Searcharxiv⌕ Search

arXiv subjects

Ke Wang

Publications and source records attributed to Ke Wang.

At least 595 records · Page 33Linked to original sources

Universal finite size scaling around tricriticality between topologically ordered, SPT, and trivial phases

A quantum tricritical point is shown to exists in coupled time-reversal symmetry (TRS) broken Majorana chains. The tricriticality separates topologically ordered, symmetry protected topological (SPT), and trivial phases of the system. Here we demonstrate that the breaking of the TRS manifests itself in an emergence of a new dimensionless scale, $g = α(ξ) B \sqrt{N}$, where $N$ is the system size, $B$ is a generic TRS breaking field, and $α(ξ)$, with $α(0)\equiv 1$, is a model-dependent function of the localization length, $ξ$, of boundary Majorana zero modes at the tricriticality. This scale determines the scaling of the finite size corrections around the tricriticality, which are shown to be {\it universal}, and independent of the nature of the breaking of the TRS. We show that the single variable scaling function, $f(w)$, $w\propto m N$, where $m$ is the excitation gap, that defines finite-size corrections to the ground state energy of the system around topological phase transition at $B=0$, becomes double-scaling, $f=f(w,g)$, at finite $B$. We realize TRS breaking through three different methods with completely different lattice details and find the same universal behavior of $f(w,g)$. In the critical regime, $m=0$, the function $f(0,g)$ is nonmonotonic, and reproduces the Ising conformal field theory scaling only in limits $g=0$ and $g\rightarrow \infty$. The obtained result sets a scale of $N \gg 1/(αB)^2$ for the system to reach the thermodynamic limit in the presence of the TRS breaking. We derive the effective low-energy theory describing the tricriticality and analytically find the asymptotic behavior of the finite-size scaling function. Our results show that the boundary entropy around the tricriticality is also a universal function of $g$ at $m=0$.

cond-mat.stat-mech↗

Pressure-induced superconductivity in topological type II Dirac semimetal NiTe2

Very recently, NiTe2 has been reported to be a type II Dirac semimetal with Dirac nodes near the Fermi surface. Furthermore, it is unveiled that NiTe2 presents the Hall Effect, which is ascribed to orbital magnetoresistance. The physical properties behavior of NiTe2 under high pressure attracts us. In this paper, we investigate the electrical properties of polycrystalline NiTe2 by application of pressure ranging from 3.4GPa to 54.45Gpa. Superconductivity emerges at critical pressure 12GPa with a transition temperature of 3.7K, and Tc reaches its maximum, 6.4 K, at the pressure of 52.8GPa. Comparing with the superconductivity in MoP, we purposed the possibility of topological superconductivity in NiTe2. Two superconductivity transitions are observed with pressure increasing in single crystal.

cond-mat.supr-con↗

Controllable Unsupervised Text Attribute Transfer via Editing Entangled Latent Representation

Unsupervised text attribute transfer automatically transforms a text to alter a specific attribute (e.g. sentiment) without using any parallel data, while simultaneously preserving its attribute-independent content. The dominant approaches are trying to model the content-independent attribute separately, e.g., learning different attributes' representations or using multiple attribute-specific decoders. However, it may lead to inflexibility from the perspective of controlling the degree of transfer or transferring over multiple aspects at the same time. To address the above problems, we propose a more flexible unsupervised text attribute transfer framework which replaces the process of modeling attribute with minimal editing of latent representations based on an attribute classifier. Specifically, we first propose a Transformer-based autoencoder to learn an entangled latent representation for a discrete text, then we transform the attribute transfer task to an optimization problem and propose the Fast-Gradient-Iterative-Modification algorithm to edit the latent representation until conforming to the target attribute. Extensive experimental results demonstrate that our model achieves very competitive performance on three public data sets. Furthermore, we also show that our model can not only control the degree of transfer freely but also allow to transfer over multiple aspects at the same time.

cs.CL↗

IStego100K: Large-scale Image Steganalysis Dataset

In order to promote the rapid development of image steganalysis technology, in this paper, we construct and release a multivariable large-scale image steganalysis dataset called IStego100K. It contains 208,104 images with the same size of 1024*1024. Among them, 200,000 images (100,000 cover-stego image pairs) are divided as the training set and the remaining 8,104 as testing set. In addition, we hope that IStego100K can help researchers further explore the development of universal image steganalysis algorithms, so we try to reduce limits on the images in IStego100K. For each image in IStego100K, the quality factors is randomly set in the range of 75-95, the steganographic algorithm is randomly selected from three well-known steganographic algorithms, which are J-uniward, nsF5 and UERD, and the embedding rate is also randomly set to be a value of 0.1-0.4. In addition, considering the possible mismatch between training samples and test samples in real environment, we add a test set (DS-Test) whose source of samples are different from the training set. We hope that this test set can help to evaluate the robustness of steganalysis algorithms. We tested the performance of some latest steganalysis algorithms on IStego100K, with specific results and analysis details in the experimental part. We hope that the IStego100K dataset will further promote the development of universal image steganalysis technology. The description of IStego100K and instructions for use can be found at https://github.com/YangzlTHU/IStego100K

cs.CR↗

Ferromagnetic van der Waals compound MnSb$_{1.8}$Bi$_{0.2}$Te$_4$

The intersection of topology and magnetism represents a new playground to discover novel quantum phenomena and device concepts. In this work, we show that a van der Waals compound MnSb$_{1.8}$Bi$_{0.2}$Te$_4$ exhibits a ferromagnetic ground state with a Curie temperature of 26 K, in contrast to the antiferromagnetic order previously found for other members of the Mn(Sb, Bi)$_2$Te$_4$ family. We employ magneto-transport, bulk magnetization and neutron scattering studies to illustrate the magnetic and electrical properties of MnSb$_{1.8}$Bi$_{0.2}$Te$_4$ and report on the observation of an unusual anomalous Hall effect. Our results are an important step in the synthesis and understanding of ferromagnetic topological insulators.

cond-mat.mes-hall↗

FlowNorm: A Learning-based Method for Increasing Convergence Range of Direct Alignment

Many approaches have been proposed to estimate camera poses by directly minimizing photometric error. However, due to the non-convex property of direct alignment, proper initialization is still required for these methods. Many robust norms (e.g. Huber norm) have been proposed to deal with the outlier terms caused by incorrect initializations. These robust norms are solely defined on the magnitude of each error term. In this paper, we propose a novel robust norm, named FlowNorm, that exploits the information from both the local error term and the global image registration information. While the local information is defined on patch alignments, the global information is estimated using a learning-based network. Using both the local and global information, we achieve an unprecedented convergence range in which images can be aligned given large view angle changes or small overlaps. We further demonstrate the usability of the proposed robust norm by integrating it into the direct methods DSO and BA-Net, and generate more robust and accurate results in real-time.

cs.RO↗

AI Benchmark: All About Deep Learning on Smartphones in 2019

The performance of mobile AI accelerators has been evolving rapidly in the past two years, nearly doubling with each new generation of SoCs. The current 4th generation of mobile NPUs is already approaching the results of CUDA-compatible Nvidia graphics cards presented not long ago, which together with the increased capabilities of mobile deep learning frameworks makes it possible to run complex and deep AI models on mobile devices. In this paper, we evaluate the performance and compare the results of all chipsets from Qualcomm, HiSilicon, Samsung, MediaTek and Unisoc that are providing hardware acceleration for AI inference. We also discuss the recent changes in the Android ML pipeline and provide an overview of the deployment of deep learning models on mobile devices. All numerical results provided in this paper can be found and are regularly updated on the official project website: http://ai-benchmark.com.

cs.PF↗

Testing Neural Program Analyzers

Deep neural networks have been increasingly used in software engineering and program analysis tasks. They usually take a program and make some predictions about it, e.g., bug prediction. We call these models neural program analyzers. The reliability of neural programs can impact the reliability of the encompassing analyses. In this paper, we describe our ongoing efforts to develop effective techniques for testing neural programs. We discuss the challenges involved in developing such tools and our future plans. In our preliminary experiment on a neural model recently proposed in the literature, we found that the model is very brittle, and simple perturbations in the input can cause the model to make mistakes in its prediction.

cs.LG↗

Multi-scale analysis of the Monoceros OB 1 star-forming region: II. Colliding filaments in the Monoceros OB1 molecular cloud

We started a multi-scale analysis of G202.3+2.5, an intertwined filamentary region of Monoceros OB1. In Paper I, we examined the distributions of dense cores and protostars and found enhanced star formation (SF) activity in the junction region of the filaments. In this second paper, we aim to unveil the connections between the core and filament evolutions, and between the filament dynamics and the global evolution of the cloud. We characterise the gas dynamics and energy balance using Herschel and WISE observations and molecular tracers observed with the IRAM 30m and TRAO 14m telescopes. The velocity field of the cloud is examined and velocity-coherent structures are put in perspective with the cloud environment. Two main velocity components (VCs) are revealed, well separated in the north and merged around the location of intense N2H+ emission where Paper I found the peak of SF activity. The relative position of the two VCs along the sightline, and the velocity gradient in N2H+ emission imply that the VCs have been undergoing collision for ~10^5 yrs. The dense gas where N2H+ is detected is interpreted as the compressed region between the two filaments, which corresponds to a high mass inflow rate of ~1e-3 Msun/yr and possibly leads to an increase in its SF efficiency. We identify a protostar in the junction region that possibly powers two crossed intermittent outflows. We show that the HII region around the nearby cluster NCG 2264 is still expanding and its role in the collision is examined. However, we cannot rule out the idea that the collision arises mostly from the global collapse of the cloud. The (sub-)filament-scale observables examined in this paper reveal a collision between G202.3+2.5 sub-structures and its probable role in feeding the cores in the junction region. One must now characterise the cloud morphology, its fragmentation, and magnetic field, all at high resolution.

astro-ph.GA↗

Multi-scale analysis of the Monoceros OB 1 star-forming region: I. The dense core population

Current theories and models attempt to explain star formation globally, from core scales to giant molecular cloud scales. A multi-scale observational characterisation of an entire molecular complex is necessary to constrain them. We investigate star formation in G202.3+2.5, a ~10x3 pc sub-region of the Monoceros OB1 cloud with a complex morphology harbouring interconnected filamentary structures. We aim to connect the evolution of cores and filaments in G202.3+2.5 with the global evolution of the cloud and to identify the engines of the cloud dynamics. In this first paper, the star formation activity is evaluated by surveying the distributions of dense cores and protostars, and their evolutionary state, as characterised using both infrared observations from the Herschel and WISE telescopes and molecular line observations with the IRAM 30-m telescope. We find ongoing star formation in the whole cloud, with a local peak in star formation activity around the centre of G202.3+2.5 where a chain of massive cores (10-50 Msun) forms a massive ridge (>150 Msun). All evolutionary stages from starless cores to Class II protostars are found in G202.3+2.5, including a possibly starless, large column density (8x10^{22} cm-2), and massive (52 Msun) core. All the core-scale observables examined in this paper point to an enhanced star formation activity centred on the junction between the three main branches of the ramified structure of G202.3+2.5. This suggests that the increased star-formation activity results from the convergence of these branches. To further investigate the origin of this enhancement, it is now necessary to extend the analysis to larger scales, in order to examine the relationship between cores, filaments and their environment. We address these points through the analysis of the dynamics of G202.3+2.5 in a joint paper.

astro-ph.GA↗

A Census of Early Phase High-Mass Star Formation in the Central Molecular Zone

We present new observations of C-band continuum emission and masers to assess high-mass ($>$8 $M_\odot$) star formation at early evolutionary phases in the inner 200 pc of the Central Molecular Zone (CMZ) of the Galaxy. The continuum observation is complete to free-free emission from stars above 10-11 $M_\odot$ in 91% of the covered area. We identify 104 compact sources in the continuum emission, among which five are confirmed ultracompact H II regions, 12 are candidates of ultra-compact H II regions, and the remaining 87 sources are mostly massive stars in clusters, field stars, evolved stars, pulsars, extragalactic sources, or of unknown nature that is to be investigated. We detect class II CH$_3$OH masers at 23 positions, among which six are new detections. We confirm six known H$_2$CO masers in two high-mass star forming regions, and detect two new H$_2$CO masers toward the Sgr C cloud, making it the ninth region in the Galaxy that contains masers of this type. In spite of these detections, we find that current high-mass star formation in the inner CMZ is only taking place in seven isolated clouds. The results suggest that star formation at early evolutionary phases in the CMZ is about 10 times less efficient than expected by the dense gas star formation relation, which is in line with previous studies that focus on more evolved phases of star formation. This means that if there will be any impending, next burst of star formation in the CMZ, it has not yet begun.

astro-ph.GA↗

Influence of Wolf-Rayet stars on surrounding star-forming molecular clouds

We investigate the influence of Wolf-Rayet (W-R) stars on their surrounding star-forming molecular clouds. We study five regions containing W-R stars in the inner Galactic plane ($l\sim$[14$^\circ$-52$^\circ$]), using multi-wavelength data from near-infrared to radio wavelengths. Analysis of $^{13}$CO line data reveals that these W-R stars have developed gas-deficient cavities in addition to molecular shells with expansion velocities of a few km s$^{-1}$. The pressure owing to stellar winds primarily drives these expanding shells and sweeps up the surrounding matter to distances of a few pc. The column densities of shells are enhanced by a minimum of 14% for one region to a maximum of 88% for another region with respect to the column densities within their central cavities. No active star formation - including molecular condensations, protostars, or ionized gas - is found inside the cavities, whereas such features are observed around the molecular shells. Although the expansion of ionized gas is considered an effective mechanism to trigger star formation, the dynamical ages of the HII regions in our sample are generally not sufficiently long to do so efficiently. Overall, our results hint at the possible importance of negative W-R wind-driven feedback on the gas-deficient cavities, where star formation is quenched as a consequence. In addition, the presence of active star formation around the molecular shells indicates that W-R stars may also assist in accumulating molecular gas, and that they could initiate star formation around those shells.

astro-ph.GA↗

Atomic Scale Measurement of Polar Entropy

Entropy is a fundamental thermodynamic quantity that is a measure of the accessible microstates available to a system, with the stability of a system determined by the magnitude of the total entropy of the system. This is valid across truly mind boggling length scales - from nanoparticles to galaxies. However, quantitative measurements of entropy change using calorimetry are predominantly macroscopic, with direct atomic scale measurements being exceedingly rare. Here for the first time, we experimentally quantify the polar configurational entropy (in meV/K) using sub-ångström resolution aberration corrected scanning transmission electron microscopy. This is performed in a single crystal of the prototypical ferroelectric $\mathsf{LiNbO_3}$ through the quantification of the niobium and oxygen atom column deviations from their paraelectric positions. Significant excursions of the niobium - oxygen polar displacement away from its symmetry constrained direction is seen in single domain regions which increases in the proximity of domain walls. Combined with first principles theory plus mean field effective Hamiltonian methods, we demonstrate the variability in the polar order parameter, which is stabilized by an increase in the magnitude of the configurational entropy. This study presents a powerful tool to quantify entropy from atomic displacements and demonstrates its dominant role in local symmetry breaking at finite temperatures in classic, nominally Ising ferroelectrics.

cond-mat.mtrl-sci↗

Galactic continuum surveys with the new 850 micron MKID camera at the EAO/JCMT 15-m telescope

This white paper gives a brief summary of Galactic continuum surveys with the next generation 850 micron camera at the James Clerk Maxwell Telescope (JCMT) in the next decade. This new camera will have mapping speeds at least 10-20 times faster than the present SCUBA-2 camera, and will enable deep (<10 mJy/beam) and extensive continuum surveys such as a wider (-5 deg.<l<240 deg., |b|<2-5 deg.) Galactic Plane survey, a larger Gould Belt survey, and a follow-up survey of all Planck compact objects visible from the northern hemisphere. These surveys will provide a complete census of molecular clouds, filaments, and dense cores across the Galaxy, vital for studying star formation in various environments.

astro-ph.GA↗

Cloud G074.11+00.11: a stellar cluster in formation

We present molecular line and dust continuum observations of a Planck-detected cold cloud, G074.11+00.11. The cloud consists of a system of curved filaments and a central star-forming clump. The clump is associated with several infrared sources and H2O maser emission. We aim to determine the mass distribution and gas dynamics within the clump, to investigate if the filamentary structure seen around the clump repeats itself on a smaller scale, and to estimate the fractions of mass contained in dense cores and filaments. The velocity distribution of pristine dense gas can be used to investigate the global dynamical state of the clump, the role of filamentary inflows, filament fragmentation and core accretion. We use molecular line and continuum observations from single dish observatories and interferometric facilities to study the kinematics of the region. The molecular line observations show that the central clump may have formed as a result of a large-scale filament collision. The central clump contains three compact cores. Assuming a distance of 2.3 kpc, based on Gaia observations and a three-dimensional extinction method of background stars, the mass of the central clump exceeds 700 solar masses, which is roughly 25% of the total mass of the cloud. Our virial analysis suggests that the central clump and all identified substructures are collapsing. We find no evidence for small-scale filaments associated with the cores. Our observations indicate that the clump is fragmented into three cores with masses in the range [10,50] solar masses and that all three are collapsing. The presence of an H2O maser emission suggests active star formation. However, the CO lines show only weak signs of outflows. We suggest that the region is young and any processes leading to star formation have just recently begun.

astro-ph.GA↗

Learning Blended, Precise Semantic Program Embeddings

Learning neural program embeddings is key to utilizing deep neural networks in program languages research --- precise and efficient program representations enable the application of deep models to a wide range of program analysis tasks. Existing approaches predominately learn to embed programs from their source code, and, as a result, they do not capture deep, precise program semantics. On the other hand, models learned from runtime information critically depend on the quality of program executions, thus leading to trained models with highly variant quality. This paper tackles these inherent weaknesses of prior approaches by introducing a new deep neural network, \liger, which learns program representations from a mixture of symbolic and concrete execution traces. We have evaluated \liger on \coset, a recently proposed benchmark suite for evaluating neural program embeddings. Results show \liger (1) is significantly more accurate than the state-of-the-art syntax-based models Gated Graph Neural Network and code2vec in classifying program semantics, and (2) requires on average 10x fewer executions covering 74\% fewer paths than the state-of-the-art dynamic model \dypro. Furthermore, we extend \liger to predict the name for a method from its body's vector representation. Learning on the same set of functions (more than 170K in total), \liger significantly outperforms code2seq, the previous state-of-the-art for method name prediction.

cs.SE↗

Möbius disjointness for skew products on a circle and a nilmanifold

Let $\mathbb{T}$ be the unit circle and $Γ\backslash G$ the $3$-dimensional Heisenberg nilmanifold. We prove that a class of skew products on $\mathbb{T} \times Γ\backslash G$ are distal, and that the Möbius function is linearly disjoint from these skew products. This verifies the Möbius Disjointness Conjecture of Sarnak.

math.NT↗

Computational MRI with Physics-based Constraints: Application to Multi-contrast and Quantitative Imaging

Compressed sensing takes advantage of low-dimensional signal structure to reduce sampling requirements far below the Nyquist rate. In magnetic resonance imaging (MRI), this often takes the form of sparsity through wavelet transform, finite differences, and low rank extensions. Though powerful, these image priors are phenomenological in nature and do not account for the mechanism behind the image formation. On the other hand, MRI signal dynamics are governed by physical laws, which can be explicitly modeled and used as priors for reconstruction. {1}These explicit and implicit signal priors can be synergistically combined in an inverse problem framework to recover sharp, multi-contrast images from highly accelerated scans. Furthermore, the physics-based constraints provide a recipe for recovering quantitative, bio-physical parameters from the data. This article introduces physics-based modeling constraints in MRI and shows how they can be used in conjunction with compressed sensing for image reconstruction and quantitative imaging. We describe model-based quantitative MRI, as well as its linear subspace approximation. We also discuss approaches to selecting user-controllable scan parameters given knowledge of the physical model. We present several MRI applications that take advantage of this framework for the purpose of multi-contrast imaging and quantitative mapping.

eess.IV↗