arXiv subjects

Kun Xu

Publications and source records attributed to Kun Xu.

At least 235 records · Page 13Linked to original sources

arXiv2021

High-order gas-kinetic scheme in general curvilinear coordinate for iLES of compressible wall-bounded turbulent flows

In this paper, a high-order gas-kinetic scheme in general curvilinear coordinate (HGKS-cur) is developed for the numerical simulation of compressible turbulence. Based on the coordinate transformation, the Bhatnagar-Gross-Krook (BGK) equation is transformed from physical space to computational space. To deal with the general mesh given by discretized points, the geometrical metrics need to be constructed by the dimension-by-dimension Lagrangian interpolation. The multidimensional weighted essentially non-oscillatory (WENO) reconstruction is adopted in the computational domain for spatial accuracy, where the reconstructed variables are the cell averaged Jacobian and the Jacobian-weighted conservative variables. The two-stage fourth-order method, which was developed for spatial-temporal coupled flow solvers, is used for temporal discretization. The numerical examples for inviscid and laminar flows validate the accuracy and geometrical conservation law of HGKS-cur. As a direct application, HGKS-cur is implemented for the implicit large eddy simulation (iLES) in compressible wall-bounded turbulent flows, including the compressible turbulent channel flow and compressible turbulent flow over periodic hills. The iLES results with HGKS-cur are in good agreement with the refereed spectral methods and high-order finite volume methods. The performance of HGKS-cur demonstrates its capability as a powerful tool for the numerical simulation of compressible wall-bounded turbulent flows and massively separated flows.

physics.flu-dyn↗

arXiv2021

Direct Modeling for Computational Fluid Dynamics and the Construction of High-order Compact Scheme for Compressible Flow Simulations

Computational fluid dynamics is a direct modeling of physical laws in a discretized space. The basic physical laws include the mass, momentum and energy conservations, physically consistent transport process, and similar domain of dependence and influence between the physical reality and the numerical representation. Therefore, a physically soundable numerical scheme must be a compact one which involves the closest neighboring cells within the domain of dependence for the solution update under a CFL number $(\sim 1 )$. In the construction of explicit high-order compact scheme, subcell flow distributions or the equivalent degree of freedoms beyond the cell averaged flow variables must be evolved and updated, such as the gradients of the flow variables inside each control volume. The direct modeling of flow evolution under generalized initial condition will be developed in this paper. The direct modeling will provide the updates of flow variables differently on both sides of a cell interface and limit high-order time derivatives of the flux function nonlinearly in case of discontinuity in time, such as a shock wave moving across a cell interface within a time step. The direct modeling unifies the nonlinear limiters in both space for the data reconstruction and time for the time-dependent flux transport. Under the direct modeling framework, as an example, the high-order compact gas-kinetic scheme (GKS) will be constructed. The scheme shows significant improvement in terms of robustness, accuracy, and efficiency in comparison with the previous high-order compact GKS.

physics.flu-dyn↗

arXiv2021

A gradient-compression-based compact high-order gas-kinetic scheme on three-dimensional hybrid unstructured mesh

In this paper, the compact gas-kinetic scheme for compressible flow is extended to hybrid unstructured mesh. Based on both cell-averaged flow variables and their gradients updated from time accurate gas evolution model at cell interfaces, a compact third-order least-square-constrained reconstruction can be obtained on unstructured mesh and a multi-resolution WENO reconstruction is adopted in case of discontinuous solutions. Moreover, a compression factor for the cell-averaged gradients is proposed to take into account the possible discontinuity in flow variables at cell interface, which significantly improves the robustness of the compact scheme for high-speed flow computation on irregular mesh and preserves the accuracy. Numerical tests from incompressible to hypersonic flow are presented to demonstrate the broad applicability of the gradient-compression-based high-order compact scheme.

math.NA↗

arXiv2021

Unified gas-kinetic wave-particle methods VI: Disperse dilute gas-particle multiphase flow

In this paper, a unified gas-kinetic wave-particle scheme (UGKWP) for the disperse dilute gas-particle multiphase flow is proposed. The gas phase is always in the hydrodynamic regime. However, the particle phase covers different flow regimes from particle trajectory crossing to the hydrodynamic wave interaction with the variation of local particle phase Knudsen number. The UGKWP is an appropriate method for the capturing of the multiscale transport mechanism in the particle phase through its coupled wave-particle formulation. In the regime with intensive particle collision, the evolution of solid particle will be followed by the analytic wave with quasi-equilibrium distribution; while in the rarefied regime the non-equilibrium particle phase will be captured through particle tracking and collision, which plays a decisive role in recovering particle trajectory crossing behavior. The gas-kinetic scheme (GKS) is employed for the simulation of gas flow. In the highly collision regime for the particles, no particles will be sampled in UGKWP and the wave formulation for solid particle with the hydrodynamic gas phase will reduce the system to the two-fluid Eulerian model. On the other hand, in the collisionless regime for the solid particle, the free transport of solid particle will be followed in UGKWP, and coupled system will return to the Eulerian-Lagrangian formulation for the gas and particle. The scheme will be tested for in all flow regimes, which include the non-equilibrium particle trajectory crossing, the particle concentration under different Knudsen number, and the dispersion of particle flow with the variation of Stokes number. A experiment of shock-induced particle bed fluidization is simulated and the results are compared with experimental measurements. These numerical solutions validate suitability of the proposed scheme for the simulation of gas-particle multiphase flow.

physics.comp-ph↗

arXiv2021

An Accurate P$^{3}$M Algorithm for Gravitational Lensing Studies in Simulations

We present a two-dimensional (2D) Particle-Particle-Particle-Mesh (P$^3$M) algorithm with an optimized Green function and adaptive softening length for gravitational lensing studies in N-Body simulations. The analytical form of the optimized Green function $\hat{G}(\rm{k})$ is given. The softening schemes ($S$) are studied for both the PM and the PP calculations in order for accurate force calculation and suppression of the particle discreteness effect. Our method is two orders of magnitude more accurate than the simple PM algorithm with the {\it poor man's} Green function ($\propto1/k^2$) at a scale of a few mesh cells or smaller. The force anisotropy is also much smaller than the conventional PM calculation. We can achieve a force accuracy better than 0.1 percent at all scales with our algorithm, which makes it an ideal (accurate and fast) algorithm for {\textit{micro}} lensing studies . When we apply the algorithm to computing {\textit{weak}} and {\textit{strong}} lensing quantities in N-Body simulations, the errors are dominated by the Poisson noise caused by particle discreteness. The Poisson noise can be suppressed by smoothing out the particle distribution, which can be achieved by simply choosing an adaptive softening length in the PP calculation. We have presented a criterion to set the adaptive softening length. Our algorithm is also applicable to cosmological simulations. We provide a \textsc{python} implementation \texttt{P3Mlens} for this algorithm.

astro-ph.CO↗

arXiv2021

Variational (Gradient) Estimate of the Score Function in Energy-based Latent Variable Models

The learning and evaluation of energy-based latent variable models (EBLVMs) without any structural assumptions are highly challenging, because the true posteriors and the partition functions in such models are generally intractable. This paper presents variational estimates of the score function and its gradient with respect to the model parameters in a general EBLVM, referred to as VaES and VaGES respectively. The variational posterior is trained to minimize a certain divergence to the true model posterior and the bias in both estimates can be bounded by the divergence theoretically. With a minimal model assumption, VaES and VaGES can be applied to the kernelized Stein discrepancy (KSD) and score matching (SM)-based methods to learn EBLVMs. Besides, VaES can also be used to estimate the exact Fisher divergence between the data and general EBLVMs.

cs.LG↗

arXiv2021

Domain-Adaptive Pretraining Methods for Dialogue Understanding

Language models like BERT and SpanBERT pretrained on open-domain data have obtained impressive gains on various NLP tasks. In this paper, we probe the effectiveness of domain-adaptive pretraining objectives on downstream tasks. In particular, three objectives, including a novel objective focusing on modeling predicate-argument relations, are evaluated on two challenging dialogue understanding tasks. Experimental results demonstrate that domain-adaptive pretraining with proper objectives can significantly improve the performance of a strong baseline on these tasks, achieving the new state-of-the-art performances.

cs.CL↗

arXiv2021

Do the periodic activities of repeating fast radio bursts represent the spins of neutron stars?

Fast radio bursts (FRBs) are mysterious radio transients with millisecond durations. Recently, a periodic activity of $\sim$16 day and a possible periodicity of $\sim$159 day were detected to arise from FRB 180916.J0158+65 and FRB 121102, respectively, and the spin period of a slow-rotation magnetar was further considered to be one of possibilities to explain the periodic activities of repeating FRBs. For isolated neutron stars, the spin evolution suggests that it's difficult to reach several hours. In this work, we mainly focus on the possible maximum spin period of isolated NSs / magnetars dominated by an interaction between star's magnetic field and the disk. We find that the disk wind plays an important role in spin evolution, whose influence varies the power law index in the evolution equation of mass flow rate. For a magnetar without disk wind, the longest spin period is tens of hours. When the disk wind with a classical parameter is involved, the maximum spin period can reach hundreds of hours. But for a much extremely large index of mass flow rate due to disk wind or other angular momentum extraction processes, a spin period of $\sim$(16-160) days is still possible.

astro-ph.HE↗

arXiv2021

Rethinking and Reweighting the Univariate Losses for Multi-Label Ranking: Consistency and Generalization

(Partial) ranking loss is a commonly used evaluation measure for multi-label classification, which is usually optimized with convex surrogates for computational efficiency. Prior theoretical work on multi-label ranking mainly focuses on (Fisher) consistency analyses. However, there is a gap between existing theory and practice -- some pairwise losses can lead to promising performance but lack consistency, while some univariate losses are consistent but usually have no clear superiority in practice. In this paper, we attempt to fill this gap through a systematic study from two complementary perspectives of consistency and generalization error bounds of learning algorithms. Our results show that learning algorithms with the consistent univariate loss have an error bound of $O(c)$ ($c$ is the number of labels), while algorithms with the inconsistent pairwise loss depend on $O(\sqrt{c})$ as shown in prior work. This explains that the latter can achieve better performance than the former in practice. Moreover, we present an inconsistent reweighted univariate loss-based learning algorithm that enjoys an error bound of $O(\sqrt{c})$ for promising performance as well as the computational efficiency of univariate losses. Finally, experimental results validate our theoretical analyses.

cs.LG↗

arXiv2021

Video-aided Unsupervised Grammar Induction

We investigate video-aided grammar induction, which learns a constituency parser from both unlabeled text and its corresponding video. Existing methods of multi-modal grammar induction focus on learning syntactic grammars from text-image pairs, with promising results showing that the information from static images is useful in induction. However, videos provide even richer information, including not only static objects but also actions and state changes useful for inducing verb phrases. In this paper, we explore rich features (e.g. action, object, scene, audio, face, OCR and speech) from videos, taking the recent Compound PCFG model as the baseline. We further propose a Multi-Modal Compound PCFG model (MMC-PCFG) to effectively aggregate these rich features from different modalities. Our proposed MMC-PCFG is trained end-to-end and outperforms each individual modality and previous state-of-the-art systems on three benchmarks, i.e. DiDeMo, YouCook2 and MSRVTT, confirming the effectiveness of leveraging video information for unsupervised grammar induction.

cs.CV↗

arXiv2021

General Implicit Iterative Method for Unified Gas-kinetic Scheme

In order to further enhance the computational efficiency of the implicit unified gas-kinetic scheme (IUGKS, JCP 315 (2016) 16-38) for multi-scale flow simulation, a two-step IUGKS is proposed in this paper. The multiscale solution of the UGKS is determined by the integral solution of the kinetic model equation, which is composed of the Lagrangian integration of the equilibrium and the free particle transport of the nonequilibrium state. With the implicit evaluation of the macroscopic variables in the first iterative step, the integration of the equilibrium can be directly used in the flux calculation of macroscopic flow variables in the second iterative step. This is equivalent to include the viscous flux in the implicit scheme to accelerate the convergence of the solution instead of using the Euler flux in the iterative process of the original IUGKS. In the present IUGKS, the update of macroscopic flow variables are closely coupled with the implicit evolution of the gas distribution function. At the same time, in order to get the more accurate solution, the full Boltzmann collision term is incorporated into the current scheme through the penalty method. Different iterative techniques, such as LU-SGS and multi-grid, are used for solving the linear algebraic system of coupled macroscopic and microscopic equations. The efficiency of the IUGKS has reached a favorable level among all implicit schemes for the kinetic equations in the literature. Several numerical examples are used to validate the performance of the IUGKS. Accurate solutions have been obtained efficiently in all flow regimes from low speed to hypersonic ones.

physics.comp-ph↗

arXiv2021

Improving Weakly Supervised Visual Grounding by Contrastive Knowledge Distillation

Weakly supervised phrase grounding aims at learning region-phrase correspondences using only image-sentence pairs. A major challenge thus lies in the missing links between image regions and sentence phrases during training. To address this challenge, we leverage a generic object detector at training time, and propose a contrastive learning framework that accounts for both region-phrase and image-sentence matching. Our core innovation is the learning of a region-phrase score function, based on which an image-sentence score function is further constructed. Importantly, our region-phrase score function is learned by distilling from soft matching scores between the detected object names and candidate phrases within an image-sentence pair, while the image-sentence score function is supervised by ground-truth image-sentence pairs. The design of such score functions removes the need of object detection at test time, thereby significantly reducing the inference cost. Without bells and whistles, our approach achieves state-of-the-art results on visual phrase grounding, surpassing previous methods that require expensive object detectors at test time.

cs.CV↗

arXiv2021

Conversational Semantic Role Labeling

Semantic role labeling (SRL) aims to extract the arguments for each predicate in an input sentence. Traditional SRL can fail to analyze dialogues because it only works on every single sentence, while ellipsis and anaphora frequently occur in dialogues. To address this problem, we propose the conversational SRL task, where an argument can be the dialogue participants, a phrase in the dialogue history or the current sentence. As the existing SRL datasets are in the sentence level, we manually annotate semantic roles for 3,000 chit-chat dialogues (27,198 sentences) to boost the research in this direction. Experiments show that while traditional SRL systems (even with the help of coreference resolution or rewriting) perform poorly for analyzing dialogues, modeling dialogue histories and participants greatly helps the performance, indicating that adapting SRL to conversations is very promising for universal dialogue understanding. Our initial study by applying CSRL to two mainstream conversational tasks, dialogue response generation and dialogue context rewriting, also confirms the usefulness of CSRL.

cs.CL↗

arXiv2021

GraphGallery: A Platform for Fast Benchmarking and Easy Development of Graph Neural Networks Based Intelligent Software

Graph Neural Networks (GNNs) have recently shown to be powerful tools for representing and analyzing graph data. So far GNNs is becoming an increasingly critical role in software engineering including program analysis, type inference, and code representation. In this paper, we introduce GraphGallery, a platform for fast benchmarking and easy development of GNNs based software. GraphGallery is an easy-to-use platform that allows developers to automatically deploy GNNs even with less domain-specific knowledge. It offers a set of implementations of common GNN models based on mainstream deep learning frameworks. In addition, existing GNNs toolboxes such as PyG and DGL can be easily incorporated into the platform. Experiments demonstrate the reliability of implementations and superiority in fast coding. The official source code of GraphGallery is available at https://github.com/EdisonLeeeee/GraphGallery and a demo video can be found at https://youtu.be/mv7Zs1YeaYo.

cs.AI↗

arXiv2021

Structural Information Preserving for Graph-to-Text Generation

The task of graph-to-text generation aims at producing sentences that preserve the meaning of input graphs. As a crucial defect, the current state-of-the-art models may mess up or even drop the core structural information of input graphs when generating outputs. We propose to tackle this problem by leveraging richer training signals that can guide our model for preserving input information. In particular, we introduce two types of autoencoding losses, each individually focusing on different aspects (a.k.a. views) of input graphs. The losses are then back-propagated to better calibrate our model via multi-task training. Experiments on two benchmarks for graph-to-text generation show the effectiveness of our approach over a state-of-the-art baseline. Our code is available at \url{http://github.com/Soistesimmer/AMR-multiview}.

cs.CL↗

arXiv2021

A giant central red disk galaxy at redshift $z=0.76$: challenge to theories of galaxy formation

We report a giant red central disk galaxy in the XMM-LSS north region. The region is covered with a rich variety of multiband photometric and spectroscopic observations. Using the photometric data of the Canada-France-Hawaii Telescope Legacy Survey (CFHTLS) and spectroscopic observation of the Baryon Oscillation Spectroscopic Survey (BOSS), we find that the galaxy has a stellar mass of $\sim 10^{11.6}$ times of the solar mass $M_\odot$. The galaxy has a red color and has an old stellar population, and thus its star formation has stopped. With the photometric image data of Hyper Suprime-Cam (HSC) Subaru Strategic Program, we demonstrate that its luminosity profile is perfectly described by a Sérsic form with $n=1.22$ indicating disk morphology. We also analyze its environment based on the VIMOS Public Extragalactic Redshift Survey (VIPERS) photometric catalog, and find that its close neighbors are all less massive, indicating that our observed galaxy is sitting at the center of its host halo. Existence of the giant red central disk galaxy seriously challenges the current standard paradigm of galaxy formation, as there is no known physical mechanism to explain the quenching of its star formation. This conclusion is supported by state-of-the-art hydrodynamical simulations of galaxy formation.

astro-ph.GA↗

arXiv2021

Effect of the anomalous magnetic moment of quarks on magnetized QCD matter and meson spectra

We systematically investigate the effect of the anomalous magnetic moment(AMM) of quarks on the magnetized QCD matter, including the magnetic susceptibility, the inverse magnetic catalysis around the critical temperature and the neutral/charged pion and rho meson spectra under magnetic fields. The dynamical AMM of quarks, its coupling with magnetic field causes Zeeman splitting of the dispersion relation of quarks thus changes the magnetism properties and meson mass spectra under magnetic fields. It is found that including the AMM of quarks cannot fully understand lattice results of the magnetized matter: The AMM of quarks reduces the dynamical quark mass thus causes the inverse magnetic catalysis around $T_c$. The neutral pion mass is very sensitive to the AMM, it decreases with magnetic field quickly, and the charged pion mass shows a nonlinear behavior, i.e., firstly linearly increases with the magnetic field and then saturates at strong magnetic field. For rho meson, it is observed that AMM reduces the mass of neutral rho meson mass with different $s_z$, and reduces the mass of $s_z=+1,0$ component charged rho meson mass but enhances the $s_z=-1$ component charged rho meson mass. The magnetic susceptibility at low temperature can be either positive or negative with different AMM.

hep-ph↗

arXiv2021

The first decade of unified gas kinetic scheme

In 2010, the unified gas kinetic scheme (UGKS) was proposed by Xu et al . (A unified gas-kinetic scheme for continuum and rarefied flows, Journal of Computational Physics, 2010). In the past decade, many numerical techniques have been developed to improve the capability of the UGKS in the aspects of efficiency increment, memory reduction, and physical modeling. The methodology of the direct modeling of the UGKS on discretization scale provides a general framework for construction of multiscale method for multiscale transport processes. This paper reviews the development and extension of the UGKS in its first decade.

physics.comp-ph↗