arXiv subjects

Jie Liu

Publications and source records attributed to Jie Liu.

At least 631 records · Page 35Linked to original sources

arXiv2018

Performance Degradation Assessment for Electrical Machines Based on SOM and Hybrid DHMM

Aimed at the timely detection of the degradation of electrical machines and the organization of active maintenance, numerous studies on performance degradation assessment have been conducted. However, previous research still suffers from two deficiencies: 1) determining the relevant relationship among diverse machine degradation states and assessing the specific degree of deterioration and 2) determining the evolutionary relationships among degradation and failure modes and assessing the failure modes corresponding to different degradation scenarios. To address these two deficiencies, a novel performance degradation assessment method is proposed. First, the self-organizing feature map (SOM) network is used to mine the latent degradation states of electrical machines. Second, the latent states are quantified according to established statistical health indexes, and by analyzing the distribution of extracted health indexes corresponding to different degradation states, the relevant transition relationships of the valid degradation states and the final evolving fault types are determined. Third, a hybrid discrete HMM is developed to fully describe the transition process among different states and assess the degradation scenario of a machine in an online manner. The results of a real application of an electric point machine show that the proposed method can identify valid degradation states and obtain a superior assessment accuracy.

eess.SP↗

arXiv2018

Degradation Detection Method for Railway Point Machines

Point machines (PMs) are used for switching and locking railway turnouts, and are considered one of the most critical elements of a railway signal system. The failure of the point mechanism directly affects the operation of the railway and may cause serious safety accidents. Hence, there is a need for early detection of the anomalies in PMs. From normal operation to complete failure, the machine usually undergoes a series of degradation states. If the degradation states are detected in time, maintenance can be organized in advance to prevent the malfunction. This paper presents a degradation detection method that can effectively mine and identify the degradation state of the PM. First, power data is processed to obtain the feature set that can describe the PM characteristics effectively. Then, a clustering analysis of the feature set is carried out by self-organizing feature-mapping network, and various degradation states are mined. Finally, the optimized support vector machine is used to build the state classifier to identify the degradation state of the PM. The experimental results obtained with the Siemens S700K PM show that the proposed method could not only mine the effective degradation states, but also obtain high identification accuracy.

eess.SP↗

arXiv2018

Note on quasi-polarized canonical Calabi-Yau threefolds

Let $(X,L)$ be a quasi-polarized canonical Calabi-Yau threefold. In this note, we show that $\vert mL\vert$ is basepoint free for $m\geq 4$. Moreover, if the morphism $Φ_{\vert 4L\vert}$ is not birational onto its image and $h^0(X,L)\geq 2$, then $L^3=1$. As an application, if $Y$ is a $n$-dimensional Fano manifold such that $-K_Y=(n-3)H$ for some ample divisor $H$, then $\vert mH\vert$ is basepoint free for $m\geq 4$ and if the morphism $Φ_{\vert 4H\vert}$ is not birational onto its image, then $Y$ is either a weighted hypersurface of degree $10$ in the weighted projective space $\mathbb{P}(1,\cdots,1,2,5)$ or $h^0(Y,H)=n-2$.

math.AG↗

arXiv2018

On the Acceleration of L-BFGS with Second-Order Information and Stochastic Batches

This paper proposes a framework of L-BFGS based on the (approximate) second-order information with stochastic batches, as a novel approach to the finite-sum minimization problems. Different from the classical L-BFGS where stochastic batches lead to instability, we use a smooth estimate for the evaluations of the gradient differences while achieving acceleration by well-scaling the initial Hessians. We provide theoretical analyses for both convex and nonconvex cases. In addition, we demonstrate that within the popular applications of least-square and cross-entropy losses, the algorithm admits a simple implementation in the distributed environment. Numerical experiments support the efficiency of our algorithms.

cs.LG↗

arXiv2018

A Spatial and Temporal Features Mixture Model with Body Parts for Video-based Person Re-Identification

The video-based person re-identification is to recognize a person under different cameras, which is a crucial task applied in visual surveillance system. Most previous methods mainly focused on the feature of full body in the frame. In this paper we propose a novel Spatial and Temporal Features Mixture Model (STFMM) based on convolutional neural network (CNN) and recurrent neural network (RNN), in which the human body is split into $N$ parts in horizontal direction so that we can obtain more specific features. The proposed method skillfully integrates features of each part to achieve more expressive representation of each person. We first split the video sequence into $N$ part sequences which include the information of head, waist, legs and so on. Then the features are extracted by STFMM whose $2N$ inputs are obtained from the developed Siamese network, and these features are combined into a discriminative representation for one person. Experiments are conducted on the iLIDS-VID and PRID-2011 datasets. The results demonstrate that our approach outperforms existing methods for video-based person re-identification. It achieves a rank-1 CMC accuracy of 74\% on the iLIDS-VID dataset, exceeding the the most recently developed method ASTPN by 12\%. For the cross-data testing, our method achieves a rank-1 CMC accuracy of 48\% exceeding the ASTPN method by 18\%, which shows that our model has significant stability.

cs.CV↗

arXiv2018

A New High Performance and Scalable SVD algorithm on Distributed Memory Systems

This paper introduces a high performance implementation of \texttt{Zolo-SVD} algorithm on distributed memory systems, which is based on the polar decomposition (PD) algorithm via the Zolotarev's function (\texttt{Zolo-PD}), originally proposed by Nakatsukasa and Freund [SIAM Review, 2016]. Our implementation highly relies on the routines of ScaLAPACK and therefore it is portable. Compared with the other PD algorithms such as the QR-based dynamically weighted Halley method (\texttt{QDWH-PD}), \texttt{Zolo-PD} is naturally parallelizable and has better scalability though performs more floating-point operations. When using many processes, \texttt{Zolo-PD} is usually 1.20 times faster than \texttt{QDWH-PD} algorithm, and \texttt{Zolo-SVD} can be about two times faster than the ScaLAPACK routine \texttt{\texttt{PDGESVD}}. These numerical experiments are performed on Tianhe-2 supercomputer, one of the fastest supercomputers in the world, and the tested matrices include some sparse matrices from particular applications and some randomly generated dense matrices with different dimensions. Our \texttt{QDWH-SVD} and \texttt{Zolo-SVD} implementations are freely available at https://github.com/shengguolsg/Zolo-SVD.

cs.DC↗

arXiv2018

Planar Superconductor-Insulator-Superconductor Mixer Array Receivers for Wide Field of View Astronomical Observation

We present a conceptual framework of planar SIS mixer array receivers and the studies on the required techniques. This concept features membrane-based on-chip waveguide probes and a quasi-two-dimensional local-oscillator distribution waveguide network. This concept allows sophisticated functions, such as dual-polarization, balanced mixing and sideband separation, easily implemented with the SIS mixer array in the same planar circuit. We have developed a single-pixel prototype receiver by implementing the concept in the design. Initial measurement results show good evidences that support the feasibility of the concept.

astro-ph.IM↗

arXiv2018

Techniques for Efficiently Handling Power Surges in Fuel Cell Powered Data Centers: Modeling, Analysis, Results

Fuel cells are a promising power source for future data centers, offering high energy efficiency, low greenhouse gas emissions, and high reliability. However, due to mechanical limitations related to fuel delivery, fuel cells are slow to adjust to sudden increases in data center power demands, which can result in temporary power shortfalls. To mitigate the impact of power shortfalls, prior work has proposed to either perform power capping by throttling the servers, or to leverage energy storage devices (ESDs) that can temporarily provide enough power to make up for the shortfall while the fuel cells ramp up power generation. Both approaches have disadvantages: power capping conservatively limits server performance and can lead to service level agreement (SLA) violations, while ESD-only solutions must significantly overprovision the energy storage device capacity to tolerate the shortfalls caused by the worst-case (i.e., largest) power surges, which greatly increases the total cost of ownership (TCO). We propose SizeCap, the first ESD sizing framework for fuel cell powered data centers, which coordinates ESD sizing with power capping to enable a cost-effective solution to power shortfalls in data centers. SizeCap sizes the ESD just large enough to cover the majority of power surges, but not the worst-case surges that occur infrequently, to greatly reduce TCO. It then uses the smaller capacity ESD in conjunction with power capping to cover the power shortfalls caused by the worst-case power surges. As part of our new flexible framework, we propose multiple power capping policies with different degrees of awareness of fuel cell and workload behavior, and evaluate their impact on workload performance and ESD size. Using traces from Microsoft's production data center systems, we demonstrate that SizeCap significantly reduces the ESD size without violating any SLAs.

cs.DC↗

arXiv2018

On sound insulation of pyramidal lattice sandwich structure

Pyramidal lattice sandwich structure (PLSS) exhibits high stiffness and strength-to-weight ratio which can be effectively utilized for designing light-weight load bearing structures for ranging from ground to aerospace vehicles. While these structures provide superior strength to weigh ratio, their sound insulation capacity has not been well understood. The aim of this study is to develop numerical and experimental methods to fundamentally investigate the sound insulation property of the pyramidal lattice sandwich structure with solid trusses (PLSSST). A finite element model has been developed to predict the sound transmission loss (STL) of PLSSST and simulation results have been compared with those obtained experimentally. Parametric studies is then performed using the validated finite element model to investigate the effect of different parameters in pyramidal lattice sandwich structure with hollow trusses (PLSSHT), revealing that the pitching angle, the uniform thickness and the length of the hollow truss and the lattice constant have considerable effects on the sound transmission loss. Finally a design optimization strategy has been formulated to optimize PLSSHT in order to maximize STL while meeting mechanical property requirements. It has been shown that STL of the optimal PLSSHT can be increased by almost 10% at the low-frequency band. The work reported here provides useful information for the noise reduction design of periodic lattice structures.

physics.app-ph↗

arXiv2018

Question Answering over Freebase via Attentive RNN with Similarity Matrix based CNN

With the rapid growth of knowledge bases (KBs), question answering over knowledge base, a.k.a. KBQA has drawn huge attention in recent years. Most of the existing KBQA methods follow so called encoder-compare framework. They map the question and the KB facts to a common embedding space, in which the similarity between the question vector and the fact vectors can be conveniently computed. This, however, inevitably loses original words interaction information. To preserve more original information, we propose an attentive recurrent neural network with similarity matrix based convolutional neural network (AR-SMCNN) model, which is able to capture comprehensive hierarchical information utilizing the advantages of both RNN and CNN. We use RNN to capture semantic-level correlation by its sequential modeling nature, and use an attention mechanism to keep track of the entities and relations simultaneously. Meanwhile, we use a similarity matrix based CNN with two-directions pooling to extract literal-level words interaction matching utilizing CNNs strength of modeling spatial correlation among data. Moreover, we have developed a new heuristic extension method for entity detection, which significantly decreases the effect of noise. Our method has outperformed the state-of-the-arts on SimpleQuestion benchmark in both accuracy and efficiency.

cs.CL↗

arXiv2018

Second Chern class of Fano manifolds and anti-canonical geometry

Let $X$ be a Fano manifold of Picard number one. We establish a lower bound for the second Chern class of $X$ in terms of its index and degree. As an application, if $Y$ is a $n$-dimensional Fano manifold with $-K_Y=(n-3)H$ for some ample divisor $H$, we prove that $h^0(Y,H)\geq n-2$. Moreover, we show that the rational map defined by $\vert mH\vert$ is birational for $m\geq 5$, and the linear system $\vert mH\vert$ is basepoint free for $m\geq 7$. As a by-product, the pluri-anti-canonical systems of singular weak Fano varieties of dimension at most $4$ are also investigated.

math.AG↗

arXiv2018

Heterogeneous-Reliability Memory: Exploiting Application-Level Memory Error Tolerance

This paper summarizes our work on characterizing application memory error vulnerability to optimize datacenter cost via Heterogeneous-Reliability Memory (HRM), which was published in DSN 2014, and examines the work's significance and future potential. Memory devices represent a key component of datacenter total cost of ownership (TCO), and techniques used to reduce errors that occur on these devices increase this cost. Existing approaches to providing reliability for memory devices pessimistically treat all data as equally vulnerable to memory errors. Our key insight is that there exists a diverse spectrum of tolerance to memory errors in new data-intensive applications, and that traditional one-size-fits-all memory reliability techniques are inefficient in terms of cost. This presents an opportunity to greatly reduce server hardware cost by provisioning the right amount of memory reliability for different applications. Toward this end, in our DSN 2014 paper, we make three main contributions to enable highly-reliable servers at low datacenter cost. First, we develop a new methodology to quantify the tolerance of applications to memory errors. Second, using our methodology, we perform a case study of three new data-intensive workloads (an interactive web search application, an in-memory key--value store, and a graph mining framework) to identify new insights into the nature of application memory error vulnerability. Third, based on our insights, we propose several new hardware/software heterogeneous-reliability memory system designs to lower datacenter cost while achieving high reliability and discuss their trade-offs. We show that our new techniques can reduce server hardware cost by 4.7% while achieving 99.90% single server availability.

cs.DC↗

arXiv2018

MLE-induced Likelihood for Markov Random Fields

Due to the intractable partition function, the exact likelihood function for a Markov random field (MRF), in many situations, can only be approximated. Major approximation approaches include pseudolikelihood and Laplace approximation. In this paper, we propose a novel way of approximating the likelihood function through first approximating the marginal likelihood functions of individual parameters and then reconstructing the joint likelihood function from these marginal likelihood functions. For approximating the marginal likelihood functions, we derive a particular likelihood function from a modified scenario of coin tossing which is useful for capturing how one parameter interacts with the remaining parameters in the likelihood function. For reconstructing the joint likelihood function, we use an appropriate copula to link up these marginal likelihood functions. Numerical investigation suggests the superior performance of our approach. Especially as the size of the MRF increases, both the numerical performance and the computational cost of our approach remain consistently satisfactory, whereas Laplace approximation deteriorates and pseudolikelihood becomes computationally unbearable.

cs.LG↗

arXiv2018

Mixed Spectral Element Method for the Waveguide Problem with Bloch Periodic Boundary

The mixed spectral element method (MSEM) is applied to solve the waveguide problem with Bloch periodic boundary condition (BPBC). Based on the BPBC for the original Helmholtz equation and the periodic boundary condition (PBC) for the equivalent but modified Helmholtz equation, two equivalent mixed variational formulations are applied for the MSEM. Unlike the traditional finite element method and SEM, both these mixed SEM schemes are completely free of spurious modes because of their use of the Gauss' law and the curl-conforming vector basis functions structured by the Gauss-Legendre-Lobatto (GLL) points. A simple implementation method is used to deal with the BPBC and the PBC for the mixed variational formulations so that both schemes can save computational costs over the traditional methods. Several numerical results are also provided to verify that both schemes are free of spurious modes and have high accuracy with the propagation constants.

physics.comp-ph↗

arXiv2018

Doubled Shapiro Steps in a Topological Josephson Junction

We study the transport properties of a superconductor-quantum spin Hall insulator-superconductor (S-QSHI-S) hybrid system in the presence of a microwave radiation. Instead of adiabatic analysis or using the resistively shunted junction model, we start from the microscopic Hamiltonian and calculate the DC current directly with the help of the non-equilibrium Green's Functions method. The numerical results show that (i) the I-V curves of background current due to multiple Andreev reflections (MAR) exhibit a different structure with that in the conventional junctions, (ii) all Shapiro steps are visible and appear one by one at high frequency, while at low frequency, the steps evolve exactly as the Bessel functions and the odd steps are completely suppressed, implying a fractional Josephson effect.

cond-mat.mes-hall↗

arXiv2018

The Role of Interaction in the Pairing of Two Spin-orbit Coupled Fermions

We investigate the role of a repulsive s-wave interaction in the two-body problem in the presence of spin orbit couplings, motivated by current interests in exploring exotic superfluid phases in spin-orbit coupled Fermi gases. For weak spin orbit coupling where the density of states is not significantly altered, we analytically show that the high-energy states become more important in determining the binding energy when the interaction strength decreases. Consequently, tuning the interaction gives rise to a rich ground state behavior, including a zigzag of the ground state momentum or inducing transitions among the meta-stable states. By exactly solving the two-body problem for a spin-orbit coupled Fermi mixture, we demonstrate that our analysis can also apply to the case when the density of states is significantly modified by the spin-orbit coupling. Our findings pave the way for understanding and controlling the paring of fermions in the presence of spin orbit couplings.

cond-mat.quant-gas↗

arXiv2018

Simultaneous 13 cm/3 cm single-pulse observations of PSR B0329+54

We have investigated the mode-changing properties of PSR B0329+54 using 31 epochs of simultaneous 13 cm/3 cm single-pulse observations obtained with Shanghai Tian Ma 65 m telescope. The pulsar was found in the abnormal emission mode 17 times, accounting for ~13% of the 41.6 hours total observation time. Single pulse analyses indicate that mode changes took place simultaneously at 13 cm/3 cm within a few rotational periods. We detected occasional bright and narrow pulses whose peak flux densities were 10 times higher than that of the integrated profile in both bands. At 3 cm, about 0.66% and 0.27% of single pulses were bright in the normal mode and abnormal mode respectively, but at 13 cm the occurrence rate was only about 0.007%. We divided the pulsar radiation window into three components (C1, C2 and C3) corresponding to the main peaks of the integrated profile. The bright pulses preferentially occurred at pulse phases corresponding to the peaks of C2 and C3. Fluctuation spectra showed that C2 had excess red noise in the normal mode, but broad quasi-periodic features with central frequencies around 0.12 cycles/period in the abnormal mode. At 3 cm, C3 had a stronger quasi-periodic modulation centered around 0.06 cycles/period in the abnormal mode. Although there were some asymmetries in the two-dimensional fluctuation spectra, we found no clear evidence for systematic subpulse drifting. Consistent with previous low-frequency observations, we found a very low nulling probability for B0329+54 with upper limits of 0.13% and 1.68% at 13 cm/3 cm respectively.

astro-ph.HE↗

arXiv2018

Obstacle evasion in free-space optical communications utilizing Airy beams

A high speed free-space optical communication system capable of self-bending signal transmission around line-of-sight obstacles is proposed and demonstrated. Airy beams are generated and controlled to achieve different propagating trajectories, and the signal transmission characteristics of these beams around the obstacle are investigated. Our results confirm that, by optimising their ballistic trajectories, Airy beams are able to bypass obstacles with more signal energy and thus improve the communication performance compared with normal Gaussian beams.

physics.app-ph↗