arXiv subjects

Kai Zhang

Publications and source records attributed to Kai Zhang.

At least 595 records · Page 33Linked to original sources

arXiv2020

LyaCoLoRe: Synthetic Datasets for Current and Future Lyman-$α$ Forest BAO Surveys

The statistical power of Lyman-$α$ forest Baryon Acoustic Oscillation (BAO) measurements is set to increase significantly in the coming years as new instruments such as the Dark Energy Spectroscopic Instrument deliver progressively more constraining data. Generating mock datasets for such measurements will be important for validating analysis pipelines and evaluating the effects of systematics. With such studies in mind, we present LyaCoLoRe: a package for producing synthetic Lyman-$α$ forest survey datasets for BAO analyses. LyaCoLoRe transforms initial Gaussian random field skewers into skewers of transmitted flux fraction via a number of fast approximations. In this work we explain the methods of producing mock datasets used in LyaCoLoRe, and then measure correlation functions on a suite of realisations of such data. We demonstrate that we are able to recover the correct BAO signal, as well as large-scale bias parameters similar to literature values. Finally, we briefly describe methods to add further astrophysical effects to our skewers - high column density systems and metal absorbers - which act as potential complications for BAO analyses.

astro-ph.CO↗

arXiv2020

Cascaded Semantic and Positional Self-Attention Network for Document Classification

Transformers have shown great success in learning representations for language modelling. However, an open challenge still remains on how to systematically aggregate semantic information (word embedding) with positional (or temporal) information (word orders). In this work, we propose a new architecture to aggregate the two sources of information using cascaded semantic and positional self-attention network (CSPAN) in the context of document classification. The CSPAN uses a semantic self-attention layer cascaded with Bi-LSTM to process the semantic and positional information in a sequential manner, and then adaptively combine them together through a residue connection. Compared with commonly used positional encoding schemes, CSPAN can exploit the interaction between semantics and word positions in a more interpretable and adaptive manner, and the classification performance can be notably improved while simultaneously preserving a compact model size and high convergence rate. We evaluate the CSPAN model on several benchmark data sets for document classification with careful ablation studies, and demonstrate the encouraging results compared with state of the art.

cs.CL↗

arXiv2020

AIM 2020 Challenge on Efficient Super-Resolution: Methods and Results

This paper reviews the AIM 2020 challenge on efficient single image super-resolution with focus on the proposed solutions and results. The challenge task was to super-resolve an input image with a magnification factor x4 based on a set of prior examples of low and corresponding high resolution images. The goal is to devise a network that reduces one or several aspects such as runtime, parameter count, FLOPs, activations, and memory consumption while at least maintaining PSNR of MSRResNet. The track had 150 registered participants, and 25 teams submitted the final results. They gauge the state-of-the-art in efficient single image super-resolution.

eess.IV↗

arXiv2020

Correspondence between winding numbers and skin modes in non-hermitian systems

We establish exact relations between the winding of "energy" (eigenvalue of Hamiltonian) on the complex plane as momentum traverses the Brillouin zone with periodic boundary condition, and the presence of "skin modes" with open boundary condition in non-hermitian systems. We show that the nonzero winding with respect to any complex reference energy leads to the presence of skin modes, and vice versa. We also show that both the nonzero winding and the presence of skin modes share the common physical origin that is the non-vanishing current through the system.

cond-mat.mes-hall↗

arXiv2020

DHP: Differentiable Meta Pruning via HyperNetworks

Network pruning has been the driving force for the acceleration of neural networks and the alleviation of model storage/transmission burden. With the advent of AutoML and neural architecture search (NAS), pruning has become topical with automatic mechanism and searching based architecture optimization. Yet, current automatic designs rely on either reinforcement learning or evolutionary algorithm. Due to the non-differentiability of those algorithms, the pruning algorithm needs a long searching stage before reaching the convergence. To circumvent this problem, this paper introduces a differentiable pruning method via hypernetworks for automatic network pruning. The specifically designed hypernetworks take latent vectors as input and generate the weight parameters of the backbone network. The latent vectors control the output channels of the convolutional layers in the backbone network and act as a handle for the pruning of the layers. By enforcing $\ell_1$ sparsity regularization to the latent vectors and utilizing proximal gradient solver, sparse latent vectors can be obtained. Passing the sparsified latent vectors through the hypernetworks, the corresponding slices of the generated weight parameters can be removed, achieving the effect of network pruning. The latent vectors of all the layers are pruned together, resulting in an automatic layer configuration. Extensive experiments are conducted on various networks for image classification, single image super-resolution, and denoising. And the experimental results validate the proposed method.

cs.CV↗

arXiv2020

Fast and efficient critical state modelling of field-cooled bulk high-temperature superconductors using a backward computation method

A backward computation method has been developed to accelerate modelling of the critical state magnetization current in a staggered-array bulk high-temperature superconducting (HTS) undulator. The key concept is as follows: i) a large magnetization current is first generated on the surface of the HTS bulks after rapid field-cooling (FC) magnetization; ii) the magnetization current then relaxes inwards step-by-step obeying the critical state model; iii) after tens of backward iterations the magnetization current reaches a steady state. The simulation results show excellent agreement with the H-formulation method for both the electromagnetic and electromagnetic-mechanical coupled analyses, but with significantly faster computation speed. Solving the FEA model with 1.8 million degrees of freedom (DOFs), the backward computation method takes less than 1.4 hours, an order of magnitude or higher faster than other state-of-the-art numerical methods. Finally, the models are used to investigate the influence of the mechanical stress on the distribution of the critical state magnetization current and the undulator field along the central axis.

cond-mat.supr-con↗

arXiv2020

The Effect of Bars on the Ionized ISM: Optical Emission Lines from Milky Way Analogs

Gas interior to the bar of the Milky Way has recently been shown as the closest example of a Low Ionization (Nuclear) Emission Region--LI(N)ER--in the universe. To better understand the nature of this gas, a sample of face-on galaxies with integral field spectroscopy are used to study the ionized gas conditions of 240 barred and 250 nonbarred galaxies, focusing on those that are most similar to the Milky Way. Strong optical line emission of $[NII]$ $λ6584$, H$α$, $[OIII]$ $λ5007$, and H$β$ are used to diagnose the dominant ionization mechanisms of gas across galaxies and the Galaxy via Baldwin-Phillips-Terlevich (BPT) Diagrams. Barred galaxies show a strong suppression of star formation and an increase in composite and LI(N)ER like spectra in their inner regions when compared with similar nonbarred counterparts. This effect is lessened in galaxies of very low ($\log_{10}(M_\star/M_\odot) \lesssim 10.4$) or very high ($\log_{10}(M_\star/M_\odot) \gtrsim 11.1$) total stellar mass. Bar masks from Galaxy Zoo:3D show the bar's non-axisymmetric effect on the ionized gas and help predict the face-on distribution of ionized gas conditions near the bar of the Milky Way.

astro-ph.GA↗

arXiv2020

Structural Landmarking and Interaction Modelling: on Resolution Dilemmas in Graph Classification

Graph neural networks are promising architecture for learning and inference with graph-structured data. Yet difficulties in modelling the ``parts'' and their ``interactions'' still persist in terms of graph classification, where graph-level representations are usually obtained by squeezing the whole graph into a single vector through graph pooling. From complex systems point of view, mixing all the parts of a system together can affect both model interpretability and predictive performance, because properties of a complex system arise largely from the interaction among its components. We analyze the intrinsic difficulty in graph classification under the unified concept of ``resolution dilemmas'' with learning theoretic recovery guarantees, and propose ``SLIM'', an inductive neural network model for Structural Landmarking and Interaction Modelling. It turns out, that by solving the resolution dilemmas, and leveraging explicit interacting relation between component parts of a graph to explain its complexity, SLIM is more interpretable, accurate, and offers new insight in graph representation learning.

cs.LG↗

arXiv2020

Upper boundaries of AGN regions in optical diagnostic diagrams

The distribution of galaxies in optical diagnostic diagrams can provide information about their physical parameters when compared with ionization models under proper assumptions. By using a sample of central emitting regions from the MaNGA survey, we find evidence of the existence of upper boundaries for narrow-line regions (NLRs) of active galactic nuclei (AGN) in optical BPT diagrams, especially in the diagrams involving [S II]$λλ$6716, 6731/H$α$. Photoionization models can well reproduce the boundaries as a consequence of the decrease of [S II]$λλ$6716, 6731/H$α$ and [O III]$λ$5007/H$β$ ratios at very high metallicity. Whilst the exact location of the upper boundary in the [S II] BPT diagram only weakly depends on the electron density of the ionized cloud and the secondary nitrogen prescription, its dependence on the shapes of the input SEDs is much stronger. This allows us to constrain the power-law index of the AGN SED between 1 Ryd and $\sim100$ Ryd to be less than or equal to $-1.40\pm 0.05$. The coverage of the photoionization models in the [N II] BPT diagram has a stronger dependence on the electron density and the secondary nitrogen prescription. With the density constrained by the [S II] doublet ratio and the input SED constrained by the [S II] BPT diagram, we find that the extent of the data in the [N II] BPT diagram favors those prescriptions with high N/O ratios. Although shock-ionized cloud can produce similar line ratios as those by photoionization, the resulting shapes of the upper boundaries, if exist, would likely be different from those of a photoionizing origin.

astro-ph.GA↗

arXiv2020

Boundary Hölder Regularity for Elliptic Equations

This paper investigates the relation between the boundary geometric properties and the boundary regularity of the solutions of elliptic equations. We prove by a new unified method the pointwise boundary Hölder regularity under proper geometric conditions. "Unified" means that our method is applicable for the Laplace equation, linear elliptic equations in divergence and non-divergence form, fully nonlinear elliptic equations, the $p-$Laplace equations and the fractional Laplace equations etc. In addition, these geometric conditions are quite general. In particular, for local equations, the measure of the complement of the domain near the boundary point concerned could be zero. The key observation in the method is that the strong maximum principle implies a decay for the solution, then a scaling argument leads to the Hölder regularity. Moreover, we also give a geometric condition, which guarantees the solvability of the Dirichlet problem for the Laplace equation. The geometric meaning of this condition is more apparent than that of the Wiener criterion.

math.AP↗

arXiv2020

Laminar separated flows over finite-aspect-ratio swept wings

We perform direct numerical simulations of laminar separated flows over finite-aspect-ratio swept wings at a chord-base Reynolds number of $Re = 400$ to reveal a variety of wake structures generated for a range of aspect ratios ($sAR=0.5-4$), angles of attack ($α=16^{\circ}-30^{\circ}$), and sweep angles ($Λ=0^{\circ}-45^{\circ}$). Flows behind swept wings exhibit increased complexity in their dynamical features compared to unswept-wing wakes. For unswept wings, the wake dynamics are predominantly influenced by the tip effects. Steady wakes are mainly limited to low-aspect-ratio wings. Unsteady vortex shedding takes place near the midspan of higher-$AR$ wings due to weakened downwash induced by the tip vortices. With increasing sweep angle, the source of three dimensionality transitions from the tip to the midspan. A pair of symmetric vortical structures forms along the two sides of the midspan, imposing downward velocity upon each other in a manner similar to the downwash induced by tip vortices. Such stabilizing midspan effects not only expand the steady wake region to higher aspect ratios, but also enhance lift. At higher aspect ratios, the midspan effects of swept wings diminish at outboard region, allowing unsteady vortex shedding to develop near the tip. In the wakes of highly swept wings, streamwise finger-like structures form repetitively along the wing span, providing a stabilizing effect. The insights revealed from this study can aid the design of high-lift devices and serve as a stepping stone for understanding the complex wake dynamics at higher Reynolds numbers and those generated by unsteady wing maneuvers.

physics.flu-dyn↗

arXiv2020

The backbone-residual model. Accurately characterising the instrumental profile of a fibre-fed echelle spectrograph

Context: Instrumental profile (IP) is the basic property of a spectrograph. Accurate IP characterisation is the prerequisite of accurate wavelength solution. It also facilitates new spectral acquisition methods such as the forward modeling and deconvolution. Aims: We investigate an IP modeling method for the fibre-fed echelle spectrograph with the emission lines of the ThAr lamp, and explore the method to evaluate the accuracy of IP characterisation. Methods: The backbone-residual (BR) model is put forward and tested on the fibre-fed High Resolution Spectrograph (HRS) at the Chinese Xinglong 2.16-m Telescope, which is the sum of the backbone function and the residual function. The backbone function is a bell-shaped function to describe the main component and the spatial variation of IP. The residual function, which is expressed as the cubic spline function, accounts for the difference between the bell-shaped function and the actual IP. The method of evaluating the accuracy of IP characterisation is based on the spectral reconstruction and Monte Carlo simulation. Results: The IP of HRS is characterised with the BR model, and the accuracy of the characterised IP reaches 0.006 of the peak value of the backbone function. This result demonstrates that the accurate IP characterisation has been achieved on HRS with the BR model, and the BR model is an excellent choice for accurate IP characterisation of fibre-fed echelle spectrographs.

astro-ph.IM↗

arXiv2020

Learning Context-Based Non-local Entropy Modeling for Image Compression

The entropy of the codes usually serves as the rate loss in the recent learned lossy image compression methods. Precise estimation of the probabilistic distribution of the codes plays a vital role in the performance. However, existing deep learning based entropy modeling methods generally assume the latent codes are statistically independent or depend on some side information or local context, which fails to take the global similarity within the context into account and thus hinder the accurate entropy estimation. To address this issue, we propose a non-local operation for context modeling by employing the global similarity within the context. Specifically, we first introduce the proxy similarity functions and spatial masks to handle the missing reference problem in context modeling. Then, we combine the local and the global context via a non-local attention block and employ it in masked convolutional networks for entropy modeling. The entropy model is further adopted as the rate loss in a joint rate-distortion optimization to guide the training of the analysis transform and the synthesis transform network in transforming coding framework. Considering that the width of the transforms is essential in training low distortion models, we finally produce a U-Net block in the transforms to increase the width with manageable memory consumption and time complexity. Experiments on Kodak and Tecnick datasets demonstrate the superiority of the proposed context-based non-local attention block in entropy modeling and the U-Net block in low distortion compression against the existing image compression standards and recent deep image compression models.

eess.IV↗

arXiv2020

Glass formation in binary alloys with different atomic symmetries

Prediction of the glass forming ability (GFA) of alloys remains a major challenge. We are not able to predict the composition dependence of the GFA of even binary alloys. To investigate the effect of each element's propensity to form particular crystal structures on glass formation, we focus on binary alloys composed of elements with the same size, but different atomic symmetries using the patchy-particle model. For mixtures with atomic symmetries that promote different crystal structures, the minimum critical cooling rate $R_c$ is only a factor of $5$ lower than that for the pure substances. For mixtures with different atomic symmetries that promote local crystalline and icosahedral order, the minimum $R_c$ is more than $3$ orders of magnitude lower than that for pure substances. Results for $R_c$ for the patchy-particle model are in agreement with those from embedded atom method simulations and sputtering experiments of NiCu, TiAl, and high entropy alloys.

cond-mat.mtrl-sci↗

arXiv2020

Helical damping and anomalous critical non-Hermitian skin effect

Non-Hermitian skin effect and critical skin effect are unique features of non-Hermitian systems. In this Letter, we study an open system with its dynamics of single-particle correlation function effectively dominated by a non-Hermitian damping matrix, which exhibits $\mathbb{Z}_2$ skin effect, and uncover the existence of a novel phenomenon of helical damping. When adding perturbations that break anomalous time reversal symmetry to the system, the critical skin effect occurs, which causes the disappearance of the helical damping in the thermodynamic limit although it can exist in small size systems. We also demonstrate the existence of anomalous critical skin effect when we couple two identical systems with $\mathbb{Z}_2$ skin effect. With the help of non-Bloch band theory, we unveil that the change of generalized Brillouin zone equation is the necessary condition of critical skin effect.

cond-mat.mes-hall↗

arXiv2020

NTIRE 2020 Challenge on Perceptual Extreme Super-Resolution: Methods and Results

This paper reviews the NTIRE 2020 challenge on perceptual extreme super-resolution with focus on proposed solutions and results. The challenge task was to super-resolve an input image with a magnification factor 16 based on a set of prior examples of low and corresponding high resolution images. The goal is to obtain a network design capable to produce high resolution results with the best perceptual quality and similar to the ground truth. The track had 280 registered participants, and 19 teams submitted the final results. They gauge the state-of-the-art in single image super-resolution.

eess.IV↗

arXiv2020

Depth Sensing Beyond LiDAR Range

Depth sensing is a critical component of autonomous driving technologies, but today's LiDAR- or stereo camera-based solutions have limited range. We seek to increase the maximum range of self-driving vehicles' depth perception modules for the sake of better safety. To that end, we propose a novel three-camera system that utilizes small field of view cameras. Our system, along with our novel algorithm for computing metric depth, does not require full pre-calibration and can output dense depth maps with practically acceptable accuracy for scenes and objects at long distances not well covered by most commercial LiDARs.

cs.CV↗

arXiv2020

A Method for Curation of Web-Scraped Face Image Datasets

Web-scraped, in-the-wild datasets have become the norm in face recognition research. The numbers of subjects and images acquired in web-scraped datasets are usually very large, with number of images on the millions scale. A variety of issues occur when collecting a dataset in-the-wild, including images with the wrong identity label, duplicate images, duplicate subjects and variation in quality. With the number of images being in the millions, a manual cleaning procedure is not feasible. But fully automated methods used to date result in a less-than-ideal level of clean dataset. We propose a semi-automated method, where the goal is to have a clean dataset for testing face recognition methods, with similar quality across men and women, to support comparison of accuracy across gender. Our approach removes near-duplicate images, merges duplicate subjects, corrects mislabeled images, and removes images outside a defined range of pose and quality. We conduct the curation on the Asian Face Dataset (AFD) and VGGFace2 test dataset. The experiments show that a state-of-the-art method achieves a much higher accuracy on the datasets after they are curated. Finally, we release our cleaned versions of both datasets to the research community.

cs.CV↗