arXiv subjects

Bin Liu

Publications and source records attributed to Bin Liu.

At least 415 records · Page 23Linked to original sources

arXiv2020

Cross-modality Person re-identification with Shared-Specific Feature Transfer

Cross-modality person re-identification (cm-ReID) is a challenging but key technology for intelligent video analysis. Existing works mainly focus on learning common representation by embedding different modalities into a same feature space. However, only learning the common characteristics means great information loss, lowering the upper bound of feature distinctiveness. In this paper, we tackle the above limitation by proposing a novel cross-modality shared-specific feature transfer algorithm (termed cm-SSFT) to explore the potential of both the modality-shared information and the modality-specific characteristics to boost the re-identification performance. We model the affinities of different modality samples according to the shared features and then transfer both shared and specific features among and across modalities. We also propose a complementary feature learning strategy including modality adaption, project adversarial learning and reconstruction enhancement to learn discriminative and complementary shared and specific features of each modality, respectively. The entire cm-SSFT algorithm can be trained in an end-to-end manner. We conducted comprehensive experiments to validate the superiority of the overall algorithm and the effectiveness of each component. The proposed algorithm significantly outperforms state-of-the-arts by 22.5% and 19.3% mAP on the two mainstream benchmark datasets SYSU-MM01 and RegDB, respectively.

cs.CV↗

arXiv2020

Uncovering Insurance Fraud Conspiracy with Network Learning

Fraudulent claim detection is one of the greatest challenges the insurance industry faces. Alibaba's return-freight insurance, providing return-shipping postage compensations over product return on the e-commerce platform, receives thousands of potentially fraudulent claims every day. Such deliberate abuse of the insurance policy could lead to heavy financial losses. In order to detect and prevent fraudulent insurance claims, we developed a novel data-driven procedure to identify groups of organized fraudsters, one of the major contributions to financial losses, by learning network information. In this paper, we introduce a device-sharing network among claimants, followed by developing an automated solution for fraud detection based on graph learning algorithms, to separate fraudsters from regular customers and uncover groups of organized fraudsters. This solution applied at Alibaba achieves more than 80% precision while covering 44% more suspicious accounts compared with a previously deployed rule-based classifier after human expert investigations. Our approach can easily and effectively generalizes to other types of insurance.

cs.CR↗

arXiv2020

The Role of Long-lived Excitons in the Dynamics of Strongly Coupled Molecular Polaritons

The concept of modifying molecular dynamics in strongly coupled exciton-polariton systems is an emerging topic in photonics due to its potential to produce customized chemical systems with tailored photophysical properties. However, before such systems can be realized, it is essential to address the open questions concerning the nature and strength of electronic interactions between exciton-polaritons and localized excited states in chemical system as well as the proper way to measure such interactions. Here, we use transient optical spectroscopy to investigate dynamical interactions between exciton-polaritons, singlet excitons, and triplet excitons in a molecular singlet fission system that is strongly coupled to an optical microcavity. We identify some of the major limitations to modify molecular dynamics in the strong coupling regime. Simultaneous excitation of cavity polaritons and 'reservoir' states, defined as dark polaritons and dark excitons (e.g. triplets) from coupled molecules and excitons from uncoupled molecules, always occurs. In addition, slow conversion from reservoir states to cavity polaritons results in minimal changes to the overall population dynamics. Furthermore, we demonstrate how in addition to the usual population dynamics, transient optical measurements on microcavities reveal information pertaining to modification of the exciton-polariton transition energies due to changes in the population of molecular excited states and the exciton-photon coupling conditions. As a consequence of weak interactions between reservoir states and cavity polaritons, judicious design considerations are required to achieve modified chemical dynamics, necessitating the use of molecular systems with long excited-state lifetimes or strong coupling approaches that require a small number of molecules.

cond-mat.mtrl-sci↗

arXiv2020

Sequential online prediction in the presence of outliers and change points: an instant temporal structure learning approach

In this paper, we consider sequential online prediction (SOP) for streaming data in the presence of outliers and change points. We propose an INstant TEmporal structure Learning (INTEL) algorithm to address this problem. Our INTEL algorithm is developed based on a full consideration of the duality between online prediction and anomaly detection. We first employ a mixture of weighted Gaussian process models (WGPs) to cover the expected possible temporal structures of the data. Then, based on the rich modeling capacity of this WGP mixture, we develop an efficient technique to instantly learn (capture) the temporal structure of the data that follows a regime shift. This instant learning is achieved only by adjusting one hyper-parameter value of the mixture model. A weighted generalization of the product of experts (POE) model is used for fusing predictions yielded from multiple GP models. An outlier is declared once a real observation seriously deviates from the fused prediction. If a certain number of outliers are consecutively declared, then a change point is declared. Extensive experiments are performed using a diverse of real datasets. Results show that the proposed algorithm is significantly better than benchmark methods for SOP in the presence of outliers and change points.

cs.LG↗

arXiv2020

Spatial and spectral deep attention fusion for multi-channel speech separation using deep embedding features

Multi-channel deep clustering (MDC) has acquired a good performance for speech separation. However, MDC only applies the spatial features as the additional information. So it is difficult to learn mutual relationship between spatial and spectral features. Besides, the training objective of MDC is defined at embedding vectors, rather than real separated sources, which may damage the separation performance. In this work, we propose a deep attention fusion method to dynamically control the weights of the spectral and spatial features and combine them deeply. In addition, to solve the training objective problem of MDC, the real separated sources are used as the training objectives. Specifically, we apply the deep clustering network to extract deep embedding features. Instead of using the unsupervised K-means clustering to estimate binary masks, another supervised network is utilized to learn soft masks from these deep embedding features. Our experiments are conducted on a spatialized reverberant version of WSJ0-2mix dataset. Experimental results show that the proposed method outperforms MDC baseline and even better than the oracle ideal binary mask (IBM).

eess.AS↗

arXiv2020

The Fundamental Performance of FAST with 19-beam Receiver at L Band

The Five-hundred-meter Aperture Spherical radio Telescope (FAST) passed national acceptance and is taking pilot cycle of 'Shared-Risk' observations. The 19-beam receiver covering 1.05-1.45 GHz was used for most of these observations. The electronics gain fluctuation of the system is better than 1\% over 3.5 hours, enabling enough stability for observations. Pointing accuracy, aperture efficiency and system temperature are three key parameters of FAST. The measured standard deviation of pointing accuracy is 7.9$''$, which satisfies the initial design of FAST. When zenith angle is less than 26.4$^\circ$, the aperture efficiency and system temperature around 1.4 GHz are $\sim$ 0.63 and less than 24 K for central beam, respectively. The measured value of these two parameters are better than designed value of 0.6 and 25 K, respectively. The sensitivity and stability of the 19-beam backend are confirmed to satisfy expectation by spectral HI observations toward N672 and polarization observations toward 3C286. The performance allows FAST to take sensitive observations in various scientific goals, from studies of pulsar to galaxy evolution.

astro-ph.IM↗

arXiv2020

Thermal relaxation of magnons and phonons near resonance points in magnetic insulators

We theoretically investigate the energy relaxation rate of magnons and phonons near the resonance points to clarify the underlying mechanism of heat transport in ferromagnetic materials. We find that the simple two-temperature model is valid for the one-phonon/one-magnon process, as the rate of energy exchange between magnons and phonons is proportional to the temperature difference between them, and it is independent of temperature in the high temperature limit. We found that the magnon-phonon relaxation time due to the one-phonon/one-magnon interaction could be reduced to 1.48 $μs$ at the resonance point by applying an external magnetic field. It means that the resonance effect plays a significant role in enhancing the total magnon-phonon energy exchange rate, apart from the higher order interaction processes.

cond-mat.mes-hall↗

arXiv2020

Unbalance Mitigation via Phase-switching Device and Static Var Compensator in Low-voltage Distribution Network

As rooftop solar PVs installed by residential customers penetrate in low voltage distribution network (LVDN), some issues, e.g. over/under voltage and unbalances, which may undermine the network's operational performance, need to be effectively addressed. To mitigate unbalances in LVDN, dynamic switching devices (PSDs) and static var compensator (SVC) are two equipment that are cost-effective and efficient. However, most existing research on operating PSDs are based on inflexible heuristic algorithms or without considering the network formulation, which may lead to strategies that violate operational requirements. Moreover, few pieces of literature have been reported on mitigating unbalances in LVDN via SVC and PSDs together. This paper, after presenting the dispatch model of SVC, formulates the decision-making process as a mixed-integer non-convex programming (MINCP) problem considering all practical operational requirements. To efficiently solve the challenging problem, the MINCP is reformulated as a mixed-integer second order-cone programming (MISOCP) problem based on either exact reformulations or accurate approximations, making it possible to employ efficient off-the-shelf solvers. Simulations based on a modified IEEE system and a practical system in Australia demonstrates the efficiency of the proposed method in mitigating unbalances in LVDN.

math.OC↗

arXiv2020

Data-Driven Model Set Design for Model Averaged Particle Filter

This paper is concerned with sequential state filtering in the presence of nonlinearity, non-Gaussianity and model uncertainty. For this problem, the Bayesian model averaged particle filter (BMAPF) is perhaps one of the most efficient solutions. Major advances of BMAPF have been made, while it still lacks a generic and practical approach to design the model set. This paper fills in this gap by proposing a generic data-driven method for BMAPF model set design. Unlike existent methods, the proposed solution does not require any prior knowledge on the parameter value of the true model; it only assumes that a small number of noisy observations are pre-obtained. The Bayesian optimization (BO) method is adapted to search the model components, each of which is associated with a specific segment of the pre-obtained dataset.The average performance of these model components is guaranteed since each one's parameter value is elaborately tuned via BO to maximize the marginal likelihood. The diversity in the model components is also ensured, as different components match the different segments of the pre-obtained dataset, respectively. Computer simulations are used to demonstrate the effectiveness of the proposed method.

stat.CO↗

arXiv2020

Polymorphism and superconductivity in the V-Nb-Mo-Al-Ga high-entropy alloys

High-entropy alloys (HEAs) are at the focus of current research for their diverse properties, including superconductivity and structural polymorphism. However, the polymorphic transition has been observed only in nonsuperconducting HEAs and mostly under high pressure. Here we report the discovery of superconductivity and temperature-driven polymorphism in the (V$_{0.5}$Nb$_{0.5}$)$_{3-x}$Mo$_{x}$Al$_{0.5}$Ga$_{0.5}$ (0.2 $\leq$ $x$ $\leq$ 1.4) HEAs. It is found that the as-cast HEA is of a single body-centered cubic (bcc) structure for $x$ = 0.2 and a mixture of the bcc and A15 structures for higher $x$ values. Upon annealing, the bcc structure undergoes a polymorphic transformation to the A15 one and all HEAs exhibits bulk superconductivity. For $x$ = 0.2, whereas the bcc polymorph is not superconducting down to 1.8 K, the A15 polymorph has a superconducting transition temperature $T_{\rm c}$ of 10.2 K and an estimated zero-temperature upper critical field $B_{\rm c2}$(0) of 20.1 T, both of which are the highest among HEA superconductors. With increasing Mo content $x$, both $T_{\rm c}$ and $B_{\rm c2}$(0) of the A15-type HEAs decrease, yet the large ratio of $B_{\rm c2}$(0)/$T_{\rm c}$ signifies a disorder-induced enhancement of the upper critical field over a wide $x$ range. The decrease in $T_{\rm c}$ is attributed to the decrease in both the electronic specific-heat coefficient and electron-phonon coupling strength. Furthermore, the valence electron count dependence of $T_{\rm c}$, which is different from both the binary A15 and other structurally different HEA superconductors, suggests that $T_{\rm c}$ may be increased further by reducing the number of valence electrons. Our results not only uncover HEA superconductors of a new structural type, but also provide the first example of polymorphism dependent superconductivity in HEAs.

cond-mat.supr-con↗

arXiv2020

A Survey on Trust Modeling from a Bayesian Perspective

In this paper, we are concerned with trust modeling for agents in networked computing systems. As trust is a subjective notion that is invisible, implicit and uncertain in nature, many attempts have been made to model trust with aid of Bayesian probability theory, while the field lacks a global comprehensive analysis for variants of Bayesian trust models. We present a study to fill in this gap by giving a comprehensive review of the literature. A generic Bayesian trust (GBT) modeling perspective is highlighted here. It is shown that all models under survey can cast into a GBT based computing paradigm as special cases. We discuss both capabilities and limitations of the GBT perspective and point out open questions to answer, with a hope to advance GBT to become a pragmatic infrastructure for analyzing intrinsic relationships among variants of trust models and developing novel tools for trust evaluation.

cs.CR↗

arXiv2020

Particle Filtering Methods for Stochastic Optimization with Application to Large-Scale Empirical Risk Minimization

This paper is concerned with sequential filtering based stochastic optimization (FSO) approaches that leverage a probabilistic perspective to implement the incremental proximity method (IPM). The present FSO methods are derived based on the Kalman filter (KF) and the extended KF (EKF). In contrast with typical methods such as stochastic gradient descent (SGD) and IPMs, they do not need to pre-schedule the learning rate for convergence. Nevertheless, they have limitations that inherit from the KF mechanism. As the particle filtering (PF) method outperforms KF and its variants remarkably for nonlinear non-Gaussian sequential filtering problems, it is natural to ask if FSO methods can benefit from PF to get around of their limitations. We provide an affirmative answer to this question by developing two PF based stochastic optimizers (PFSOs). For performance evaluation, we apply them to address nonlinear least-square fitting with simulated data, and empirical risk minimization for binary classification of real data sets. Experimental results demonstrate that PFSOs outperform remarkably a benchmark SGD algorithm, the vanilla IPM, and KF-type FSO methods in terms of numerical stability, convergence speed, and flexibility in handling diverse types of loss functions.

cs.LG↗

arXiv2019

A general criterion for solid instability and its application to creases

A general force-perturbation-based criterion for solid instability is proposed, which can predict instability including crease without priori knowledge of instability configuration. The crease instability is analyzed in detail, we found that the occurrence of solid instability does not always correspond to the non-positive definiteness of global stiffness matrix. An element stiffness-based criterion based on material stiffness is proposed as a stronger criterion in order to fast determine the occurrence of instability. This criterion has been shown to degenerate into the criterion for judging instability of certain known phenomena, such as necking and shear band phenomena. Besides, instability in strongly anisotropic materials is also predicted by the element stiffness-based criterion.

cond-mat.soft↗

arXiv2019

The GBT Diffuse Ionized Gas Survey: Tracing the Diffuse Ionized Gas around the Giant HII Region W43

The Green Bank Telescope (GBT) Diffuse Ionized Gas Survey (GDIGS) is a fully-sampled radio recombination line (RRL) survey of the inner Galaxy at C-band (4-8 GHz). We average together ~15 Hn$α$ RRLs within the receiver bandpass to improve the spectral signal-to-noise ratio. The average beam size for the RRL observations at these frequencies is ~2'. We grid these data to have spatial and velocity spacings of 30" and 0.5 km/s, respectively. Here we discuss the first RRL data from GDIGS: a six square-degree-area surrounding the Galactic HII region complex W43. We attempt to create a map devoid of emission from discrete HII regions and detect RRL emission from the diffuse ionized gas (DIG) across nearly the entire mapped area. We estimate the intensity of the DIG emission by a simple empirical model, taking only the HII region locations, angular sizes, and RRL intensities into account. The DIG emission is predominantly found at two distinct velocities: ~40 km/s and ~100 km/s. While the 100 km/s component is associated with W43 at a distance of ~6 kpc, the origin of the 40 km/s component is less clear. Since the distribution of the 40 km/s emission cannot be adequately explained by ionizing sources at the same velocity, we hypothesize that the plasma at the two velocity components is interacting, placing the 40 km/s DIG at a similar distance as the 100 km/s emission. We find a correlation between dust temperature and integrated RRL intensity, suggesting that the same radiation field that heats the dust also maintains the ionization of the DIG.

astro-ph.GA↗

arXiv2019

Harnessing Low-Fidelity Data to Accelerate Bayesian Optimization via Posterior Regularization

Bayesian optimization (BO) is a powerful paradigm for derivative-free global optimization of a black-box objective function (BOF) that is expensive to evaluate. However, the overhead of BO can still be prohibitive for problems with highly expensive function evaluations. In this paper, we investigate how to reduce the required number of function evaluations for BO without compromise in solution quality. We explore the idea of posterior regularization to harness low fidelity (LF) data within the Gaussian process upper confidence bound (GP-UCB) framework. The LF data can arise from previous evaluations of an LF approximation of the BOF or of a related optimization task. An extra GP model called LF-GP is trained to fit the LF data. We develop an operator termed dynamic weighted product of experts (DW-POE) fusion. The regularization is induced by this operator on the posterior of the BOF. The impact of the LF GP model on the resulting regularized posterior is adaptively adjusted via Bayesian formalism. Extensive experimental results on benchmark BOF optimization tasks demonstrate the superior performance of the proposed algorithm over state-of-the-art.

cs.LG↗

arXiv2019

Lunar impact craters identification and age estimation with Chang'E data by deep and transfer learning

Impact craters, as "lunar fossils", are the most dominant lunar surface features and occupy most of the Moon's surface. Their formation and evolution record the history of the Solar System. Sixty years of triumphs in the lunar exploration projects accumulated a large amount of lunar data. Currently, there are 9137 existing recognized craters. However, only 1675 of them have been determined age, which is obviously not satisfactory to reveal the evolution of the Moon. Identifying craters is a challenging task due to their enormous difference in size, large variations in shape and vast presence. Furthermore, estimating the age of craters is extraordinarily difficult due to their complex and different morphologies. Here, in order to effectively identify craters and estimate their age, we convert the crater identification problem into a target detection task and crater age estimation into a taxonomy structure. From an initial small number of available craters, we progressively identify craters and estimate their age from Chang'E data by transfer learning (TL) using deep neural networks. For comprehensive identification of multi-scale craters, a two-stage craters detection approach is developed. Thus 117240 unrecognized lunar craters that range in diameter from 532 km to 1 km are identified. Then, a two-stage classification approach is developed to estimate the age of craters by simultaneously extracting their morphological features and stratigraphic information. The age of 79243 craters larger than 3 km in diameter is estimated. These identified and aged craters throughout the mid and low-latitude regions of the Moon are crucial for reconstructing the dynamic evolution process of the Solar System.

astro-ph.EP↗

arXiv2019

An Efficient MILP Formulation of Economic Dispatch with Adjustable Transformer Ratio and Phase Shifter

In this short paper, we study the economic dispatch with adjustable transformer ratio and phase shifter, both of which, along with the transmission line, are formulated into a generalized branch model. Resulted nonlinear parts are thereafter exactly linearized using the piecewise liner technique to make the derived ED problem computationally tractable. Numerical studies based on modified IEEE systems demonstrate the effectiveness of the proposed method to efficiency and flexibility of power system operation.

math.OC↗

arXiv2019

Improving Operational Feasibility of Low-voltage Distribution Network by Phase-switching Devices

High penetration of residential photovoltaic (PV) in low-voltage distribution networks (LVDN) makes the unbalance issue more sever due to the asymmetry of generation/load characteristic in different phases, and may lead to infeasible operation of the whole network. Phase switching device (PSD), which can switch the connected phase of a residential load as required, is viable and efficient equipment to help address this issue. This paper, based on three-phase power flow (TUPF) formulation, aims to investigate the benefit of PSD on improving the operation feasibility of LVDN. The linear model of TUPF with PSD is presented to effectively take the flexible device into account, which can be conveniently used to seek the PSD positions that lead to a viable operation strategy of the original problem when infeasibility is reported by the traditional iteration-based algorithm. Case study based on a practical LVDN in Australia demonstrates the efficacy of the proposed method.

math.OC↗