arXiv subjects

Zhe Wang

Publications and source records attributed to Zhe Wang.

At least 577 records · Page 32Linked to original sources

arXiv2019

FocusNet: Imbalanced Large and Small Organ Segmentation with an End-to-End Deep Neural Network for Head and Neck CT Images

In this paper, we propose an end-to-end deep neural network for solving the problem of imbalanced large and small organ segmentation in head and neck (HaN) CT images. To conduct radiotherapy planning for nasopharyngeal cancer, more than 10 organs-at-risk (normal organs) need to be precisely segmented in advance. However, the size ratio between large and small organs in the head could reach hundreds. Directly using such imbalanced organ annotations to train deep neural networks generally leads to inaccurate small-organ label maps. We propose a novel end-to-end deep neural network to solve this challenging problem by automatically locating, ROI-pooling, and segmenting small organs with specifically designed small-organ sub-networks while maintaining the accuracy of large organ segmentation. A strong main network with densely connected atrous spatial pyramid pooling and squeeze-and-excitation modules is used for segmenting large organs, where large organs' label maps are directly output. For small organs, their probabilistic locations instead of label maps are estimated by the main network. High-resolution and multi-scale feature volumes for each small organ are ROI-pooled according to their locations and are fed into small-organ networks for accurate segmenting small organs. Our proposed network is extensively tested on both collected real data and the \emph{MICCAI Head and Neck Auto Segmentation Challenge 2015} dataset, and shows superior performance compared with state-of-the-art segmentation methods.

eess.IV↗

arXiv2019

NeuralCubes: Deep Representations for Visual Data Exploration

Visual exploration of large multidimensional datasets has seen tremendous progress in recent years, allowing users to express rich data queries that produce informative visual summaries, all in real time. Techniques based on data cubes are some of the most promising approaches. However, these techniques usually require a large memory footprint for large datasets. To tackle this problem, we present NeuralCubes: neural networks that predict results for aggregate queries, similar to data cubes. NeuralCubes learns a function that takes as input a given query, for instance, a geographic region and temporal interval, and outputs the result of the query. The learned function serves as a real-time, low-memory approximator for aggregation queries. NeuralCubes models are small enough to be sent to the client side (e.g. the web browser for a web-based application) for evaluation, enabling data exploration of large datasets without database/network connection. We demonstrate the effectiveness of NeuralCubes through extensive experiments on a variety of datasets and discuss how NeuralCubes opens up opportunities for new types of visualization and interaction.

cs.DB↗

arXiv2019

Cubic Regularization with Momentum for Nonconvex Optimization

Momentum is a popular technique to accelerate the convergence in practical training, and its impact on convergence guarantee has been well-studied for first-order algorithms. However, such a successful acceleration technique has not yet been proposed for second-order algorithms in nonconvex optimization.In this paper, we apply the momentum scheme to cubic regularized (CR) Newton's method and explore the potential for acceleration. Our numerical experiments on various nonconvex optimization problems demonstrate that the momentum scheme can substantially facilitate the convergence of cubic regularization, and perform even better than the Nesterov's acceleration scheme for CR. Theoretically, we prove that CR under momentum achieves the best possible convergence rate to a second-order stationary point for nonconvex optimization. Moreover, we study the proposed algorithm for solving problems satisfying an error bound condition and establish a local quadratic convergence rate. Then, particularly for finite-sum problems, we show that the proposed algorithm can allow computational inexactness that reduces the overall sample complexity without degrading the convergence rate.

math.OC↗

arXiv2019

Model-free posterior inference on the area under the receiver operating characteristic curve

The area under the receiver operating characteristic curve (AUC) serves as a summary of a binary classifier's performance. Methods for estimating the AUC have been developed under a binormality assumption which restricts the distribution of the score produced by the classifier. However, this assumption introduces an infinite-dimensional nuisance parameter and can be inappropriate, especially in the context of machine learning. This motivates us to adopt a model-free Gibbs posterior distribution for the AUC. We present the asymptotic Gibbs posterior concentration rate, and a strategy for tuning the learning rate so that the corresponding credible intervals achieve the nominal frequentist coverage probability. Simulation experiments and a real data analysis demonstrate the Gibbs posterior's strong performance compared to existing methods based on a rank likelihood.

stat.ME↗

arXiv2019

Augmenting Model Robustness with Transformation-Invariant Attacks

The vulnerability of neural networks under adversarial attacks has raised serious concerns and motivated extensive research. It has been shown that both neural networks and adversarial attacks against them can be sensitive to input transformations such as linear translation and rotation, and that human vision, which is robust against adversarial attacks, is invariant to natural input transformations. Based on these, this paper tests the hypothesis that model robustness can be further improved when it is adversarially trained against transformed attacks and transformation-invariant attacks. Experiments on MNIST, CIFAR-10, and restricted ImageNet show that while transformations of attacks alone do not affect robustness, transformation-invariant attacks can improve model robustness by 2.5\% on MNIST, 3.7\% on CIFAR-10, and 1.1\% on restricted ImageNet. We discuss the intuition behind this phenomenon.

cs.CV↗

arXiv2019

Coordinate-wise descent methods for leading eigenvalue problem

Leading eigenvalue problems for large scale matrices arise in many applications. Coordinate-wise descent methods are considered in this work for such problems based on a reformulation of the leading eigenvalue problem as a non-convex optimization problem. The convergence of several coordinate-wise methods is analyzed and compared. Numerical examples of applications to quantum many-body problems demonstrate the efficiency and provide benchmarks of the proposed coordinate-wise descent methods.

math.NA↗

arXiv2019

Momentum Schemes with Stochastic Variance Reduction for Nonconvex Composite Optimization

Two new stochastic variance-reduced algorithms named SARAH and SPIDER have been recently proposed, and SPIDER has been shown to achieve a near-optimal gradient oracle complexity for nonconvex optimization. However, the theoretical advantage of SPIDER does not lead to substantial improvement of practical performance over SVRG. To address this issue, momentum technique can be a good candidate to improve the performance of SPIDER. However, existing momentum schemes used in variance-reduced algorithms are designed specifically for convex optimization, and are not applicable to nonconvex scenarios. In this paper, we develop novel momentum schemes with flexible coefficient settings to accelerate SPIDER for nonconvex and nonsmooth composite optimization, and show that the resulting algorithms achieve the near-optimal gradient oracle complexity for achieving a generalized first-order stationary condition. Furthermore, we generalize our algorithm to online nonconvex and nonsmooth optimization, and establish an oracle complexity result that matches the state-of-the-art. Our extensive experiments demonstrate the superior performance of our proposed algorithm over other stochastic variance-reduced algorithms.

math.OC↗

arXiv2019

Training Multi-Task Adversarial Network for Extracting Noise-Robust Speaker Embedding

Under noisy environments, to achieve the robust performance of speaker recognition is still a challenging task. Motivated by the promising performance of multi-task training in a variety of image processing tasks, we explore the potential of multi-task adversarial training for learning a noise-robust speaker embedding. In this paper we present a novel framework which consists of three components: an encoder that extracts noise-robust speaker embedding; a classifier that classifies the speakers; a discriminator that discriminates the noise type of the speaker embedding. Besides, we propose a training strategy using the training accuracy as an indicator to stabilize the multi-class adversarial optimization process. We conduct our experiments on the English and Mandarin corpus and the experimental results demonstrate that our proposed multi-task adversarial training method could greatly outperform the other methods without adversarial training in noisy environments. Furthermore, experiments indicate that our method is also able to improve the speaker verification performance the clean condition.

cs.SD↗

arXiv2019

Coordinate descent full configuration interaction

We develop an efficient algorithm, coordinate descent FCI (CDFCI), for the electronic structure ground state calculation in the configuration interaction framework. CDFCI solves an unconstrained non-convex optimization problem, which is a reformulation of the full configuration interaction eigenvalue problem, via an adaptive coordinate descent method with a deterministic compression strategy. CDFCI captures and updates appreciative determinants with different frequencies proportional to their importance. We show that CDFCI produces accurate variational energy for both static and dynamic correlation by benchmarking the binding curve of nitrogen dimer in the cc-pVDZ basis with $10^{-3}$ mHa accuracy. We also demonstrate the efficiency and accuracy of CDFCI for strongly correlated chromium dimer in the Ahlrichs VDZ basis and produces state-of-the-art variational energy.

physics.chem-ph↗

arXiv2019

Quantum Critical Dynamics of a Heisenberg-Ising Chain in a Longitudinal Field: Many-Body Strings versus Fractional Excitations

We report a high-resolution terahertz spectroscopic study of quantum spin dynamics in the antiferromagnetic Heisenberg-Ising spin-chain compound BaCo$_2$V$_2$O$_8$ as a function of temperature and longitudinal magnetic field. Confined spinon excitations are observed in an antiferromagnetic phase below $T_N\simeq 5.5$ K. In a field-induced gapless phase above $B_c=3.8$ T, we identify many-body string excitations as well as low-energy fractional psinon/antipsinon excitations by comparing to Bethe-Ansatz calculations. In the vicinity of $B_c$, the high-energy string excitations are found to be dynamically dominant over the fractional excitations.

cond-mat.str-el↗

arXiv2019

THz excitations in α-RuCl3: Majorana fermions, rigid-plane shear and compression modes

Spin liquids may host emergent quasiparticles, collective excitations of the spin degrees of freedom with characteristic features of Majorana fermions, which experimentally are detectable by broad excitation continua due to spin fractionalization. The latter is predicted for the Kitaev spin liquid, an exactly solvable model with bond-dependent interactions on a two-dimensional honeycomb lattice. Here we report on detailed THz experiments in α-RuCl3, identifying these characteristic fingerprints of Majorana fermions. The continuum intensity decreases and finally vanishes on increasing temperature. It partly overlaps with phonon modes representing characteristic sliding and compression modes of the van der Waals bonded molecular layers.

cond-mat.str-el↗

arXiv2019

Slow Liquid Scintillator Candidates for MeV-scale Neutrino Experiments

Slow liquid scintillator Cherenkov detectors have been proposed as part of several future neutrino experiments because they can provide both directionality and energy measurements. This feature is expected to enhance the sensitivities for MeV-scale neutrino physics, including solar physics, the search for supernova relic neutrino, and the study of geo-sciences. In this study, the characteristics of a slow liquid scintillator were investigated, along with the light yields and decay time constants for various combinations of linear alkylbenzene (LAB), 2,5-diphenyloxazole (PPO), and 1,4-bis (2-methylstyryl)-benzene (bis-MSB). The results of our study indicated that LAB with 0.07 g/L of PPO and 13 mg/L of bis-MSB was the best candidate for an effective separation between Cherenkov and scintillation lights with a reasonably high light yield.

physics.ins-det↗

arXiv2019

A^2-Net: Molecular Structure Estimation from Cryo-EM Density Volumes

Constructing of molecular structural models from Cryo-Electron Microscopy (Cryo-EM) density volumes is the critical last step of structure determination by Cryo-EM technologies. Methods have evolved from manual construction by structural biologists to perform 6D translation-rotation searching, which is extremely compute-intensive. In this paper, we propose a learning-based method and formulate this problem as a vision-inspired 3D detection and pose estimation task. We develop a deep learning framework for amino acid determination in a 3D Cryo-EM density volume. We also design a sequence-guided Monte Carlo Tree Search (MCTS) to thread over the candidate amino acids to form the molecular structure. This framework achieves 91% coverage on our newly proposed dataset and takes only a few minutes for a typical structure with a thousand amino acids. Our method is hundreds of times faster and several times more accurate than existing automated solutions without any human intervention.

cs.LG↗

arXiv2019

Three-dimensional laser printing of macro-scale glass objects at a micro-scale resolution

Three-dimensional (3D) printing has allowed for production of geometrically complex 3D objects with extreme flexibility, which is currently undergoing rapid expansions in terms of materials, functionalities, as well as areas of application. When attempting to print 3D microstructures in glass, femtosecond laser induced chemical etching (FLICE) has proved itself a powerful approach. Here, we demonstrate fabrication of macro-scale 3D glass objects of large heights up to ~3.8 cm with a well-balanced (i.e., lateral vs longitudinal) spatial resolution of ~20 μm. The remarkable accomplishment is achieved by revealing an unexplored regime in the interaction of ultrafast laser pulses with fused silica which results in aberration-free focusing of the laser pulses deeply inside fused silica.

physics.app-ph↗

arXiv2019

Terahertz oscilloscope for recording time information of ultrashort electron beams

We propose and demonstrate a Terahertz (THz) oscilloscope for recording time information of an ultrashort electron beam. By injecting a laser-driven THz pulse with circular polarization into a dielectric tube, the electron beam is swept helically such that the time information is uniformly encoded into the angular distribution that allows one to characterize both the temporal profile and timing jitter of an electron beam. The dynamic range of the measurement in such a configuration is significantly increased compared to deflection with a linearly polarized THz pulse. With this THz oscilloscope, nearly 50-fold longitudinal compression of a relativistic electron beam to about 15 fs (rms) is directly visualized with its arrival time determined with 3 fs accuracy. This technique bridges the gap between streaking of photoelectrons with optical lasers and deflection of relativistic electron beams with radio-frequency deflectors, and should have wide applications in many ultrashort electron beam based facilities.

physics.acc-ph↗

arXiv2019

Probing magnetism in 2D materials at the nanoscale with single spin microscopy

The recent discovery of ferromagnetism in 2D van der Waals (vdw) crystals has generated widespread interest, owing to their potential for fundamental and applied research. Advancing the understanding and applications of vdw magnets requires methods to quantitatively probe their magnetic properties on the nanoscale. Here, we report the study of atomically thin crystals of the vdw magnet CrI$_3$ down to individual monolayers using scanning single-spin magnetometry, and demonstrate quantitative, nanoscale imaging of magnetisation, localised defects and magnetic domains. We determine the magnetisation of CrI$_3$ monolayers to be $\approx16~μ_B/$nm$^2$ and find comparable values in samples with odd numbers of layers, whereas the magnetisation vanishes when the number of layers is even. We also establish that this inscrutable even-odd effect is intimately connected to the material structure, and that structural modifications can induce switching between ferro- and anti-ferromagnetic interlayer ordering. Besides revealing new aspects of magnetism in atomically thin CrI$_3$ crystals, these results demonstrate the power of single-spin scanning magnetometry for the study of magnetism in 2D vdw magnets.

cond-mat.mes-hall↗

arXiv2018

A high-energy density antiferroelectric made by interfacial electrostatic engineering

Dielectric capacitors hold a tremendous advantage for energy storage due to their fast charge/discharge times and stability in comparison to batteries and supercapacitors. A key limitation to today's dielectric capacitors, however, is the low storage capacity of conventional dielectric materials. To mitigate this issue, antiferroelectric materials have been proposed, but relatively few families of antiferroelectric materials have been identified to date. Here, we propose a new design strategy for the construction of lead-free antiferroelectric materials using interfacial electrostatic engineering. We begin with a ferroelectric material with one of the highest known bulk polarizations, BiFeO3. We show that by confining atomically-precise thin layers of BiFeO3 in a dielectric matrix that we can induce a metastable antiferroelectric structure. Application of an electric field reversibly switches between this new phase and a ferroelectric state, in addition, tuning of the dielectric layer causes coexistence of the ferroelectric and antiferroelectric states. Precise engineering of the structure generates an antiferroelectric phase with energy storage comparable to that of the best lead-based materials. The use of electrostatic confinement provides a new pathway for the design of engineered antiferroelectric materials with large and potentially coupled responses.

cond-mat.mtrl-sci↗

arXiv2018

Quantitative imaging of the complexity in liquid bubbles' evolution reveals the dynamics of film retraction

The dynamics and stability of thin liquid films have fascinated scientists over many decades. Thin film flows are central to numerous areas of engineering, geophysics, and biophysics and occur over a wide range of length, velocity, and liquid properties scales. In spite of many significant developments in this area, we still lack appropriate quantitative experimental tools with the spatial and temporal resolution necessary for a comprehensive study of film evolution. We propose tackling this problem with a holographic technique that combines quantitative phase imaging with a custom setup designed to form and manipulate bubbles. The results, gathered on a model aqueous polymeric solution, provide an unparalleled insight into bubble dynamics through the combination of full-field thickness estimation, three-dimensional imaging, and fast acquisition time. The unprecedented level of detail offered by the proposed methodology will promote a deeper understanding of the underlying physics of thin film dynamics.

cond-mat.soft↗