arXiv subjects

Hanwen Wang

Publications and source records attributed to Hanwen Wang.

32 records · Page 2Linked to original sources

arXiv2023

An Expert's Guide to Training Physics-informed Neural Networks

Physics-informed neural networks (PINNs) have been popularized as a deep learning framework that can seamlessly synthesize observational data and partial differential equation (PDE) constraints. Their practical effectiveness however can be hampered by training pathologies, but also oftentimes by poor choices made by users who lack deep learning expertise. In this paper we present a series of best practices that can significantly improve the training efficiency and overall accuracy of PINNs. We also put forth a series of challenging benchmark problems that highlight some of the most prominent difficulties in training PINNs, and present comprehensive and fully reproducible ablation studies that demonstrate how different architecture choices and training strategies affect the test accuracy of the resulting models. We show that the methods and guiding principles put forth in this study lead to state-of-the-art results and provide strong baselines that future studies should use for comparison purposes. To this end, we also release a highly optimized library in JAX that can be used to reproduce all results reported in this paper, enable future research studies, as well as facilitate easy adaptation to new use-case scenarios.

cs.LG↗

arXiv2023

Probing the fractional quantum Hall phases in valley-layer locked bilayer MoS$_{2}$

Semiconducting transition-metal dichalcogenides (TMDs) exhibit high mobility, strong spin-orbit coupling, and large effective masses, which simultaneously leads to a rich wealth of Landau quantizations and inherently strong electronic interactions. However, in spite of their extensively explored Landau levels (LL) structure, probing electron correlations in the fractionally filled LL regime has not been possible due to the difficulty of reaching the quantum limit. Here, we report evidence for fractional quantum Hall (FQH) states at filling fractions 4/5 and 2/5 in the lowest LL of bilayer MoS$_{2}$, manifested in fractionally quantized transverse conductance plateaus accompanied by longitudinal resistance minima. We further show that the observed FQH states sensitively depend on the dielectric and gate screening of the Coulomb interactions. Our findings establish a new FQH experimental platform which are a scarce resource: an intrinsic semiconducting high mobility electron gas, whose electronic interactions in the FQH regime are in principle tunable by Coulomb-screening engineering, and as such, could be the missing link between atomically thin graphene and semiconducting quantum wells.

cond-mat.mes-hall↗

arXiv2023

Probing the Anomalous Hall Transport and Magnetic Reversal of Chiral-Lattice Antiferromagnet Co$_{1/3}$NbS$_2$

Antiferromagnets exhibiting giant anomalous Hall effect (AHE) and anomalous Nernst effect (ANE) have recently aroused broad interest, not only for their potential applications in future electronic devices, but also because of the rich physics arising from the Berry curvature near the Fermi level. $\rm{Co_{1/3}NbS_2}$, by intercalating $\rm{Co^{2+}}$ ions between $\rm{NbS_2}$ layers, is a quasi-two-dimensional layered antiferromagnet with a chiral lattice. A large AHE has been observed in $\rm{Co_{1/3}NbS_2}$, but its origin is under debate. In this letter, we report the large AHE and ANE in exfoliated $\rm{Co_{1/3}NbS_2}$ flakes. By analyzing the thermoelectric data via the Mott relation, we determined that the observed large AHE and ANE primarily result from the intrinsic Berry curvature. We also observed the magnetic domains in $\rm{Co_{1/3}NbS_2}$ by reflective magnetic circular dichroism measurements. Combined with electrical transport measurements, we confirmed that the magnetic reversal in $\rm{Co_{1/3}NbS_2}$ is determined by domain wall motion, and the critical field ($H_c$) exhibits a memory effect of consecutive magnetic sweeps. Our work provides insight into the topological properties of $\rm{Co_{1/3}NbS_2}$ and paves the way to studying the spin configuration and magnetic domain dynamics in this fascinating antiferromagnet.

cond-mat.mes-hall↗

arXiv2022

Random Weight Factorization Improves the Training of Continuous Neural Representations

Continuous neural representations have recently emerged as a powerful and flexible alternative to classical discretized representations of signals. However, training them to capture fine details in multi-scale signals is difficult and computationally expensive. Here we propose random weight factorization as a simple drop-in replacement for parameterizing and initializing conventional linear layers in coordinate-based multi-layer perceptrons (MLPs) that significantly accelerates and improves their training. We show how this factorization alters the underlying loss landscape and effectively enables each neuron in the network to learn using its own self-adaptive learning rate. This not only helps with mitigating spectral bias, but also allows networks to quickly recover from poor initializations and reach better local minima. We demonstrate how random weight factorization can be leveraged to improve the training of neural representations on a variety of tasks, including image regression, shape representation, computed tomography, inverse rendering, solving partial differential equations, and learning operators between function spaces.

cs.LG↗

arXiv2022

A gate-programmable van der Waals metal-ferroelectric-semiconductor memory

Ferroelecticity, one of the keys to realize nonvolatile memories owing to the remanent electric polarization, has been an emerging phenomenon in the two-dimensional (2D) limit. Yet the demonstrations of van der Waals (vdW) memories using 2D ferroelectric materials as an ingredient are very limited. Especially, gate-tunable ferroelectric vdW memristive device, which holds promises in future neuromorphic applications, remains challenging. Here, we show a prototype gate-programmable memory by vertically assembling graphite, CuInP2S6, and MoS2 layers into a metal-ferroelectric-semiconductor architecture. The resulted devices exhibit two-terminal switchable electro-resistance with on-off ratios exceeding 105 and long-term retention, akin to a conventional memristor but strongly coupled to the ferroelectric characteristics of the CuInP2S6 layer. By controlling the top gate, Fermi level of MoS2 can be set inside (outside) of its band gap to quench (enable) the memristive behaviour, yielding a three-terminal gate programmable nonvolatile vdW memory. Our findings pave the way for the engineering of ferroelectric-mediated memories in future implementations of nanoelectronics.

cond-mat.mtrl-sci↗

arXiv2021

Realization of graphene logics in an exciton-enhanced insulating phase

For two decades, two-dimensional carbon species, including graphene, have been the core of research in pursuing next-generation logic applications beyond the silicon technology. Yet the opening of a gap in a controllable range of doping, whilst keeping high conductance outside of this gapped state, has remained a grand challenge in them thus far. Here we show that, by bringing Bernal-stacked bilayer graphene in contact with an anti-ferromagnetic insulator CrOCl, a strong insulating behavior is observed in a wide range of positive total electron doping $n_\mathrm{tot}$ and effective displacement field $D_\mathrm{eff}$ at low temperatures. Transport measurements further prove that such an insulating phase can be well described by the picture of an inter-layer excitonic state in bilayer graphene owing to electron-hole interactions. The consequential over 1 $\mathrm{G\Omega}$ excitonic insulator can be readily killed by tuning $D_\mathrm{eff}$ and/or $n_\mathrm{tot}$, and the system recovers to a high mobility graphene with a sheet resistance of less than 100 $\mathrm{\Omega}$. It thus yields transistors with "ON-OFF" ratios reaching 10$^{7}$, and a CMOS-like graphene logic inverter is demonstrated. Our findings of the robust insulating phase in bilayer graphene may be a leap forward to fertilize the future carbon computing.

cond-mat.mes-hall↗

arXiv2021

Improved architectures and training algorithms for deep operator networks

Operator learning techniques have recently emerged as a powerful tool for learning maps between infinite-dimensional Banach spaces. Trained under appropriate constraints, they can also be effective in learning the solution operator of partial differential equations (PDEs) in an entirely self-supervised manner. In this work we analyze the training dynamics of deep operator networks (DeepONets) through the lens of Neural Tangent Kernel (NTK) theory, and reveal a bias that favors the approximation of functions with larger magnitudes. To correct this bias we propose to adaptively re-weight the importance of each training example, and demonstrate how this procedure can effectively balance the magnitude of back-propagated gradients during training via gradient descent. We also propose a novel network architecture that is more resilient to vanishing gradient pathologies. Taken together, our developments provide new insights into the training of DeepONets and consistently improve their predictive accuracy by a factor of 10-50x, demonstrated in the challenging setting of learning PDE solution operators in the absence of paired input-output observations. All code and data accompanying this manuscript are publicly available at \url{https://github.com/PredictiveIntelligenceLab/ImprovedDeepONets.}

cs.LG↗

arXiv2021

Learning the solution operator of parametric partial differential equations with physics-informed DeepOnets

Deep operator networks (DeepONets) are receiving increased attention thanks to their demonstrated capability to approximate nonlinear operators between infinite-dimensional Banach spaces. However, despite their remarkable early promise, they typically require large training data-sets consisting of paired input-output observations which may be expensive to obtain, while their predictions may not be consistent with the underlying physical principles that generated the observed data. In this work, we propose a novel model class coined as physics-informed DeepONets, which introduces an effective regularization mechanism for biasing the outputs of DeepOnet models towards ensuring physical consistency. This is accomplished by leveraging automatic differentiation to impose the underlying physical laws via soft penalty constraints during model training. We demonstrate that this simple, yet remarkably effective extension can not only yield a significant improvement in the predictive accuracy of DeepOnets, but also greatly reduce the need for large training data-sets. To this end, a remarkable observation is that physics-informed DeepONets are capable of solving parametric partial differential equations (PDEs) without any paired input-output observations, except for a set of given initial or boundary conditions. We illustrate the effectiveness of the proposed framework through a series of comprehensive numerical studies across various types of PDEs. Strikingly, a trained physics informed DeepOnet model can predict the solution of $\mathcal{O}(10^3)$ time-dependent PDEs in a fraction of a second -- up to three orders of magnitude faster compared a conventional PDE solver. The data and code accompanying this manuscript are publicly available at \url{https://github.com/PredictiveIntelligenceLab/Physics-informed-DeepONets}.

cs.LG↗

arXiv2021

Strong bulk-surface interaction dominated in-plane anisotropy of electronic structure in GaTe

Recently, intriguing physical properties have been unraveled in anisotropic layered semiconductors, in which the in-plane electronic band structure anisotropy often originates from the low crystallographic symmetry and thus a thickness-independent character emerges. Here, we apply high-resolution angle-resolved photoemission spectroscopy to directly image the in-plane anisotropic energy bands in monoclinic gallium telluride (GaTe). Our first-principles calculations reveal the in-plane anisotropic energy band structure of GaTe measured experimentally is dominated by a strong bulk-surface interaction rather than geometric factors, surface effect and quantum confinement effect. Furthermore, accompanied by the thickness of GaTe increasing from mono- to few-layers, the strong interlayer coupling of GaTe induces direct-indirect-direct band gap transitions and the in-plane anisotropy of hole effective mass is reversed. Our results shed light on the physical origins of in-plane anisotropy of electronic structure in GaTe, paving the way for the design and device applications of nanoelectronics and optoelectronics based on anisotropic layered semiconductors.

cond-mat.mtrl-sci↗

arXiv2020

On the eigenvector bias of Fourier feature networks: From regression to solving multi-scale PDEs with physics-informed neural networks

Physics-informed neural networks (PINNs) are demonstrating remarkable promise in integrating physical models with gappy and noisy observational data, but they still struggle in cases where the target functions to be approximated exhibit high-frequency or multi-scale features. In this work we investigate this limitation through the lens of Neural Tangent Kernel (NTK) theory and elucidate how PINNs are biased towards learning functions along the dominant eigen-directions of their limiting NTK. Using this observation, we construct novel architectures that employ spatio-temporal and multi-scale random Fourier features, and justify how such coordinate embedding layers can lead to robust and accurate PINN models. Numerical examples are presented for several challenging cases where conventional PINN models fail, including wave propagation and reaction-diffusion dynamics, illustrating how the proposed methods can be used to effectively tackle both forward and inverse problems involving partial differential equations with multi-scale behavior. All code an data accompanying this manuscript will be made publicly available at \url{https://github.com/PredictiveIntelligenceLab/MultiscalePINNs}.

cs.LG↗

arXiv2020

Ultra-efficient magnetism modulation in a Weyl ferromagnet by current-assisted domain wall motion

Flexible and efficient manipulation of magnetic configurations can be challenging. In the design of practical devices, achieving a high effective magnetic field with a low working current is under tight demand. Here, we report a unique method for efficient magnetism modulation by direct current injection in magnetic Weyl semimetal Co3Sn2S2. We demonstrate that the modulation process stems from current-assisted domain wall motion. Through two independent methods, we reveal that the spin-transfer torque efficiency of Co3Sn2S2 reaches as high as 2.4-5.6 kOe MA^(-1) cm^2, and the threshold current density for driving the magnetic domain walls is as low as <5.1*10^5 A/cm^2 without an external field, and <1.5*10^5 A/cm^2 with a moderate external field. Our findings manifest a new and powerful approach for sub-micron magnetism manipulation, and also open the door towards a new paradigm of spintronics that combines magnetism, topology, and metallicity for low-energy consumption memory and computing.

cond-mat.mtrl-sci↗

arXiv2019

Room temperature 2D ferromagnetism in few-layered 1$T$-CrTe$_{2}$

Spin-related electronics using two dimensional (2D) van der Waals (vdW) materials as a platform are believed to hold great promise for revolutionizing the next generation spintronics. Although many emerging new phenomena have been unravelled in 2D electronic systems with spin long-range orderings, the scarcely reported room temperature magnetic vdW material has thus far hindered the related applications. Here, we show that intrinsic ferromagnetically aligned spin polarization can hold up to 316 K in a metallic phase of 1$T$-CrTe$_{2}$ in the few-layer limit. This room temperature 2D long range spin interaction may be beneficial from an itinerant enhancement. Spin transport measurements indicate an in-plane room temperature negative anisotropic magnetoresistance (AMR) in few-layered CrTe$_{2}$, but a sign change in the AMR at lower temperature, with -0.6$\%$ at 300 K and +5$\%$ at 10 K, respectively. This behavior may originate from the specific spin polarized band structure of CrTe$_{2}$. Our findings provide insights into magnetism in few-layered CrTe$_{2}$, suggesting potential for future room temperature spintronic applications of such 2D vdW magnets.

cond-mat.mes-hall↗

arXiv2019

Field-free switching of a Spin-orbit torque device aided by interlayer-coupling induced domain walls

The spin-orbit torque device is promising as a candidate for next generation magnetic memory, while the static in-plane field needed to induce deterministic switching is a main obstacle for its application in highly integrated circuits. Instead of introducing effective field into the device, in this work we present an alternative way to achieve the field-free current-driven magnetization switching. By adding Tb/Co multilayers at two ends of the current channel, assisting domain wall is created by interlayer exchange coupling. The field-free deterministic switching is achieved by the movement of the domain wall driven by current. By loop shift measurement we find the driven force exerted on the domain wall is determined by the direction of the in-plane moment in the domain wall. Finally, we modify the device into a synthetic antiferromagnetic structure to solve the problem of reading out signal. The present work broadens the choice of field-free spin-orbit torque device design and clearly depicts the difference between two different switching mechanism.

physics.app-ph↗

arXiv2019

Gate tunable giant anisotropic resistance in ultra-thin GaTe

In crystals, the duplication of atoms often follows different periodicity along different directions. It thus gives rise to the so called anisotropy, which is usually even more pronounced in two dimensional (2D) materials due to the absence of $\textbf{z}$ dimension. Indeed, in the emerging 2D materials, electrical anisotropy has been one of the focuses in recent experimental efforts. However, key understandings of the in-plane anisotropic resistance in low-symmetry 2D materials, as well as demonstrations of model devices taking advantage of it, have proven difficult. Here, we show that, in few-layered semiconducting GaTe, electrical conductivity along $\textbf{x}$ and $\textbf{y}$ directions of the 2D crystal can be gate tuned from a ratio of less than one order to as large as 10$^{3}$. This effect is further demonstrated to yield an anisotropic memory resistor behaviour in ultra-thin GaTe, when equipped with an architecture of van der Waals floating gate. Our findings of gate tunable giant anisotropic resistance (GAR) effect pave the way for potential applications in nano-electronics such as multifunctional directional memories in the 2D limit.

cond-mat.mes-hall↗