Optical techniques adopted in optical computing rely on spatial multiplexing, requiring numerous integrated elements and restricting the architecture to perform a single kernel convolution per layer. The authors demonstrate a fiber-optic computing architecture based on temporal multiplexing that performs multiple convolutions in a single layer.

The mechanism behind the transient emergence of superconductivity upon irradiation with light in some materials is highly debated, which is in part due to the strong correlations at play. Here, the authors investigate the dynamical emergence of superconductivity in the strongly correlated, yet exactly solvable, Yukawa-Sachdev-Ye-Kitaev model and discuss differences and similarities to the relaxation dynamics of conventional superconductors.

Developing cryogenic memory is a crucial undertaking in advancing superconducting logic systems. Here, the authors demonstrate the realization of a superconducting memory cell, whose state is encoded by the number of current vortices within a Josephson junction, and the readout employs microwave currents for an energy-efficient, non-destructive process.

Optical beams carrying orbital angular momentum (OAM) are promising candidates for free-space optical communication. The authors devise a hybrid optical-electronic convolutional neural network approach reaching a 4-bit OAM-coded signal demultiplexing accuracy of 72.84% under strong atmospheric turbulence conditions with 3.2 times faster training time than all electronic convolutional neural network.

Spatial Kramers–Kronig (KK) media are inhomogeneous materials enabling omnidirectional light absorption, but the successful experimental realizations are polarization-dependent, i.e., they absorb either transverse electric or transverse magnetic fields. Using a matryoshka metamaterial, the authors report the experimental realization of a polarization-independent omnidirectional absorber.

The theoretical description of ultra-cold Fermi gases is challenging due to the presence of strong, short-ranged interactions. This work introduces a Pfaffian-Jastrow neural-network quantum state that outperforms existing Slater-Jastrow frameworks and diffusion Monte Carlo methods.

The strongly correlated interactions found in twisted bilayer systems give rise to a range of exotic phenomena but due to range of factors it is challenging to develop theoretical models that can faithfully describe the underlying physics. Here, the authors provide an analysis of the chiral limits of perturbed Dirac field theories that are relevant to C3-symmetric twisted bilayer graphene and transition metal dichalcogenides homobilayers. They identify two possible scenarios and show that flat bands are a more likely phenomenon in the latter system

Quantum simulators study important models of condensed matter and high-energy physics. Research on synthetic dimensions has paved the way for studying exotic phenomena, such as curved space-times, topological phases of matter, lattice gauge theories, twistronics without a twist, and more

The frameworks to simulate pathogen, malware and failure spreading are computationally demanding, and they are subject to large statistical uncertainty. The authors develop efficient inference and control algorithms based on dynamic message passing to study a two-layer spreading process, where the spreading infection triggers cascading failures and leads to secondary disasters.

Quantum Speed Limit (QSL) is a lower bound on the time evolution for quantum systems. Still, experimental studies for open systems are few due to the lack of control over their environment’s interaction. The authors control the qubitreservoir interaction in an ensemble of chloroform molecules, observing crossovers between different QSLs.

From the cardiac system of various human patients, to changes in sea surface temperature across different oceans, dynamical systems often exhibit many common characteristics. Here we develop a framework for jointly learning the dynamics of multiple interrelated systems while leveraging their shared traits.

Reconstructing unstable heavy particles is a crucial aspect of many analyses at the Large Hadron Collider (LHC). We introduce SPA-Net, a machine-learning approach to this problem which outperforms existing baseline methods, performs several auxiliary tasks, and leads to significant improvements in three example flagship LHC analyses

Non-Hermitian (NH) systems have recently attracted great attention, and the NH bulk-boundary correspondence is a key question. The authors propose an approach to restore NH bulk-boundary correspondence by a “doubling and swapping" construction which eliminates the NH skin effect, resulting in a new NH Hamiltonian with the same energy spectrum.