arXiv subjects

Chang Liu

Publications and source records attributed to Chang Liu.

At least 1,099 records · Page 61Linked to original sources

arXiv2018

Statistical Analysis of Torus and Kink Instabilities in Solar Eruptions

A recent laboratory experiment of ideal magnetohydrodynamic (MHD) instabilities reveals four distinct eruption regimes readily distinguished by the torus instability (TI) and helical kink instability (KI) parameters \citep{Myers2015}. To establish its observational counterpart, we collect 38 solar flares (stronger than GOES class M5 in general) that took place within 45$^{\circ}$ of disk center during 2011$-$2017, 26 of which are associated with a halo or partial halo coronal mass ejection (CME) (i.e., ejective events), while the others are CMEless (i.e., confined events). This is a complete sample of solar events satisfying our selection criteria detailed in the paper. For each event, we calculate decay index $n$ of the potential strapping field above the magnetic flux rope (MFR) in and around the flaring magnetic polarity inversion line (a TI parameter), and the unsigned twist number $T_w$ of the non-linear force-free (NLFF) field lines forming the same MFR (a KI parameter). We then construct a $n-T_w$ diagram to investigate how the eruptiveness depends on these parameters. We find: (1) $T_w$ appears to play little role in discriminating between confined and ejective events; (2) the events with $n\gtrsim0.8$ are all ejective and all confined events have $n\lesssim0.8$. However, $n\gtrsim0.8$ is not a necessary condition for eruption, because some events with $n\lesssim0.8$ also erupted. In addition, we investigate the MFR's geometrical parameters, apex height and distance between footpoints, as a possible factor for the eruptiveness. We briefly discuss the difference of the present result for solar eruptions with that of the laboratory result in terms of the role played by magnetic reconnection.

astro-ph.SR↗

arXiv2018

Neural Network-based Graph Embedding for Cross-Platform Binary Code Similarity Detection

The problem of cross-platform binary code similarity detection aims at detecting whether two binary functions coming from different platforms are similar or not. It has many security applications, including plagiarism detection, malware detection, vulnerability search, etc. Existing approaches rely on approximate graph matching algorithms, which are inevitably slow and sometimes inaccurate, and hard to adapt to a new task. To address these issues, in this work, we propose a novel neural network-based approach to compute the embedding, i.e., a numeric vector, based on the control flow graph of each binary function, then the similarity detection can be done efficiently by measuring the distance between the embeddings for two functions. We implement a prototype called Gemini. Our extensive evaluation shows that Gemini outperforms the state-of-the-art approaches by large margins with respect to similarity detection accuracy. Further, Gemini can speed up prior art's embedding generation time by 3 to 4 orders of magnitude and reduce the required training time from more than 1 week down to 30 minutes to 10 hours. Our real world case studies demonstrate that Gemini can identify significantly more vulnerable firmware images than the state-of-the-art, i.e., Genius. Our research showcases a successful application of deep learning on computer security problems.

cs.CR↗

arXiv2018

Linear Span Network for Object Skeleton Detection

Robust object skeleton detection requires to explore rich representative visual features and effective feature fusion strategies. In this paper, we first re-visit the implementation of HED, the essential principle of which can be ideally described with a linear reconstruction model. Hinted by this, we formalize a Linear Span framework, and propose Linear Span Network (LSN) modified by Linear Span Units (LSUs), which minimize the reconstruction error of convolutional network. LSN further utilizes subspace linear span beside the feature linear span to increase the independence of convolutional features and the efficiency of feature integration, which enlarges the capability of fitting complex ground-truth. As a result, LSN can effectively suppress the cluttered backgrounds and reconstruct object skeletons. Experimental results validate the state-of-the-art performance of the proposed LSN.

cs.CV↗

arXiv2018

An efficient Moving Morphable Component (MMC)-based approach for multi-resolution topology optimization

In the present work, a highly efficient Moving Morphable Component (MMC) based approach for multi-resolution topology optimization is proposed. In this approach, high-resolution optimization results can be obtained with much less number of degrees of freedoms (DOFs) and design variables since the finite element analysis model and the design optimization model are totally decoupled in the MMC-based problem formulation. This is achieved by introducing super-elements for structural response analysis and adopting a domain decomposition strategy to preserve the topology complexity of optimized structures. Both two-and three-dimensional numerical results demonstrate that substantial computational efforts can be saved with use of the proposed approach.

cs.CE↗

arXiv2018

Message Passing Stein Variational Gradient Descent

Stein variational gradient descent (SVGD) is a recently proposed particle-based Bayesian inference method, which has attracted a lot of interest due to its remarkable approximation ability and particle efficiency compared to traditional variational inference and Markov Chain Monte Carlo methods. However, we observed that particles of SVGD tend to collapse to modes of the target distribution, and this particle degeneracy phenomenon becomes more severe with higher dimensions. Our theoretical analysis finds out that there exists a negative correlation between the dimensionality and the repulsive force of SVGD which should be blamed for this phenomenon. We propose Message Passing SVGD (MP-SVGD) to solve this problem. By leveraging the conditional independence structure of probabilistic graphical models (PGMs), MP-SVGD converts the original high-dimensional global inference problem into a set of local ones over the Markov blanket with lower dimensions. Experimental results show its advantages of preventing vanishing repulsive force in high-dimensional space over SVGD, and its particle efficiency and approximation flexibility over other inference methods on graphical models.

stat.ML↗

arXiv2018

Cell Size Effect on Computational Fluid Dynamics: The Limitation Principle for Flow Simulation

For theoretical gas dynamics, the flow regimes are classified according to the Knudsen number. For computational fluid dynamics (CFD), the numerical flow field is the projection of the physical flow field onto the discrete space and time, which is related to the cell Knudsen number. The real representable flow regimes are controlled by these two parameters. According to the values of Knudsen number and cell Knudsen number, we study the classification of the numerical flow regimes. In the process of mesh refinement, the numerical experiments show the change of numerical flow regime from continuum, to near-continuum, and to non-equilibrium one. The change of flow regime with different cell resolution is the limitation principle for the numerical simulation, which is the best a multiscale method can do. In other words, we should have changeable numerical governing equations in different mesh size scale, and they are coupled with the traditional physical equations in different scales, such as the Navier-Stokes and Boltzmann. Under the multiscale modeling, a mesh refinement is a process in resolving the flow physics in different scale. The verification and validation (V$\&$V) need include the physical modeling mechanism in the mesh refinement process. The traditional idea of mesh refinement for targeting a fixed partial differential equation cannot achieve the final goal of computation, which is to recover the flow physics as truthfully as possible under the limitation of the cell resolution.

physics.comp-ph↗

arXiv2018

Origami Inspired Reconfigurable Antenna for Wireless Communication Systems

This paper presents the design, fabrication, and experimental validation of an origami-inspired reconfigurable antenna. The proposed antenna can operate as a monopole or an inverted-L antenna, by changing its configuration. Doing so changes its operational frequency, principal radiation mode, and directivity. Measurements show that the antenna is able to change its resonance frequency from 750 MHz to 920 MHz---equivalent to 22.6\% frequency shift. Simulations are carried out to compare the radiation characteristics of the antenna (gain and radiation pattern) in both configurations. The results validate that the antenna has a reconfigurable bandwidth, which enables a better channel characterization for RF systems.

physics.app-ph↗

arXiv2018

Multi-task Learning for Macromolecule Classification, Segmentation and Coarse Structural Recovery in Cryo-Tomography

Cellular Electron Cryo-Tomography (CECT) is a powerful 3D imaging tool for studying the native structure and organization of macromolecules inside single cells. For systematic recognition and recovery of macromolecular structures captured by CECT, methods for several important tasks such as subtomogram classification and semantic segmentation have been developed. However, the recognition and recovery of macromolecular structures are still very difficult due to high molecular structural diversity, crowding molecular environment, and the imaging limitations of CECT. In this paper, we propose a novel multi-task 3D convolutional neural network model for simultaneous classification, segmentation, and coarse structural recovery of macromolecules of interest in subtomograms. In our model, the learned image features of one task are shared and thereby mutually reinforce the learning of other tasks. Evaluated on realistically simulated and experimental CECT data, our multi-task learning model outperformed all single-task learning methods for classification and segmentation. In addition, we demonstrate that our model can generalize to discover, segment and recover novel structures that do not exist in the training data.

cs.CV↗

arXiv2018

Curriculum Adversarial Training

Recently, deep learning has been applied to many security-sensitive applications, such as facial authentication. The existence of adversarial examples hinders such applications. The state-of-the-art result on defense shows that adversarial training can be applied to train a robust model on MNIST against adversarial examples; but it fails to achieve a high empirical worst-case accuracy on a more complex task, such as CIFAR-10 and SVHN. In our work, we propose curriculum adversarial training (CAT) to resolve this issue. The basic idea is to develop a curriculum of adversarial examples generated by attacks with a wide range of strengths. With two techniques to mitigate the forgetting and the generalization issues, we demonstrate that CAT can improve the prior art's empirical worst-case accuracy by a large margin of 25% on CIFAR-10 and 35% on SVHN. At the same, the model's performance on non-adversarial inputs is comparable to the state-of-the-art models.

cs.LG↗

arXiv2018

Formation of Large Scale Coronal Loops Interconnecting Two Active Regions Through Gradual Magnetic Reconnection and Associated Heating Process

Coronal loops interconnecting two active regions, called as interconnecting loops (ILs), are prominent large-scale structures in the solar atmosphere. They carry a significant amount of magnetic flux, therefore are considered to be an important element of the solar dynamo process. Earlier observations show that eruptions of ILs are an important source of CMEs. It is generally believed that ILs are formed through magnetic reconnection in the high corona (>150-200"), and several scenarios have been proposed to explain their brightening in soft X-rays (SXRs). Yet, the detailed IL formation process has not been fully explored and the associated energy release in the corona still remains unresolved. Here we report the complete formation process of a set of ILs connecting two nearby active regions, with successive observations by STEREO-A on the far side of the Sun and SDO and Hinode on the Earth side. We conclude that ILs are formed by gradual reconnection high in the corona, in line with earlier postulations. In addition, we show evidence supporting that ILs become brightened in SXRs and EUVs through heating at or close to the reconnection site in the corona (i.e., through direct heating process of reconnection), a process that has been largely overlooked in earlier studies on ILs

astro-ph.SR↗

arXiv2018

Direct evidence of ferromagnetism in a quantum anomalous Hall system

Quantum anomalous Hall (QAH) systems are of great fundamental interest and potential application because of their dissipationless conduction without the need for external magnetic field. The QAH effect has been realized in magnetically doped topological insulator thin films. However, full quantization requires extremely low temperature ($T< 50\,$mK) in the initial works, though it has been significantly improved with modulation doping or co-doping of magnetic elements. Improved ferromagnetism has been shown in these thin films, yet direct evidence of long-range ferromagnetic order is lacking. Herein, we present direct visualization of long-range ferromagnetic order in thin films of Cr and V co-doped (Bi,Sb)$_2$Te$_3$ using low-temperature magnetic force microscopy with $\textit{in-situ}$ transport. The magnetization reversal process reveals typical ferromagnetic domain behavior, i.e., domain nucleation and possibly domain wall propagation, in contrast to much weaker magnetic signals observed in the end members, possibly due to superparamagnetic behavior. The observed long-range ferromagnetic order resolves one of the major challenges in QAH systems, and paves the way to high-temperature dissipationless conduction by exploring magnetic topological insulators.

cond-mat.mtrl-sci↗

arXiv2018

Tunable ferroelectricity and anisotropic electric transport in monolayer $β$-GeSe

Low-dimensional ferroelectricity has attracted tremendous attention due to its huge potential in device applications. Here, based on first-principles calculations, we predict the existence of spontaneous in-plane electrical polarization and ferroelectricity in monolayer $β$-GeSe, a polymorph of GeSe with a boat conformation newly synthesized in experiment. The magnitude of the polarization is about $0.16$ nC/m, which is comparable to that of monolayer SnTe studied in recent experiment, and the intrinsic Curie temperature is estimated to be above 200 K. Interestingly, owing to its puckered structure, the physical properties of $β$-GeSe can be easily controlled by strain. The Curie temperature can be raised above room temperature by applying a ($\sim 1\%$) tensile strain, and the magnitude of polarization can be largely increased by strains in either armchair or zigzag directions. Furthermore, we find that for the case with electron doping, applying strain can readily tune the anisotropic electric transport with the preferred conducting direction rotated by $90^\circ$, which is connected with a strain-induced Lifshitz transition. The ratio between the effective masses along the two in-plane directions can undergo a dramatic change of two orders of magnitude even by a 2% strain. Our result reveals monolayer $β$-GeSe as a promising platform for exploring ferroelectricity in two-dimensions and for nanoscale mechano-electronic device applications.

cond-mat.mtrl-sci↗

arXiv2018

A Machine Learning Approach To Prevent Malicious Calls Over Telephony Networks

Malicious calls, i.e., telephony spams and scams, have been a long-standing challenging issue that causes billions of dollars of annual financial loss worldwide. This work presents the first machine learning-based solution without relying on any particular assumptions on the underlying telephony network infrastructures. The main challenge of this decade-long problem is that it is unclear how to construct effective features without the access to the telephony networks' infrastructures. We solve this problem by combining several innovations. We first develop a TouchPal user interface on top of a mobile App to allow users tagging malicious calls. This allows us to maintain a large-scale call log database. We then conduct a measurement study over three months of call logs, including 9 billion records. We design 29 features based on the results, so that machine learning algorithms can be used to predict malicious calls. We extensively evaluate different state-of-the-art machine learning approaches using the proposed features, and the results show that the best approach can reduce up to 90% unblocked malicious calls while maintaining a precision over 99.99% on the benign call traffic. The results also show the models are efficient to implement without incurring a significant latency overhead. We also conduct ablation analysis, which reveals that using 10 out of the 29 features can reach a performance comparable to using all features.

cs.CR↗

arXiv2018

Fooling Vision and Language Models Despite Localization and Attention Mechanism

Adversarial attacks are known to succeed on classifiers, but it has been an open question whether more complex vision systems are vulnerable. In this paper, we study adversarial examples for vision and language models, which incorporate natural language understanding and complex structures such as attention, localization, and modular architectures. In particular, we investigate attacks on a dense captioning model and on two visual question answering (VQA) models. Our evaluation shows that we can generate adversarial examples with a high success rate (i.e., > 90%) for these models. Our work sheds new light on understanding adversarial attacks on vision systems which have a language component and shows that attention, bounding box localization, and compositional internal structures are vulnerable to adversarial attacks. These observations will inform future work towards building effective defenses.

cs.AI↗

arXiv2018

The role of magnetic moment in the collisionless pitch-angle scattering of runaway electrons

Recently, the validity of the guiding-center approach to model relativistic runaway electrons in tokamaks has been challenged by full-orbit simulations that demonstrate the breakdown of the standard magnetic moment conservation. In this paper, we derive a new expression for the magnetic moment of relativistic runaway electrons, which is conserved significantly better than the standard one. The new result includes one of the second-order corrections in the standard guiding-center theory which, in case of runaway electrons with $p_{\parallel}\gg p_{\perp}$, can peculiarly be of the same order as the lowest-order term. The better conservation of the new magnetic moment also explains the collisionless pitch-angle-scattering effect observed in full-orbit simulations since it allows momentum transfer between the perpendicular and parallel directions when the runaway electron is accelerated by an electric field. While the derivation of the second-order correction to the magnetic moment in general case would require the full extent of the relativistic second-order guiding-center theory, we exploit the Lie-perturbation method at the limit $p_{\parallel}\gg p_{\perp}$ which simplifies the computations significantly. Consequently, we present the corresponding guiding-center equations applicable to the highly relativistic runaway electrons.

physics.plasm-ph↗

arXiv2018

Systematic investigation of methods for multiple freeform optimization in multi-lens imaging systems

With the development in freeform technology, it has now become more and more feasible to use freeform surfaces in real system designs. While the freeform surfaces helping optical designers achieve more and more challenging system features, the methods for multiple freeform implementations are still underdeveloped. We therefore investigate strategies to use freeform surfaces properly in imaging optical systems with one Scheimpflug system and one lithographic system. Based on the studies of the influences of the freeform normalization radius, freeform order and system eccentricity, the methods of determining the optimal location for implementing one freeform surface are discussed. Different optimization strategies to optimize two freeform surfaces are discussed to compare their resulting influences on the system performance. On top of that, ways to implement more than one freeform surface in the optical system is also investigated. In the end, a workflow is presented as guidance for implementing multiple freeform surfaces with respect to system aberration constitutions.

physics.app-ph↗

arXiv2018

Performance Analysis for Practical Unmanned Aerial Vehicle Networks with LoS/NLoS Transmissions

In this paper, we provide a performance analysis for practical unmanned aerial vehicle (UAV)-enabled networks. By considering both line-of-sight (LoS) and non-line-of-sight (NLoS) transmissions between aerial base stations (BSs) and ground users, the coverage probability and the area spectral efficiency (ASE) are derived. Considering that there is no consensus on the path loss model for studying UAVs in the literature, in this paper, three path loss models, i.e., high-altitude model, low-altitude model and ultra-low-altitude model, are investigated and compared. Moreover, the lower bound of the network performance is obtained assuming that UAVs are hovering randomly according to homogeneous Poisson point process (HPPP), while the upper bound is derived assuming that UAVs can instantaneously move to the positions directly overhead ground users. From our analytical and simulation results for a practical UAV height of 50 meters, we find that the network performance of the high-altitude model and the low-altitude model exhibit similar trends, while that of the ultra-low-altitude model deviates significantly from the above two models. In addition, the optimal density of UAVs to maximize the coverage probability performance has also been investigated.

cs.NI↗

arXiv2018

The effects of kinetic instabilities on the electron cyclotron emission from runaway electrons

In this paper we show that the kinetic instabilities associated with runaway electron beams play an essential role for the production of high-level non-thermal electron-cyclotron-emission (ECE) radiation. Most of the non-thermal ECE comes from runaway electrons in the low-energy regime with large pitch angle, which are strongly scattered by the excited whistler waves. The power of ECE from runaway electrons is obtained using a synthetic diagnostic model based on the reciprocity method. The electron distribution function is calculated using a kinetic simulation model including the whistler wave instabilities and the quasilinear diffusion effects. Simulations based on DIII-D low-density discharge reproduces the rapid growth of the ECE signals observed in DIII-D experiments. Unlike the thermal ECE where radiation for a certain frequency is strongly localized inside the resonance region, the non-thermal ECE radiation from runaway electrons is nonlocal, and the emission-absorption ratio is higher than that of thermal electrons. The runaway electron tail is more significant for ECE with higher frequencies, and the ECE spectrum becomes flatter as RE population grows. The nonlinear behavior of the kinetic instabilities is illustrated in the osculations of the ECE waves. The good agreement with the DIII-D experimental observations after including the kinetic instabilities clearly illustrate the significance of the scattering effects from wave-particle interactions, which can also be important for runaway electrons produced in disruptions.

physics.plasm-ph↗