complex networks for demanding applications such as autonomous vehicles and real-time video recognition.

Read more at Nature: https://www.nature.com/articles/s41586-020-03063-0

Electrostatic interactions between nanoparticles (NPs) and functionalized ligands lead to the formation of NP surfactants (NPSs) that assemble at the water-oil interface and form jammed structures. To understand the interfacial behavior of NPSs, it is necessary to understand the mechanism by which the NPSs attach to the interface and how this attachment depends on the areal coverage of the interface. Through direct observation with high spatial and temporal resolution, using laser scanning confocal microscopy and in situ atomic force microscopy (AFM), we observe that early-stage attachment of NPs to the interface is diffusion limited and with increasing areal density of the NPSs, further attachment requires cooperative displacement of the previously assembled NPSs both laterally and vertically. The unprecedented detail provided by in situ AFM reveals the complex mechanism of attachment and the deeply nonequilibrium nature of the assembly.

The work is featured on the cover of the Nov 2020 volume of Science Advances (Volume 6, Issue 48).

Read more at Science Advances: https://advances.sciencemag.org/content/6/48/eabb8675

The paper has also featured in ScienceDaily:
https://www.sciencedaily.com/releases/2021/02/210225082555.htm

Moreover, the paper was highlighted in a press release by Lawrence Berkeley National Laboratory:
https://newscenter.lbl.gov/2021/02/25/nanoparticles-get-in-shape/