arXiv subjects

Chen Wu

Publications and source records attributed to Chen Wu.

At least 235 records · Page 13Linked to original sources

arXiv2018

Object Tracking in Satellite Videos Based on a Multi-Frame Optical Flow Tracker

Object tracking is a hot topic in computer vision. Thanks to the booming of the very high resolution (VHR) remote sensing techniques, it is now possible to track targets of interests in satellite videos. However, since the targets in the satellite videos are usually too small compared with the entire image, and too similar with the background, most state-of-the-art algorithms failed to track the target in satellite videos with a satisfactory accuracy. Due to the fact that optical flow shows the great potential to detect even the slight movement of the targets, we proposed a multi-frame optical flow tracker (MOFT) for object tracking in satellite videos. The Lucas-Kanade optical flow method was fused with the HSV color system and integral image to track the targets in the satellite videos, while multi-frame difference method was utilized in the optical flow tracker for a better interpretation. The experiments with three VHR remote sensing satellite video datasets indicate that compared with state-of-the-art object tracking algorithms, the proposed method can track the target more accurately.

cs.CV↗

arXiv2018

Multi-class Active Learning: A Hybrid Informative and Representative Criterion Inspired Approach

Labeling each instance in a large dataset is extremely labor- and time- consuming . One way to alleviate this problem is active learning, which aims to which discover the most valuable instances for labeling to construct a powerful classifier. Considering both informativeness and representativeness provides a promising way to design a practical active learning. However, most existing active learning methods select instances favoring either informativeness or representativeness. Meanwhile, many are designed based on the binary class, so that they may present suboptimal solutions on the datasets with multiple classes. In this paper, a hybrid informative and representative criterion based multi-class active learning approach is proposed. We combine the informative informativeness and representativeness into one formula, which can be solved under a unified framework. The informativeness is measured by the margin minimum while the representative information is measured by the maximum mean discrepancy. By minimizing the upper bound for the true risk, we generalize the empirical risk minimization principle to the active learning setting. Simultaneously, our proposed method makes full use of the label information, and the proposed active learning is designed based on multiple classes. So the proposed method is not suitable to the binary class but also the multiple classes. We conduct our experiments on twelve benchmark UCI data sets, and the experimental results demonstrate that the proposed method performs better than some state-of-the-art methods.

cs.LG↗

arXiv2018

High-resolution Observations of Low-luminosity Gigahertz-Peaked Spectrum and Compact Steep Spectrum Sources

We present Very Long Baseline Interferometry observations of a faint and low-luminosity ($L_{\rm 1.4 GHz} < 10^{27}~\mbox{W Hz}^{-1}$) Gigahertz-Peaked Spectrum (GPS) and Compact Steep Spectrum (CSS) sample. We select eight sources from deep radio observations that have radio spectra characteristic of a GPS or CSS source and an angular size of $θ\lesssim 2$ arcsec, and detect six of them with the Australian Long Baseline Array. We determine their linear sizes, and model their radio spectra using Synchrotron Self Absorption (SSA) and Free Free Absorption (FFA) models. We derive statistical model ages, based on a fitted scaling relation, and spectral ages, based on the radio spectrum, which are generally consistent with the hypothesis that GPS and CSS sources are young and evolving. We resolve the morphology of one CSS source with a radio luminosity of $10^{25}~\mbox{W Hz}^{-1}$, and find what appear to be two hotspots spanning 1.7 kpc. We find that our sources follow the turnover-linear size relation, and that both homogenous SSA and an inhomogeneous FFA model can account for the spectra with observable turnovers. All but one of the FFA models do not require a spectral break to account for the radio spectrum, while all but one of the alternative SSA and power law models do require a spectral break to account for the radio spectrum. We conclude that our low-luminosity sample is similar to brighter samples in terms of their spectral shape, turnover frequencies, linear sizes, and ages, but cannot test for a difference in morphology.

astro-ph.GA↗

arXiv2017

Session-aware Information Embedding for E-commerce Product Recommendation

Most of the existing recommender systems assume that user's visiting history can be constantly recorded. However, in recent online services, the user identification may be usually unknown and only limited online user behaviors can be used. It is of great importance to model the temporal online user behaviors and conduct recommendation for the anonymous users. In this paper, we propose a list-wise deep neural network based architecture to model the limited user behaviors within each session. To train the model efficiently, we first design a session embedding method to pre-train a session representation, which incorporates different kinds of user search behaviors such as clicks and views. Based on the learnt session representation, we further propose a list-wise ranking model to generate the recommendation result for each anonymous user session. We conduct quantitative experiments on a recently published dataset from an e-commerce company. The evaluation results validate the effectiveness of the proposed method, which can outperform the state-of-the-art significantly.

cs.IR↗

arXiv2017

A third-order discrete unified gas kinetic scheme for continuum and rarefied flows: low-speed isothermal case

An efficient third-order discrete unified gas kinetic scheme (DUGKS) with efficiency is presented in this work for simulating continuum and rarefied flows. By employing two-stage time-stepping scheme and the high-order DUGKS flux reconstruction strategy, third-order of accuracy in both time and space can be achieved in the present method. It is also analytically proven that the second-order DUGKS is a special case of the present method. Compared with the high-order lattice Boltzmann equation {LBE} based methods, the present method is capable to deal with the rarefied flows by adopting the Newton-Cotes quadrature to approximate the integrals of moments. Instead of being constrained by the second-order (or lower-order) of accuracy in time splitting scheme as in the conventional high-order Runge-Kutta (RK) based kinetic methods, the present method solves the original BE, which overcomes the limitation in time accuracy. Typical benchmark tests are carried out for comprehensive evaluation of the present method. It is observed in the tests that the present method is advantageous over the original DUGKS in accuracy and capturing delicate flow structures. Moreover, the efficiency of the present third-order method is also shown in simulating rarefied flows.

physics.flu-dyn↗

arXiv2017

Ensemble Methods for Personalized E-Commerce Search Challenge at CIKM Cup 2016

Personalized search has been a hot research topic for many years and has been widely used in e-commerce. This paper describes our solution to tackle the challenge of personalized e-commerce search at CIKM Cup 2016. The goal of this competition is to predict search relevance and re-rank the result items in SERP according to the personalized search, browsing and purchasing preferences. Based on a detailed analysis of the provided data, we extract three different types of features, i.e., statistic features, query-item features and session features. Different models are used on these features, including logistic regression, gradient boosted decision trees, rank svm and a novel deep match model. With the blending of multiple models, a stacking ensemble model is built to integrate the output of individual models and produce a more accurate prediction result. Based on these efforts, our solution won the champion of the competition on all the evaluation metrics.

cs.IR↗

arXiv2017

A high resolution foreground model for the MWA EoR1 field: model and implications for EoR power spectrum analysis

The current generation of experiments aiming to detect the neutral hydrogen signal from the Epoch of Reionisation (EoR) is likely to be limited by systematic effects associated with removing foreground sources from target fields. In this paper we develop a model for the compact foreground sources in one of the target fields of the MWA's EoR key science experiment: the `EoR1' field. The model is based on both the MWA's GLEAM survey and GMRT 150 MHz data from the TGSS survey, the latter providing higher angular resolution and better astrometric accuracy for compact sources than is available from the MWA alone. The model contains 5049 sources, some of which have complicated morphology in MWA data, Fornax A being the most complex. The higher resolution data show that 13% of sources that appear point-like to the MWA have complicated morphology such as double and quad structure, with a typical separation of 33~arcsec. We derive an analytic expression for the error introduced into the EoR two-dimensional power spectrum due to peeling close double sources as single point sources and show that for the measured source properties, the error in the power spectrum is confined to high $k_\bot$ modes that do not affect the overall result for the large-scale cosmological signal of interest. The brightest ten mis-modelled sources in the field contribute 90% of the power bias in the data, suggesting that it is most critical to improve the models of the brightest sources. With this hybrid model we reprocess data from the EoR1 field and show a maximum of 8% improved calibration accuracy and a factor of two reduction in residual power in $k$-space from peeling these sources. Implications for future EoR experiments including the SKA are discussed in relation to the improvements obtained.

astro-ph.IM↗

arXiv2017

DALiuGE: A Graph Execution Framework for Harnessing the Astronomical Data Deluge

The Data Activated Liu Graph Engine - DALiuGE - is an execution framework for processing large astronomical datasets at a scale required by the Square Kilometre Array Phase 1 (SKA1). It includes an interface for expressing complex data reduction pipelines consisting of both data sets and algorithmic components and an implementation run-time to execute such pipelines on distributed resources. By mapping the logical view of a pipeline to its physical realisation, DALiuGE separates the concerns of multiple stakeholders, allowing them to collectively optimise large-scale data processing solutions in a coherent manner. The execution in DALiuGE is data-activated, where each individual data item autonomously triggers the processing on itself. Such decentralisation also makes the execution framework very scalable and flexible, supporting pipeline sizes ranging from less than ten tasks running on a laptop to tens of millions of concurrent tasks on the second fastest supercomputer in the world. DALiuGE has been used in production for reducing interferometry data sets from the Karl E. Jansky Very Large Array and the Mingantu Ultrawide Spectral Radioheliograph; and is being developed as the execution framework prototype for the Science Data Processor (SDP) consortium of the Square Kilometre Array (SKA) telescope. This paper presents a technical overview of DALiuGE and discusses case studies from the CHILES and MUSER projects that use DALiuGE to execute production pipelines. In a companion paper, we provide in-depth analysis of DALiuGE's scalability to very large numbers of tasks on two supercomputing facilities.

cs.DC↗

arXiv2017

Quasinormal modes of gravitational perturbation around some well-known regular black holes

In this paper, the quasinormal modes of gravitational perturbation around some well-known regular black holes were evaluated by using the WKB approximation as well as the asymptotic iteration method. Through numerical calculation, we make a detailed analysis of the gravitational QNM frequencies by varying the characteristic parameters of the gravitational perturbation and the spacetime charge parameters of the regular black holes. It is found that the imaginary part of quasinormal modes as a function of the charge parameter has different monotonic behaviors for different black hole spacetimes. Moreover, the asymptotic expressions of gravitational QNMs for $l \gg 1$ are obtained by using the eikonal limit method. We demonstrate that the gravitational perturbation is stable in all these spacetimes.

gr-qc↗

arXiv2017

A study of halo and relic radio emission in merging clusters using the Murchison Widefield Array

We have studied radio haloes and relics in nine merging galaxy clusters using the Murchison Widefield Array (MWA). The images used for this study were obtained from the GaLactic and Extragalactic All-sky MWA (GLEAM) Survey which was carried out at 5 frequencies, viz. 88, 118, 154, 188 and 215 MHz. We detect diffuse radio emission in 8 of these clusters. We have estimated the spectra of haloes and relics in these clusters over the frequency range 80-1400 MHz; the first such attempt to estimate their spectra at low frequencies. The spectra follow a power law with a mean value of $α= -1.13\pm0.21$ for haloes and $α= -1.2\pm0.19$ for relics where, $S \propto ν^α$. We reclassify two of the cluster sources as radio galaxies. The low frequency spectra are thus an independent means of confirming the nature of cluster sources. Five of the nine clusters host radio haloes. For the remaining four clusters, we place upper limits on the radio powers of possible haloes in them. These upper limits are a factor of 2-20 below those expected from the $L_{\rm X}-P_{\rm 1.4}$ relation. These limits are the lowest ever obtained and the implications of these limits to the hadronic model of halo emission are discussed.

astro-ph.CO↗

arXiv2016

Neutron stars including the effects of chaotic magnetic fields and the anomalous magnetic moments

In this work the relativistic mean field (RMF) FSUGold model extended to include hyperons is employed to study the properties of neutron stars with strong magnetic fields. The chaotic magnetic field approximation is utilized. The effect of the anomalous magnetic moments (AMMs) is also investigated. It is shown that the equation of state (EOS)of neutron star matter is stiffened by the presence of the magnetic field, which increases the maximum mass of neutron star by around 6%. The AMMs only have a small influence on the EOS of neutron star matter, and increase the maximum mass of neutron star by 0.02 M$_{sun}$. Neutral particles are spin polarized due to the presence of the AMMs.

astro-ph.HE↗

arXiv2016

Scalar field quasinormal frequencies of Reissner-Nordström black hole surrounded by quintessence by using he continued fraction method

We evaluate the quasinormal modes of massless scalar field around Reissner-Nordstr$\ddot{\text{o}}$m black hole surrounded by a static and spherically symmetric quintessence by using the continued fraction method. The appropriate Frobenius series for three special cases of the quintessence parameter $ ε= -1/3, -2/3$ and $-1$ are derived successfully. We show that the variation of quasinormal frequencies with charge of the black hole and the quintessential parameters. The numerical results show that quintessence field decreases oscillation frequencies of all angular momentum $l$ modes and increases the damping time of $l>0$ modes.

gr-qc↗

arXiv2016

GaLactic and Extragalactic All-sky Murchison Widefield Array (GLEAM) survey I: A low-frequency extragalactic catalogue

Using the Murchison Widefield Array (MWA), the low-frequency Square Kilometre Array (SKA1 LOW) precursor located in Western Australia, we have completed the GaLactic and Extragalactic All-sky MWA (GLEAM) survey, and present the resulting extragalactic catalogue, utilising the first year of observations. The catalogue covers 24,831 square degrees, over declinations south of $+30^\circ$ and Galactic latitudes outside $10^\circ$ of the Galactic plane, excluding some areas such as the Magellanic Clouds. It contains 307,455 radio sources with 20 separate flux density measurements across 72--231MHz, selected from a time- and frequency- integrated image centred at 200MHz, with a resolution of $\approx 2$'. Over the catalogued region, we estimate that the catalogue is 90% complete at 170mJy, and 50% complete at 55mJy, and large areas are complete at even lower flux density levels. Its reliability is 99.97% above the detection threshold of $5σ$, which itself is typically 50mJy. These observations constitute the widest fractional bandwidth and largest sky area survey at radio frequencies to date, and calibrate the low frequency flux density scale of the southern sky to better than 10%. This paper presents details of the flagging, imaging, mosaicking, and source extraction/characterisation, as well as estimates of the completeness and reliability. All source measurements and images are available online (http://www.mwatelescope.org/science/gleam-survey). This is the first in a series of publications describing the GLEAM survey results.

astro-ph.GA↗

arXiv2016

Effect of stochastic transition in the fundamental diagram of traffic flow

In this work, we propose an alternative stochastic model for the fundamental diagram of traffic flow with minimal number of parameters. Our approach is based on a mesoscopic viewpoint of the traffic system in terms of the dynamics of vehicle speed transitions. A key feature of the present approach lies in its stochastic nature which makes it possible to study not only the flow-concentration relation, namely, the fundamental diagram, but also its uncertainty, namely, the variance of the fundamental diagram \textemdash an important characteristic in the observed traffic flow data. It is shown that in the simplified versions of the model consisting of only a few speed states, analytic solutions for both quantities can be obtained, which facilitate the discussion of the corresponding physical content. We also show that the effect of vehicle size can be included into the model by introducing the maximal congestion density $k_{max}$. By making use of this parameter, the free flow region and congested flow region are naturally divided, and the transition is characterized by the capacity drop at the maximum of the flow-concentration relation. The model parameters are then adjusted to the observed traffic flow on the I-80 Freeway Dataset in the San Francisco area from the NGSIM program, where both the fundamental diagram and its variance are reasonably reproduced. Despite its simplicity, we argue that the current model provides an alternative description for the fundamental diagram and its uncertainty in the study of traffic flow.

physics.soc-ph↗

arXiv2015

Imaging SKA-Scale data in three different computing environments

We present the results of our investigations into options for the computing platform for the imaging pipeline in the CHILES project, an ultra-deep HI pathfinder for the era of the Square Kilometre Array. CHILES pushes the current computing infrastructure to its limits and understanding how to deliver the images from this project is clarifying the Science Data Processing requirements for the SKA. We have tested three platforms: a moderately sized cluster, a massive High Performance Computing (HPC) system, and the Amazon Web Services (AWS) cloud computing platform. We have used well-established tools for data reduction and performance measurement to investigate the behaviour of these platforms for the complicated access patterns of real-life Radio Astronomy data reduction. All of these platforms have strengths and weaknesses and the system tools allow us to identify and evaluate them in a quantitative manner. With the insights from these tests we are able to complete the imaging pipeline processing on both the HPC platform and also on the cloud computing platform, which paves the way for meeting big data challenges in the era of SKA in the field of Radio Astronomy. We discuss the implications that all similar projects will have to consider, in both performance and costs, to make recommendations for the planning of Radio Astronomy imaging workflows.

astro-ph.IM↗

arXiv2015

Superfluidity of $Λ$ hyperons in warm strange hadronic star matter

In this work we evaluate the $^1S_0$ superfluidity of $Λ$ hyperons in $β$-stable strange hadronic matter. We investigate the equation of state (EOS) of hadronic matter at finite values of baryon density and temperature in the relativistic mean field (RMF) theory. Effects of the introduced isoscalar-isovector cross-interaction term on the $Λ$ superfluidity are investigated systematically. In addition, the temperature effects on the superfluidity of $Λ$ hyperons in hadronic matter are discussed. It is found that the density region and the magnitude of the $Λ$ pairing gap are dependent on the cross-interaction term. The obtained maximal critical temperature of $Λ$ superfluid is around $10^9$ K.

nucl-th↗

arXiv2015

Decay of massive scalar field in a black hole background immersed in magnetic field

We evaluate quasinormal modes of massive scalar field of the Ernst spacetime, an exact solution of the Einstein-Maxwell equations describing a black hole immersed in uniform magnetic field $B$. It is known that the quasinormal spectrum for massive scalar field in the vicinity of the magnetized black holes acquires an effective mass $μ_{eff}= \sqrt{4B^2 m^2+μ^2}$, where $m$ is the azimuthal number and $μ$ is the mass of scalar field. The numerical result shows that increasing of the field effective mass and the magnetic field $B$ gives rise to decreasing of the imaginary part of the quasinormal modes until reaching the vanishing damping rate.

gr-qc↗

arXiv2015

Preprint Virtual Reality GIS and Cloud Service Based Traffic Analysis Platform

This is the preprint version of our paper on The 23rd International Conference on Geoinformatics (Geoinformatics2015). City traffic data has several characteristics, such as large scale, diverse predictable and real-time, which falls in the range of definition of Big Data. This paper proposed a cloud service platform which targets for wise transportation is to carry out unified management and mining analysis of the huge number of the multivariate and heterogeneous dynamic transportation information, provides real-time transportation information, increase the utilization efficiency of transportation, promote transportation management and service level of travel information and provide decision support of transportation management by virtual reality as visual.

cs.OH↗