arXiv subjects

Wei Li

Publications and source records attributed to Wei Li.

At least 955 records · Page 53Linked to original sources

arXiv2021

Realization of Topological Mott Insulator in a Twisted Bilayer Graphene Lattice Model

Magic-angle twisted bilayer graphene has recently become a thriving material platform realizing correlated electron phenomena taking place within its topological flat bands. Several numerical and analytical methods have been applied to understand the correlated phases therein, revealing some similarity with the quantum Hall physics. In this work, we provide a Mott-Hubbard perspective for the TBG system. Employing the large-scale density matrix renormalization group on the lattice model containing the projected Coulomb interactions only, we identify a first-order quantum phase transition between the insulating stripe phase and the quantum anomalous Hall state with the Chern number of $\pm 1$. Our results not only shed light on the mechanism of the quantum anomalous Hall state discovered at three-quarters filling, but also provide an example of the topological Mott insulator, i.e., the quantum anomalous Hall state in the strong coupling limit.

cond-mat.str-el↗

arXiv2021

Privacy amplification scheme based on composite coding

Privacy amplification is an indispensable step in the post-processing of quantum key distribution, which can be used to compress the redundancy of shared key and improve the security level of the key. The commonly used privacy amplification is based on the random selection of universal hash functions, which needs the help of an additional random source, while it does not exist in general. In this paper, we propose a privacy amplification scheme based on composite coding, which is an extension of quantum CSS codes to classical linear codes. Compared with the universal hashing function, the proposed scheme does not need other random sources, and the randomness can be completely provided by the qubit string. Furthermore, the information-theoretic bound for the extraction of the key is obvious in composite coding.

quant-ph↗

arXiv2021

Tied & Reduced RNN-T Decoder

Previous works on the Recurrent Neural Network-Transducer (RNN-T) models have shown that, under some conditions, it is possible to simplify its prediction network with little or no loss in recognition accuracy (arXiv:2003.07705 [eess.AS], [2], arXiv:2012.06749 [cs.CL]). This is done by limiting the context size of previous labels and/or using a simpler architecture for its layers instead of LSTMs. The benefits of such changes include reduction in model size, faster inference and power savings, which are all useful for on-device applications. In this work, we study ways to make the RNN-T decoder (prediction network + joint network) smaller and faster without degradation in recognition performance. Our prediction network performs a simple weighted averaging of the input embeddings, and shares its embedding matrix weights with the joint network's output layer (a.k.a. weight tying, commonly used in language modeling arXiv:1611.01462 [cs.LG]). This simple design, when used in conjunction with additional Edit-based Minimum Bayes Risk (EMBR) training, reduces the RNN-T Decoder from 23M parameters to just 2M, without affecting word-error rate (WER).

cs.CL↗

arXiv2021

Cylindrical and Asymmetrical 3D Convolution Networks for LiDAR-based Perception

State-of-the-art methods for driving-scene LiDAR-based perception (including point cloud semantic segmentation, panoptic segmentation and 3D detection, \etc) often project the point clouds to 2D space and then process them via 2D convolution. Although this cooperation shows the competitiveness in the point cloud, it inevitably alters and abandons the 3D topology and geometric relations. A natural remedy is to utilize the 3D voxelization and 3D convolution network. However, we found that in the outdoor point cloud, the improvement obtained in this way is quite limited. An important reason is the property of the outdoor point cloud, namely sparsity and varying density. Motivated by this investigation, we propose a new framework for the outdoor LiDAR segmentation, where cylindrical partition and asymmetrical 3D convolution networks are designed to explore the 3D geometric pattern while maintaining these inherent properties. The proposed model acts as a backbone and the learned features from this model can be used for downstream tasks such as point cloud semantic and panoptic segmentation or 3D detection. In this paper, we benchmark our model on these three tasks. For semantic segmentation, we evaluate the proposed model on several large-scale datasets, \ie, SemanticKITTI, nuScenes and A2D2. Our method achieves the state-of-the-art on the leaderboard of SemanticKITTI (both single-scan and multi-scan challenge), and significantly outperforms existing methods on nuScenes and A2D2 dataset. Furthermore, the proposed 3D framework also shows strong performance and good generalization on LiDAR panoptic segmentation and LiDAR 3D detection.

cs.CV↗

arXiv2021

Lorentz Resonance in the Homogenization of Plasmonic Crystals

We explain the Lorentz resonances in plasmonic crystals that consist of 2D nano dielectric inclusions as the interaction between resonant material properties and geometric resonances of electrostatic nature. One example of such plasmonic crystals are graphene nanosheets that are periodically arranged within a non-magnetic bulk dielectric. We identify local geometric resonances on the length scale of the small scale period. From a materials perspective, the graphene surface exhibits a dispersive surface conductance captured by the Drude model. Together these phenomena conspire to generate Lorentz resonances at frequencies controlled by the surface geometry and the surface conductance. The Lorentz resonances found in the frequency response of the effective dielectric tensor of the bulk metamaterial is shown to be given by an explicit formula, in which material properties and geometric resonances are decoupled. This formula is rigorous and obtained directly from corrector fields describing local electrostatic fields inside the heterogeneous structure. Our analytical findings can serve as an efficient computational tool to describe the general frequency dependence of periodic optical devices. As a concrete example, we investigate two prototypical geometries composed of nanotubes and nanoribbons.

physics.optics↗

arXiv2021

Do Transformer Modifications Transfer Across Implementations and Applications?

The research community has proposed copious modifications to the Transformer architecture since it was introduced over three years ago, relatively few of which have seen widespread adoption. In this paper, we comprehensively evaluate many of these modifications in a shared experimental setting that covers most of the common uses of the Transformer in natural language processing. Surprisingly, we find that most modifications do not meaningfully improve performance. Furthermore, most of the Transformer variants we found beneficial were either developed in the same codebase that we used or are relatively minor changes. We conjecture that performance improvements may strongly depend on implementation details and correspondingly make some recommendations for improving the generality of experimental results.

cs.LG↗

arXiv2021

EPick: Multi-Class Attention-based U-shaped Neural Network for Earthquake Detection and Seismic Phase Picking

Earthquake detection and seismic phase picking not only play a crucial role in travel time estimation of body waves(P and S waves) but also in the localisation of the epicenter of the corresponding event. Generally, manual phase picking is a trustworthy and the optimum method to determine the phase arrival time, however, its capacity is restricted by available resources and time. Moreover, noisy seismic data renders an additional critical challenge for fast and accurate phase picking. In this study, a deep learning based model, EPick, is proposed which benefits both from U shaped neural network (also called UNet)and attention mechanism, as a strong alternative for seismic event detection and phase picking. On one hand, the utilization of UNet structure enables addressing different levels of deep features. On the other hand, attention mechanism promotes the decoder in the UNet structure to focus on the efficient exploitation of the low-resolution features learned from the encoder part to achieve precise phase picking. Extensive experimental results demonstrate that EPick achieves better performance over the benchmark method, and show the models robustness when tested on a different seismic dataset.

physics.geo-ph↗

arXiv2021

Shifted Quiver Yangians and Representations from BPS Crystals

We introduce a class of new algebras, the shifted quiver Yangians, as the BPS algebras for type IIA string theory on general toric Calabi-Yau three-folds. We construct representations of the shifted quiver Yangian from general subcrystals of the canonical crystal. We derive our results via equivariant localization for supersymmetric quiver quantum mechanics for various framed quivers, where the framings are determined by the shape of the subcrystals. Our results unify many known BPS state counting problems, including open BPS counting, non-compact D4-branes, and wall crossing phenomena, simply as different representations of the shifted quiver Yangians. Furthermore, most of our representations seem to be new, and this suggests the existence of a zoo of BPS state counting problems yet to be studied in detail.

hep-th↗

arXiv2021

Musical Tempo Estimation Using a Multi-scale Network

Recently, some single-step systems without onset detection have shown their effectiveness in automatic musical tempo estimation. Following the success of these systems, in this paper we propose a Multi-scale Grouped Attention Network to further explore the potential of such methods. A multi-scale structure is introduced as the overall network architecture where information from different scales is aggregated to strengthen contextual feature learning. Furthermore, we propose a Grouped Attention Module as the key component of the network. The proposed module separates the input feature into several groups along the frequency axis, which makes it capable of capturing long-range dependencies from different frequency positions on the spectrogram. In comparison experiments, the results on public datasets show that the proposed model outperforms existing state-of-the-art methods on Accuracy1.

eess.AS↗

arXiv2021

Algorithms yield upper bounds in differential algebra

Consider an algorithm computing in a differential field with several commuting derivations such that the only operations it performs with the elements of the field are arithmetic operations, differentiation, and zero testing. We show that, if the algorithm is guaranteed to terminate on every input, then there is a computable upper bound for the size of the output of the algorithm in terms of the size of the input. We also generalize this to algorithms working with models of good enough theories (including for example, difference fields). We then apply this to differential algebraic geometry to show that there exists a computable uniform upper bound for the number of components of any variety defined by a system of polynomial PDEs. We then use this bound to show the existence of a computable uniform upper bound for the elimination problem in systems of polynomial PDEs with delays.

math.AC↗

arXiv2021

Long-term, Short-term and Sudden Event: Trading Volume Movement Prediction with Graph-based Multi-view Modeling

Trading volume movement prediction is the key in a variety of financial applications. Despite its importance, there is few research on this topic because of its requirement for comprehensive understanding of information from different sources. For instance, the relation between multiple stocks, recent transaction data and suddenly released events are all essential for understanding trading market. However, most of the previous methods only take the fluctuation information of the past few weeks into consideration, thus yielding poor performance. To handle this issue, we propose a graphbased approach that can incorporate multi-view information, i.e., long-term stock trend, short-term fluctuation and sudden events information jointly into a temporal heterogeneous graph. Besides, our method is equipped with deep canonical analysis to highlight the correlations between different perspectives of fluctuation for better prediction. Experiment results show that our method outperforms strong baselines by a large margin.

q-fin.ST↗

arXiv2021

A Muon-Ion Collider at BNL: the future QCD frontier and path to a new energy frontier of $μ^+μ^-$ colliders

We propose the development and construction of a novel muon-ion collider (MuIC) at Brookhaven National Laboratory (BNL) in the USA as an upgrade to succeed the electron-ion collider (EIC) that is scheduled to commence in the early 2030s, by a joint effort of the nuclear and particle physics communities. The BNL facility could accommodate a muon storage beam with an energy up to about 1~TeV with existing magnet technology. When collided with a 275~GeV hadron beam, the MuIC center-of-mass energy of about 1~TeV will extend the kinematic coverage of deep inelastic scattering physics at the EIC (with polarized beams) by more than an order of magnitude in $Q^2$ and $x$, opening a new QCD frontier to address many fundamental scientific questions in nuclear and particle physics. This coverage is comparable to that of the proposed Large Hadron-Electron Collider (LHeC) at CERN, but with complementary lepton and hadron kinematics and beam polarization. Additionally, the developmentof a MuIC at BNL will focus the worldwide R\&D efforts on muon collider technology and serve as a demonstrator toward a future muon-antimuon collider at {$\mathcal O$}(10)~TeV energies, which is an attractive option to reach the next high energy frontier in particle physics at an affordable cost and a smaller footprint than a future circular hadron collider. We discuss here the possible design parameters of the MuIC, kinematic coverage, science cases, and detector design considerations including estimates of resolutions on DIS kinematic variables. A possible road map toward the future MuIC and muon-antimuon colliders is also presented.

physics.acc-ph↗

arXiv2021

Investigation of Experimental Observables in Search of the Chiral Magnetic Effect in Heavy-ion Collisions in the STAR experiment

The chiral magnetic effect (CME) is a novel transport phenomenon, arising from the interplay between quantum anomalies and strong magnetic fields in chiral systems. In high-energy nuclear collisions, the CME may survive the expansion of the quark-gluon plasma fireball and be detected in experiments. Over the past decade, the experimental searches for the CME have aroused extensive interest at the Relativistic Heavy Ion Collider (RHIC) and the Large Hadron Collider (LHC). The main goal of this article is to investigate three pertinent experimental approaches: the $γ$ correlator, the $R$ correlator and the signed balance functions. We will exploit both simple Monte Carlo simulations and a realistic event generator (EBE-AVFD) to verify the equivalence in the kernel-component observables among these methods and to ascertain their sensitivities to the CME signal for the isobaric collisions at RHIC.

nucl-ex↗

arXiv2021

The third- and fourth-order orbital angular momentum multiplexed amplification with ultra-low differential mode gain

In this Letter, a ring-core erbium-doped fiber (RC-EDF), with two-layer erbium-doped structure, supporting up to the fourth-order orbital angular momentum (OAM) mode is designed and fabricated for OAM mode multiplexed amplification. Using the RC-EDF, the third- and fourth-order OAM modes amplification with ultra-low differential mode gain (DMG) is demonstrated by observing both the modal intensity and phase distribution and measuring the modal gain under the fundamental mode core-pumping. The measured average gain of four modes (l=+3, -3, +4, -4) multiplexed amplification is higher than 19dB cover the C-band and the DMG is less than 1dB. Additionally, the gain of two conjugate OAM modes are almost the same under different pump power no matter they are amplified simultaneously or separately.

physics.optics↗

arXiv2021

Multi defect detection and analysis of electron microscopy images with deep learning

Electron microscopy is widely used to explore defects in crystal structures, but human detecting of defects is often time-consuming, error-prone, and unreliable, and is not scalable to large numbers of images or real-time analysis. In this work, we discuss the application of machine learning approaches to find the location and geometry of different defect clusters in irradiated steels. We show that a deep learning based Faster R-CNN analysis system has a performance comparable to human analysis with relatively small training data sets. This study proves the promising ability to apply deep learning to assist the development of automated microscopy data analysis even when multiple features are present and paves the way for fast, scalable, and reliable analysis systems for massive amounts of modern electron microscopy data.

cs.CV↗

arXiv2021

Semantic Concentration for Domain Adaptation

Domain adaptation (DA) paves the way for label annotation and dataset bias issues by the knowledge transfer from a label-rich source domain to a related but unlabeled target domain. A mainstream of DA methods is to align the feature distributions of the two domains. However, the majority of them focus on the entire image features where irrelevant semantic information, e.g., the messy background, is inevitably embedded. Enforcing feature alignments in such case will negatively influence the correct matching of objects and consequently lead to the semantically negative transfer due to the confusion of irrelevant semantics. To tackle this issue, we propose Semantic Concentration for Domain Adaptation (SCDA), which encourages the model to concentrate on the most principal features via the pair-wise adversarial alignment of prediction distributions. Specifically, we train the classifier to class-wisely maximize the prediction distribution divergence of each sample pair, which enables the model to find the region with large differences among the same class of samples. Meanwhile, the feature extractor attempts to minimize that discrepancy, which suppresses the features of dissimilar regions among the same class of samples and accentuates the features of principal parts. As a general method, SCDA can be easily integrated into various DA methods as a regularizer to further boost their performance. Extensive experiments on the cross-domain benchmarks show the efficacy of SCDA.

cs.CV↗

arXiv2021

ASFM-Net: Asymmetrical Siamese Feature Matching Network for Point Completion

We tackle the problem of object completion from point clouds and propose a novel point cloud completion network employing an Asymmetrical Siamese Feature Matching strategy, termed as ASFM-Net. Specifically, the Siamese auto-encoder neural network is adopted to map the partial and complete input point cloud into a shared latent space, which can capture detailed shape prior. Then we design an iterative refinement unit to generate complete shapes with fine-grained details by integrating prior information. Experiments are conducted on the PCN dataset and the Completion3D benchmark, demonstrating the state-of-the-art performance of the proposed ASFM-Net. Our method achieves the 1st place in the leaderboard of Completion3D and outperforms existing methods with a large margin, about 12%. The codes and trained models are released publicly at https://github.com/Yan-Xia/ASFM-Net.

cs.CV↗

arXiv2021

Infinitely many embedded eigenvalues for the Neumann-Poincaré operator in 3D

This article constructs a surface whose Neumann-Poincaré (NP) integral operator has infinitely many eigenvalues embedded in its essential spectrum. The surface is a sphere perturbed by smoothly attaching a conical singularity, which imparts essential spectrum. Rotational symmetry allows a decomposition of the operator into Fourier components. Eigenvalues of infinitely many Fourier components are constructed so that they lie within the essential spectrum of other Fourier components and thus within the essential spectrum of the full NP operator. The proof requires the perturbation to be sufficiently small, with controlled curvature, and the conical singularity to be sufficiently flat.

math.FA↗