arXiv subjects

Zheng Sun

Publications and source records attributed to Zheng Sun.

At least 91 records · Page 5Linked to original sources

arXiv2023

On a numerical artifact of solving shallow water equations with a discontinuous bottom: Analysis and a nontransonic fix

In this paper, we study a numerical artifact of solving the nonlinear shallow water equations with a discontinuous bottom topography. For various first-order schemes, the numerical solution of the momentum will form a spurious spike at the discontinuous points of the bottom, which should not exist in the exact solution. The height of the spike cannot be reduced even after the mesh is refined. For subsonic problems, this numerical artifact may cause the wrong convergence to a function far away from the exact solution. To explain the formation of the spurious spike, we perform a convergence analysis by proving a Lax--Wendroff type theorem. It is shown that the spurious spike is caused by the numerical viscosity in the computation of the water height at the discontinuous bottom. The height of the spike is proportional to the magnitude of the viscosity constant in the Lax--Friedrichs flux. Motivated by this conclusion, we propose a modified scheme by adopting the central flux at the bottom discontinuity in the equation of mass conservation, and show that this numerical artifact can be removed in many cases. For various numerical tests with nontransonic Riemann solutions, we observe that the modified scheme is able to retrieve the correct convergence.

math.NA↗

arXiv2023

Observation of two splitting processes in a partial filament eruption on the sun: the role of breakout reconnection

Partial filament eruptions have often been observed, however, the physical mechanisms that lead to filament splitting are not yet fully understood. In this study, we present a unique event of a partial filament eruption that undergoes two distinct splitting processes. The first process involves vertical splitting and is accompanied by brightenings inside the filament, which may result from internal magentic reconnection within the filament. Following the first splitting process, the filament is separated into an upper part and a lower part. Subsequently, the upper part undergoes a second splitting, which is accompanied by a coronal blowout jet. An extrapolation of the coronal magnetic field reveals a hyperbolic flux tube structure above the filament, indicating the occurrence of breakout reconnection that reduces the constraning field above. Consequently, the filament is lifted up, but at a nonuniform speed. The high-speed part reaches the breakout current sheet to generate the blowout jet, while the low-speed part falls back to the solar surface, resulting in the second splitting. In addition, continuous brightenings are observed along the flare ribbons, suggesting the occurrence of slipping reconnection process. This study presents, for the first time, the unambiguous observation of a two-stage filament splitting process, advancing our understanding of the complex dynamics of solar eruptions.

astro-ph.SR↗

arXiv2023

Robust Table Structure Recognition with Dynamic Queries Enhanced Detection Transformer

We present a new table structure recognition (TSR) approach, called TSRFormer, to robustly recognizing the structures of complex tables with geometrical distortions from various table images. Unlike previous methods, we formulate table separation line prediction as a line regression problem instead of an image segmentation problem and propose a new two-stage dynamic queries enhanced DETR based separation line regression approach, named DQ-DETR, to predict separation lines from table images directly. Compared to Vallina DETR, we propose three improvements in DQ-DETR to make the two-stage DETR framework work efficiently and effectively for the separation line prediction task: 1) A new query design, named Dynamic Query, to decouple single line query into separable point queries which could intuitively improve the localization accuracy for regression tasks; 2) A dynamic queries based progressive line regression approach to progressively regressing points on the line which further enhances localization accuracy for distorted tables; 3) A prior-enhanced matching strategy to solve the slow convergence issue of DETR. After separation line prediction, a simple relation network based cell merging module is used to recover spanning cells. With these new techniques, our TSRFormer achieves state-of-the-art performance on several benchmark datasets, including SciTSR, PubTabNet, WTW and FinTabNet. Furthermore, we have validated the robustness and high localization accuracy of our approach to tables with complex structures, borderless cells, large blank spaces, empty or spanning cells as well as distorted or even curved shapes on a more challenging real-world in-house dataset.

cs.CV↗

arXiv2023

Stecformer: Spatio-temporal Encoding Cascaded Transformer for Multivariate Long-term Time Series Forecasting

Multivariate long-term time series forecasting is of great application across many domains, such as energy consumption and weather forecasting. With the development of transformer-based methods, the performance of multivariate long-term time series forecasting has been significantly improved, however, the study of spatial features extracting in transformer-based model is rare and the consistency of different prediction periods is unsatisfactory due to the large span. In this work, we propose a complete solution to address these problems in terms of feature extraction and target prediction. For extraction, we design an efficient spatio-temporal encoding extractor including a semi-adaptive graph to acquire sufficient spatio-temporal information. For prediction, we propose a Cascaded Decoding Predictor (CDP) to strengthen the correlation between different intervals, which can also be utilized as a generic component to improve the performance of transformer-based methods. The proposed method, termed as Spatio-temporal Encoding Cascaded Transformer (Stecformer), achieving a notable gap over the baseline model and is comparable with the state-of-the-art performance of transformer-based methods on five benchmark datasets. We hope our attempt will serve as a regular configuration in multivariate long-term time series forecasting in the future.

cs.LG↗

arXiv2023

Entanglement generation and detection in split exciton-polariton condensates

We propose a method of generating and detecting entanglement in two spatially separated excitonpolariton Bose-Einstein condensates (BECs) at steady-state. In our scheme we first create a spinor polariton BEC, such that steady-state squeezing is obtained under a one-axis twisting interaction. Then the condensate is split either physically or virtually, which results in entanglement generated between the two parts. A virtual split means that the condensate is not physically split, but its near-field image is divided into two parts and the spin correlations are deduced from polarization measurements in each half. We theoretically model and examine logarithmic negativity criterion and several correlation-based criteria to show that entanglement exists under experimentally achievable parameters.

quant-ph↗

arXiv2023

Towards the Universal Defense for Query-Based Audio Adversarial Attacks

Recently, studies show that deep learning-based automatic speech recognition (ASR) systems are vulnerable to adversarial examples (AEs), which add a small amount of noise to the original audio examples. These AE attacks pose new challenges to deep learning security and have raised significant concerns about deploying ASR systems and devices. The existing defense methods are either limited in application or only defend on results, but not on process. In this work, we propose a novel method to infer the adversary intent and discover audio adversarial examples based on the AEs generation process. The insight of this method is based on the observation: many existing audio AE attacks utilize query-based methods, which means the adversary must send continuous and similar queries to target ASR models during the audio AE generation process. Inspired by this observation, We propose a memory mechanism by adopting audio fingerprint technology to analyze the similarity of the current query with a certain length of memory query. Thus, we can identify when a sequence of queries appears to be suspectable to generate audio AEs. Through extensive evaluation on four state-of-the-art audio AE attacks, we demonstrate that on average our defense identify the adversary intent with over 90% accuracy. With careful regard for robustness evaluations, we also analyze our proposed defense and its strength to withstand two adaptive attacks. Finally, our scheme is available out-of-the-box and directly compatible with any ensemble of ASR defense models to uncover audio AE attacks effectively without model retraining.

eess.AS↗

arXiv2023

Towards the Transferable Audio Adversarial Attack via Ensemble Methods

In recent years, deep learning (DL) models have achieved significant progress in many domains, such as autonomous driving, facial recognition, and speech recognition. However, the vulnerability of deep learning models to adversarial attacks has raised serious concerns in the community because of their insufficient robustness and generalization. Also, transferable attacks have become a prominent method for black-box attacks. In this work, we explore the potential factors that impact adversarial examples (AEs) transferability in DL-based speech recognition. We also discuss the vulnerability of different DL systems and the irregular nature of decision boundaries. Our results show a remarkable difference in the transferability of AEs between speech and images, with the data relevance being low in images but opposite in speech recognition. Motivated by dropout-based ensemble approaches, we propose random gradient ensembles and dynamic gradient-weighted ensembles, and we evaluate the impact of ensembles on the transferability of AEs. The results show that the AEs created by both approaches are valid for transfer to the black box API.

cs.CR↗

arXiv2023

The Solar Upper Transition Region Imager (SUTRI) onboard the SATech-01 satellite

The Solar Upper Transition Region Imager (SUTRI) onboard the Space Advanced Technology demonstration satellite (SATech-01), which was launched to a sun-synchronous orbit at a height of 500 km in July 2022, aims to test the on-orbit performance of our newly developed Sc-Si multi-layer reflecting mirror and the 2kx2k EUV CMOS imaging camera and to take full-disk solar images at the Ne VII 46.5 nm spectral line with a filter width of 3 nm. SUTRI employs a Ritchey-Chretien optical system with an aperture of 18 cm. The on-orbit observations show that SUTRI images have a field of view of 41.6'x41.6' and a moderate spatial resolution of 8" without an image stabilization system. The normal cadence of SUTRI images is 30 s and the solar observation time is about 16 hours each day because the earth eclipse time accounts for about 1/3 of SATech-01's orbit period. Approximately 15 GB data is acquired each day and made available online after processing. SUTRI images are valuable as the Ne VII 46.5 nm line is formed at a temperature regime of 0.5 MK in the solar atmosphere, which has rarely been sampled by existing solar imagers. SUTRI observations will establish connections between structures in the lower solar atmosphere and corona, and advance our understanding of various types of solar activity such as flares, filament eruptions, coronal jets and coronal mass ejections.

astro-ph.SR↗

arXiv2023

Gaugino masses from misaligned supersymmetry breaking and R-symmetry breaking spurions

In gauge mediation models with multiple spurion fields breaking SUSY and the R-symmetry separately, we show that it is possible to generate gaugino masses at one loop if the R-charge arrangement satisfies a certain condition. The resulting gaugino masses are calculated and suppressed by some power of the messenger mass scale. We present two simple examples to demonstrate this possibility, and discuss possible phenomenology implications.

hep-ph↗

arXiv2022

Stability of structure-aware Taylor methods for tents

Structure-aware Taylor (SAT) methods are a class of timestepping schemes designed for propagating linear hyperbolic solutions within a tent-shaped spacetime region. Tents are useful to design explicit time marching schemes on unstructured advancing fronts with built-in locally variable timestepping for arbitrary spatial and temporal discretization orders. The main result of this paper is that an $s$-stage SAT timestepping within a tent is weakly stable under the time step constraint $Δt \leq Ch^{1+1/s}$, where $Δt$ is the time step size and $h$ is the spatial mesh size. Improved stability properties are also presented for high order SAT time discretizations coupled with low order spatial polynomials. A numerical verification of the sharpness of proven estimates is also included.

math.NA↗

arXiv2022

Mixline: A Hybrid Reinforcement Learning Framework for Long-horizon Bimanual Coffee Stirring Task

Bimanual activities like coffee stirring, which require coordination of dual arms, are common in daily life and intractable to learn by robots. Adopting reinforcement learning to learn these tasks is a promising topic since it enables the robot to explore how dual arms coordinate together to accomplish the same task. However, this field has two main challenges: coordination mechanism and long-horizon task decomposition. Therefore, we propose the Mixline method to learn sub-tasks separately via the online algorithm and then compose them together based on the generated data through the offline algorithm. We constructed a learning environment based on the GPU-accelerated Isaac Gym. In our work, the bimanual robot successfully learned to grasp, hold and lift the spoon and cup, insert them together and stir the coffee. The proposed method has the potential to be extended to other long-horizon bimanual tasks.

cs.RO↗

arXiv2022

Novel counterexample to the Nelson-Seiberg theorem

We present a new type of counterexample to the Nelson-Seiberg theorem. It is a generic R-symmetric Wess-Zumino model with nine chiral superfields, including one field of R-charge 2 and no R-charge 0 field. As in previous counterexamples, the model gives a set of degenerate supersymmetric vacua with a non-zero expectation value for a pair of oppositely R-charged fields. However, one of these fields appears quadratically in the superpotential, and many other fields with non-zero R-charges gain non-zero expectation values at the vacuum, and so this model escapes the sufficient condition for counterexamples established in previous literature. Thus there are still open problems in the relation of R-symmetries to supersymmetry breaking in generic models.

hep-th↗

arXiv2022

Cross-loop propagation of a quasi-periodic extreme-ultraviolet wave train triggered by successive stretching of magnetic field structures during a solar eruption

Solar extreme-ultraviolet (EUV) waves generally refer to large-scale disturbances propagating outward from sites of solar eruptions in EUV imaging observations. Using the recent observations from the Atmospheric Imaging Assembly (AIA) on board the Solar Dynamics Observatory (SDO), we report a quasi-periodic wave train propagating outward at an average speed of $\sim$308 km s$^{-1}$. At least five wavefronts can be clearly identified with the period being $\sim$120 s. These wavefronts originate from the coronal loop expansion, which propagates with an apparent speed of $\sim$95 km s$^{-1}$, about 3 times slower than the wave train. In the absence of a strong lateral expansion, these observational results might be explained by the theoretical model of Chen et al. (2002), which predicted that EUV waves may have two components: a faster component that is a fast-mode magnetoacoustic wave or shock wave and a slower apparent front formed as a result of successive stretching of closed magnetic field lines. In this scenario, the wave train and the successive loop expansion we observed likely correspond to the fast and slow components in the model, respectively.

astro-ph.SR↗

arXiv2022

TSRFormer: Table Structure Recognition with Transformers

We present a new table structure recognition (TSR) approach, called TSRFormer, to robustly recognizing the structures of complex tables with geometrical distortions from various table images. Unlike previous methods, we formulate table separation line prediction as a line regression problem instead of an image segmentation problem and propose a new two-stage DETR based separator prediction approach, dubbed \textbf{Sep}arator \textbf{RE}gression \textbf{TR}ansformer (SepRETR), to predict separation lines from table images directly. To make the two-stage DETR framework work efficiently and effectively for the separation line prediction task, we propose two improvements: 1) A prior-enhanced matching strategy to solve the slow convergence issue of DETR; 2) A new cross attention module to sample features from a high-resolution convolutional feature map directly so that high localization accuracy is achieved with low computational cost. After separation line prediction, a simple relation network based cell merging module is used to recover spanning cells. With these new techniques, our TSRFormer achieves state-of-the-art performance on several benchmark datasets, including SciTSR, PubTabNet and WTW. Furthermore, we have validated the robustness of our approach to tables with complex structures, borderless cells, large blank spaces, empty or spanning cells as well as distorted or even curved shapes on a more challenging real-world in-house dataset.

cs.CV↗

arXiv2022

Supersymmetry and R-symmetries in Wess-Zumino models: properties and model dataset construction

The Nelson-Seiberg theorem and its extensions relate supersymmetry breaking and R-symmetries in Wess-Zumino models. But their applicability may be limited by previously found non-generic counterexamples. Constructing a dataset of R-symmetric Wess-Zumino models is useful for studying the occurrence of such counterexamples as well as other purposes. This work gives a pedagogical review on the basics of supersymmetry in (3+1)-dimensions, Wess-Zumino models and their supergravity extensions, the Nelson-Seiberg theorem and its extensions. We present a preliminary construction of the dataset of R-symmetric Wess-Zumino models with up to 5 chiral fields. Among 925 models in total, 20 of them with non-generic R-charges are counterexamples to both the Nelson-Seiberg theorem and its extensions. Thus the dataset gives an estimation of the accuracy of the field counting method based on these theorems. More constructions and applications of the dataset are expected in future work.

hep-th↗

arXiv2022

On Energy Laws and Stability of Runge--Kutta Methods for Linear Seminegative Problems

This paper presents a systematic theoretical framework to derive the energy identities of general implicit and explicit Runge--Kutta (RK) methods for linear seminegative systems. It generalizes the stability analysis of explicit RK methods in [Z. Sun and C.-W. Shu, SIAM J. Numer. Anal., 57 (2019), pp. 1158-1182]. The established energy identities provide a precise characterization on whether and how the energy dissipates in the RK discretization, thereby leading to weak and strong stability criteria of RK methods. Furthermore, we discover a unified energy identity for all the diagonal Pade approximations, based on an analytical Cholesky type decomposition of a class of symmetric matrices. The structure of the matrices is very complicated, rendering the discovery of the unified energy identity and the proof of the decomposition highly challenging. Our proofs involve the construction of technical combinatorial identities and novel techniques from the theory of hypergeometric series. Our framework is motivated by a discrete analogue of integration by parts technique and a series expansion of the continuous energy law. In some special cases, our analyses establish a close connection between the continuous and discrete energy laws, enhancing our understanding of their intrinsic mechanisms. Several specific examples of implicit methods are given to illustrate the discrete energy laws. A few numerical examples further confirm the theoretical properties.

math.NA↗

arXiv2021

A formal notion of genericity and term-by-term vanishing superpotentials at supersymmetric vacua from R-symmetric Wess-Zumino models

It is known in previous literature that if a Wess-Zumino model with an R-symmetry gives a supersymmetric vacuum, the superpotential vanishes at the vacuum. In this work, we establish a formal notion of genericity, and show that if the R-symmetric superpotential has generic coefficients, the superpotential vanishes term-by-term at a supersymmetric vacuum. This result constrains the form of the superpotential which leads to a supersymmetric vacuum. It may contribute to a refined classification of R-symmetric Wess-Zumino models, and find applications in string constructions of vacua with small superpotentials. A similar result for a scalar potential system with a scaling symmetry is discussed.

hep-th↗

arXiv2021

A sufficient condition for counterexamples to the Nelson-Seiberg theorem

Several counterexample models to the Nelson-Seiberg theorem have been discovered in previous literature, with generic superpotentials respecting the R-symmetry and non-generic R-charge assignments for chiral fields. This work present a sufficient condition for such counterexample models: The number of R-charge 2 fields, which is greater than the number of R-charge 0 fields, must be less than or equal to the number of R-charge 0 fields plus the number of independent field pairs with opposite R-charges and satisfying some extra requirements. We give a correct count of such field pairs when there are multiple field pairs with degenerated R-charges. These models give supersymmetric vacua with spontaneous R-symmetry breaking, thus are counterexamples to both the Nelson-Seiberg theorem and its extensions.

hep-th↗