Search

Neural networks are a central technique in machine learning. Recent years have seen a wave of interest in applying neural networks to physical systems for which the governing dynamics are known and expressed through differential equations. Two fundamental challenges facing the development of neural networks in physics applications is their lack of interpretability and their physics-agnostic design. The focus of the present work is to embed physical constraints into the structure of the neural network to address the second fundamental challenge. By constraining tunable parameters (such as weights and biases) and adding special layers to the network, the desired constraints are guaranteed to be satisfied without the need for explicit regularization terms. This is demonstrated on upervised and unsupervised networks for two basic symmetries: even/odd symmetry of a function and energy conservation. In the supervised case, the network with embedded constraints is shown to perform well on regression problems while simultaneously obeying the desired constraints whereas a traditional network fits the data but violates the underlying constraints. Finally, a new unsupervised neural network is proposed that guarantees energy conservation through an embedded symplectic structure. The symplectic neural network is used to solve a system of energy-conserving differential equations and out-performs an unsupervised, non-symplectic neural network., Comment: This is the same manuscript with version 1 (arXiv:1904.08991v1) which accidentally was replaced 16 pages, 8 figures

Chimeras and branching are two archetypical complex phenomena that appear in many physical systems; because of their different intrinsic dynamics, they delineate opposite non-trivial limits in the complexity of wave motion and present severe challenges in predicting chaotic and singular behavior in extended physical systems. We report on the long-term forecasting capability of Long Short-Term Memory (LSTM) and reservoir computing (RC) recurrent neural networks, when they are applied to the spatiotemporal evolution of turbulent chimeras in simulated arrays of coupled superconducting quantum interference devices (SQUIDs) or lasers, and branching in the electronic flow of two-dimensional graphene with random potential. We propose a new method in which we assign one LSTM network to each system node except for "observer" nodes which provide continual "ground truth" measurements as input; we refer to this method as "Observer LSTM" (OLSTM). We demonstrate that even a small number of observers greatly improves the data-driven (model-free) long-term forecasting capability of the LSTM networks and provide the framework for a consistent comparison between the RC and LSTM methods. We find that RC requires smaller training datasets than OLSTMs, but the latter requires fewer observers. Both methods are benchmarked against Feed-Forward neural networks (FNNs), also trained to make predictions with observers (OFNNs).

Van der Waals (vdW) heterostructures are an emergent class of metamaterials comprised of vertically stacked two-dimensional (2D) building blocks, which provide us with a vast tool set to engineer their properties on top of the already rich tunability of 2D materials. One of the knobs, the twist angle between different layers, plays a crucial role in the ultimate electronic properties of a vdW heterostructure and does not have a direct analog in other systems such as MBE-grown semiconductor heterostructures. For small twist angles, the moir\'e pattern produced by the lattice misorientation creates a long-range modulation. So far, the study of the effect of twist angles in vdW heterostructures has been mostly concentrated in graphene/hexagonal boron nitride (h-BN) twisted structures, which exhibit relatively weak interlayer interaction due to the presence of a large bandgap in h-BN. Here we show that when two graphene sheets are twisted by an angle close to the theoretically predicted 'magic angle', the resulting flat band structure near charge neutrality gives rise to a strongly-correlated electronic system. These flat bands exhibit half-filling insulating phases at zero magnetic field, which we show to be a Mott-like insulator arising from electrons localized in the moir\'e superlattice. These unique properties of magic-angle twisted bilayer graphene (TwBLG) open up a new playground for exotic many-body quantum phases in a 2D platform made of pure carbon and without magnetic field. The easy accessibility of the flat bands, the electrical tunability, and the bandwidth tunability though twist angle may pave the way towards more exotic correlated systems, such as unconventional superconductors or quantum spin liquids., Comment: Main Text: 14 pages, 4 figures. Supplemental Information: 24 pages, 14 figures

Twisted bilayer graphene (TwBLG) is one of the simplest van der Waals heterostructures, yet it yields a complex electronic system with intricate interplay between moir\'{e} physics and interlayer hybridization effects. We report on electronic transport measurements of high mobility small angle TwBLG devices showing clear evidence for insulating states at the superlattice band edges, with thermal activation gaps several times larger than theoretically predicted. Moreover, Shubnikov-de Haas oscillations and tight binding calculations reveal that the band structure consists of two intersecting Fermi contours whose crossing points are effectively unhybridized. We attribute this to exponentially suppressed interlayer hopping amplitudes for momentum transfers larger than the moir\'{e} wavevector., Comment: 5 pages, 4 figures

We use the results of first-principles electronic structure calculations and a strong coupling perturbation approach, together with general theoretical arguments, to illustrate the differences in super-exchange interactions between the copper-oxides and iron-pnictides. We show that the two magnetic ground states can be understood in a simple manner within the same theoretical foundation. Contrary to the emerging view that magnetic order in the iron-pnictides is of itinerant nature, we argue that the observed magnetic moment is small because of frustration introduced by the electrons of the Fe orbitals as they compete to impose their preferred magnetic ordering., Comment: 5 pages, 5 figures

We present first-principles calculations in the framework of density-functional theory and the pseudopotential approach, aiming to model the intermediate stages of the reduction of NO in the presence of MoO$_3$(010). In particular, we study the formation of dinitrosyl, which proves to be an important intermediate stage in the catalytic reduction. We find that the replacement of an oxygen of MoO$_3$ by NO is energetically favorable, and that the system lowers further its energy by the formation of (NO)$_2$. Moreover, the geometry and charge distribution for the adsorbed dinitrosyl indicates a metal-oxide mediated coupling between the two nitrogen and the two oxygen atoms. We discuss the mechanisms for the dinitrosyl formation and the role of the oxide in the reaction., Comment: 6 pages, 4 figs, RevTeX. To be published in J. Chem. Phys

The role of geometrical micro-barriers on the conversion efficiency of reactive flows in narrow three-dimensional channels of millimetric size is investigated. Using a Lattice-Boltzmann-Lax-Wendroff code, we show that micro-barriers have an appreciable effect on the effective reaction efficiency of the device. If extrapolated to macroscopic scales, these effects can result in a sizeable increase of the overall reaction efficiency., Comment: 5 pages, 7 figures

We investigate the effects of geometrical micro-irregularities on the conversion efficiency of reactive flows in narrow channels of millimetric size. Three-dimensional simulations, based upon a Lattice-Boltzmann-Lax-Wendroff code, indicate that periodic micro-barriers may have an appreciable effect on the effective reaction efficiency of the device. Once extrapolated to macroscopic scales, these effects can result in a sizeable increase of the overall reaction efficiency., Comment: 12 pages, 12 figures

The structural features of the interface between the cystalline and amorphous phases of Si solid are studied in simulations based on a combination of empirical interatomic potentials and a nonorthogonal tight-binding model. The tight-binding Hamiltonian was created and tested for the types of structures and distortions anticipated to occur at this interface. The simulations indicate the presence of a number of interesting features near the interface. The features that may lead to crystallization upon heating include <110> chains with some defects, most prominently dimers similar to those on the Si(001) 2x1 reconstructed free surface. Within the amorphous region order is lost over very short distances. By examining six different samples with two interfaces each, we find the energy of the amorphous-crystal interface to be 0.49 +/- 0.05 J/m^2, Comment: Submitted to Phys. Rev. B