Search

The symmetric mass generation (SMG) phase is an insulator in which a single-particle gap is intrinsically opened by the interaction, without involving symmetry spontaneously breaking or topological order. Here, we perform unbiased quantum Monte-Carlo simulation and systematically investigate a bilayer fermionic model hosting Fermi surface SMG in the strongly interacting regime. With increasing interaction strength, the model undergoes a quantum phase transition from an exciton insulator to an SMG phase, belonging to the (2+1)-dimensional O(4) universality class. We access the spectral properties of the SMG phase, resembling a Mott insulating phase with relatively flat dispersion and pronounced spectral broadening. The dispersion of Green's function zeros is extracted from spectral function, featuring a surface at zero frequency precisely located at the original non-interacting Fermi surface, which constitutes a hallmark of the Fermi surface SMG phase. The bilayer model we study is potentially relevant to the newly discovered high-$T_c$ superconductor $\rm{La}_3 \rm{Ni}_2 \rm{O}_7$. Our results in SMG phase qualitatively capture the salient features of spectral function unveiled in recent ARPES experiments, shedding new insight on the underlying physics of $\rm{La}_3 \rm{Ni}_2 \rm{O}_7$., Comment: 4.5 pages plus supplemental material, 4 figures

Fluid dynamics computations for tube-like geometries are important for biomedical evaluation of vascular and airway fluid dynamics. Physics-Informed Neural Networks (PINNs) have recently emerged as a good alternative to traditional computational fluid dynamics (CFD) methods. The vanilla PINN, however, requires much longer training time than the traditional CFD methods for each specific flow scenario and thus does not justify its mainstream use. Here, we explore the use of the multi-case PINN approach for calculating biomedical tube flows, where varied geometry cases are parameterized and pre-trained on the PINN, such that results for unseen geometries can be obtained in real time. Our objective is to identify network architecture, tube-specific, and regularization strategies that can optimize this, via experiments on a series of idealized 2D stenotic tube flows., Comment: 24 pages, 8 figures, 5 tables

Here we develop a new scheme of projective quantum Monte-Carlo (QMC) simulation combining unbiased zero-temperature (projective) determinant QMC and variational Monte-Carlo based on Gutzwiller projection wave function, dubbed as ``Gutzwiller projection QMC''. The numerical results demonstrate that employment of Gutzwiller projection trial wave function with minimum energy strongly speed up the convergence of computational results, thus tremendously reducing computational time in the simulation. More remarkably, we present an example that sign problem is enormously alleviated in the Gutzwiller projection QMC, especially in the regime where sign problem is severe. Hence, we believe that Gutzwiller projection QMC paves a new route to improving the efficiency, and alleviating sign problem in QMC simulation on interacting fermionic systems., Comment: 5 pages, 4 figures

This paper reports a comprehensive study on the impacts of temperature-change, process variation, flicker noise and device aging on the inference accuracy of pre-trained all-ferroelectric (FE) FinFET deep neural networks. Multiple-level-cell (MLC) operation with a novel adaptive-program-and-read algorithm with 100ns write pulse has been experimentally demonstrated in 5 nm thick hafnium zirconium oxide (HZO)-based FE-FinFET. With pre-trained neural network (NN) with 97.5% inference accuracy on MNIST dataset as baseline, device to device variation is shown to have negligible impact. Flicker noise characterization at various bias conditions depicts that drain current fluctuation is less than 0.7% with virtually no inference accuracy degradation. The conductance drift of a programmed cell, as an aftermath of temperature change, was captured by a compact model over a wide range of gate biases. Despite significant inference accuracy degradation at 233K for a NN trained at 300K, gate bias optimization for recovering the accuracy is demonstrated. Endurance above 10$^8$ cycles and extrapolated retention above 10 years are shown, which paves the way for edge device artificial intelligence with FE-FinFETs.

The compressible Navier-Stokes-Poisson system takes the form of usual Navier-Stokes equations coupled with the self-consistent Poisson equation, which is used to simulate the transport of charged particles under the electric field of electrostatic potential force. In this paper, we focus on the large time behavior of global strong solutions in the $L^{p}$ Besov spaces of critical regularity. By exploring the dissipative effect arising from Poisson potential, we posed the new regularity assumption of low frequencies and then establish a \textit{sharp} time-weighted inequality, which leads to the optimal time-decay estimates of the solution. Indeed, we see that the decay of density is faster at the half rate than that of velocity, which is a different ingredient in comparison with the situation of usual Navier-Stokes equations. Our proof mainly depends on tricky and non classical Besov product estimates with respect to various Sobolev embeddings., Comment: 28pages. arXiv admin note: text overlap with arXiv:1710.00154

As a means of human-based computation, crowdsourcing has been widely used to annotate large-scale unlabeled datasets. One of the obvious challenges is how to aggregate these possibly noisy labels provided by a set of heterogeneous annotators. Another challenge stems from the difficulty in evaluating the annotator reliability without even knowing the ground truth, which can be used to build incentive mechanisms in crowdsourcing platforms. When each instance is associated with many possible labels simultaneously, the problem becomes even harder because of its combinatorial nature. In this paper, we present new flexible Bayesian models and efficient inference algorithms for multi-label annotation aggregation by taking both annotator reliability and label dependency into account. Extensive experiments on real-world datasets confirm that the proposed methods outperform other competitive alternatives, and the model can recover the type of the annotators with high accuracy., Comment: The paper needs more refinement