arXiv Science⌕ Search

arXiv subjects

Gang Li

Publications and source records attributed to Gang Li.

At least 325 records · Page 18Linked to original sources
arXiv2023

Internal magnetic fields in 13 red giants detected by asteroseismology

While surface fields have been measured for stars across the HR diagram, internal magnetic fields remain largely unknown. The recent seismic detection of magnetic fields in the cores of several Kepler red giants has opened a new avenue to understand better the origin of magnetic fields and their impact on stellar structure and evolution. We aim to use asteroseismology to systematically search for internal magnetic fields in red giant stars and to determine the strengths and geometries of these fields. Magnetic fields are known to break the symmetry of rotational multiplets. In red giants, oscillation modes are mixed, behaving as pressure modes in the envelope and as gravity modes in the core. Magnetism-induced asymmetries are expected to be stronger for g-dominated modes than for p-dominated modes and to decrease with frequency. After collecting a sample of 2500 Kepler red giant stars with clear mixed-mode patterns, we specifically searched for targets among 1200 stars with dipole triplets. We identified 13 stars exhibiting clear asymmetric multiplets and measured their parameters, especially the asymmetry parameter and the magnetic frequency shift. By combining these estimates with best-fitting stellar models, we measured average core magnetic fields ranging from 20 to 150kG, corresponding to 5% to 30% of the critical field strengths. We showed that the detected core fields have various horizontal geometries, some of which significantly differ from a dipolar configuration. We found that the field strengths decrease with stellar evolution, despite the fact that the cores of these stars are contracting. Even though these stars have strong internal magnetic fields, they display normal core rotation rates, suggesting no significantly different histories of angular momentum transport compared to other red giant stars. We also discuss the possible origin of the detected fields.

astro-ph.SR↗
arXiv2023

When AUC meets DRO: Optimizing Partial AUC for Deep Learning with Non-Convex Convergence Guarantee

In this paper, we propose systematic and efficient gradient-based methods for both one-way and two-way partial AUC (pAUC) maximization that are applicable to deep learning. We propose new formulations of pAUC surrogate objectives by using the distributionally robust optimization (DRO) to define the loss for each individual positive data. We consider two formulations of DRO, one of which is based on conditional-value-at-risk (CVaR) that yields a non-smooth but exact estimator for pAUC, and another one is based on a KL divergence regularized DRO that yields an inexact but smooth (soft) estimator for pAUC. For both one-way and two-way pAUC maximization, we propose two algorithms and prove their convergence for optimizing their two formulations, respectively. Experiments demonstrate the effectiveness of the proposed algorithms for pAUC maximization for deep learning on various datasets.

cs.LG↗
arXiv2023

Cartoondiff: Training-free Cartoon Image Generation with Diffusion Transformer Models

Image cartoonization has attracted significant interest in the field of image generation. However, most of the existing image cartoonization techniques require re-training models using images of cartoon style. In this paper, we present CartoonDiff, a novel training-free sampling approach which generates image cartoonization using diffusion transformer models. Specifically, we decompose the reverse process of diffusion models into the semantic generation phase and the detail generation phase. Furthermore, we implement the image cartoonization process by normalizing high-frequency signal of the noisy image in specific denoising steps. CartoonDiff doesn't require any additional reference images, complex model designs, or the tedious adjustment of multiple parameters. Extensive experimental results show the powerful ability of our CartoonDiff. The project page is available at: https://cartoondiff.github.io/

cs.CV↗
arXiv2023

Variability of magnetic hot stars from the TESS observations

Magnetic hot stars refer to the stars, which effective temperatures approximately in the range from 7,000 to 50,000 K, and with large-scale globally organized magnetic fields. These magnetic fields exhibit strengths ranging from tens of Gauss to tens of kilo-Gauss. They are key in understanding the effects caused by magnetic fields in the stellar evolution. However, there are only three magnetic hot stars studied via a combination of spectropolarimetric and asteroseismic modeling. Combined with $Transiting\;Exoplanet\;Survey\;Satellite\;(TESS)$ 1-56 sectors data sets, we provided a photometric variability and stochastic low frequency (SLF) variability study of 118 magnetic hot stars. 9 new rotating variable stars are identified. Using the Bayesian Markov Chain Monte Carlo (MCMC) framework, we fitted the morphologies of SLF variability for magnetic hot stars. Our analysis reveals that the magnetic hot stars in our sample have $γ< 5.5$ with the vast majority having $1 \leq γ\leq 3$. The $ν_{\rm char}$ is primarily in the ranges of $0\;\text{d}^{-1} < ν_{\rm char} < 6.3\;\text{d}^{-1}$. The amplitude of SLF variability, log$α_{\rm 0}$, shows a dominant distribution ranging from 0.8 to 3. No significant correlations are observed between the luminosity and fitting parameters, suggesting no clear dependence of SLF variability on stellar mass for our sample of magnetic hot stars with masses between approximately $1.5 M_{\odot}< M < 20 M_{\odot}$. We found a significant negative correlation between the $B_{\rm p}$ and $ν_{char}$. This suppression effect of magnetic fields on $ν_{\rm char}$ may be a result of their inhibition of macroturbulence.

astro-ph.SR↗
arXiv2023

Nonreciprocal photon transport in indirectly coupled whispering-gallery mode resonators

We study the reflection and transmission properties of a system comprising two whispering-gallery mode resonators, each containing a Zeeman-split quantum dot and side-coupled to an optical fiber. Our results demonstrate that unidirectional reflection and transmission can be achieved by tuning the coupling strength between the resonators and the optical fiber. Furthermore, we establish a correspondence between quantum dot energy level resonance frequencies and the positions of low reflection (transmission) peaks at a phase shift of π. This research provides insights for the development of quantum optical devices like isolators, circulators, and routers.

quant-ph↗
arXiv2023

Correlated flat bands in the paramagnetic phase of triangular antiferromagnets Na$_2$BaX(PO$_4$)$_2$ (X = Mn, Co, Ni)

Flat band systems in condensed matter physics are intriguing because they can exhibit exotic phases and unconventional properties. In this work, we studied three correlated magnetic systems, Na$_2$BaX(PO$_4$)$_2$ (X = Mn, Co, Ni), and revealed their unusual electronic structure and magnetic properties. Despite their different effective angular momentum, our first-principles calculations showed a similar electronic structure among them. However, their different valence configurations led to different responses to electronic correlations in the high-temperature paramagnetic phase. Using the dynamical mean-field method, we found that all systems can be understood as a multi-band Hubbard model with Hund'ss coupling. Our calculations of spin susceptibility and the {\it ab-initio} estimation of magnetic exchange coupling indicated strong intra-plane antiferromagnetic coupling and weak inter-plane coupling in all systems. The ground states of these systems are largely degenerate. It is likely that none of these magnetic states would dominate over the others, leading to the possibility of quantum spin liquid states in these systems. Our work unifies the understanding of these three structurally similar systems and opens new avenues for exploring correlated flat bands with distinct electronic and magnetic responses.

cond-mat.str-el↗
arXiv2023

MCTNet: A Multi-Scale CNN-Transformer Network for Change Detection in Optical Remote Sensing Images

For the task of change detection (CD) in remote sensing images, deep convolution neural networks (CNNs)-based methods have recently aggregated transformer modules to improve the capability of global feature extraction. However, they suffer degraded CD performance on small changed areas due to the simple single-scale integration of deep CNNs and transformer modules. To address this issue, we propose a hybrid network based on multi-scale CNN-transformer structure, termed MCTNet, where the multi-scale global and local information are exploited to enhance the robustness of the CD performance on changed areas with different sizes. Especially, we design the ConvTrans block to adaptively aggregate global features from transformer modules and local features from CNN layers, which provides abundant global-local features with different scales. Experimental results demonstrate that our MCTNet achieves better detection performance than existing state-of-the-art CD methods.

cs.CV↗
arXiv2023

Study on Virtual Gear Hobbing Simulation and Gear Tooth Surface Accuracy

This paper presents a digital simulation method for the hobbing process of cylindrical gears. Based on the gear generation principle, taking the professional software as the tool, the problem of virtual hobbing simulation on involute helical gears was studied, and the virtual hobbing simulation of hobbing on the whole gear was completed by using macros of CATIA V5. The validity of this method was validated by analyzing the tooth surface accuracy error of the model which was below 0.001 mm between the virtual tooth surface and the theoretical tooth surface and the possible factors that affected the tooth surface accuracy during manufacturing were also carried on the discussion. It offers a fictitious three-D platform for studying the principle of manufacture errors of a gear-cutting machine as well as the finite element analysis between the ideal tooth surface and the erroneous tooth surface.

cs.GR↗
arXiv2023

The East-West Asymmetry of Particle Intensity in Energetic Storm Particle Events

We examine the East-West asymmetry of the peak intensity in energetic storm particle (ESP) events using the improved Particle Acceleration and Transport in the Heliosphere (iPATH) model. We find that injection efficiency peaks east of the nose of coronal mass ejection shock where the shock exhibits a quasi-parallel geometry. We show that the peak intensity at the eastern flank is generally larger than that at the western flank and it positively correlates with the injection efficiency. We also examine this asymmetry for heavy ions, which depends sensitively on the ion energy. Comparison between the modelling results with the measurements of ESP events at 1 au shows a reasonable agreement. We suggest that the injection efficiency can be a primary factor leading to the East-West asymmetry of the peak intensity in ESP events. Additionally, the charge-to-mass (Q/A) dependence of the maximum particle energy affects this asymmetry for heavy ions.

astro-ph.SR↗
arXiv2023

Performance Analysis and Evaluation of Post Quantum Secure Blockchained Federated Learning

Post-quantum security is critical in the quantum era. Quantum computers, along with quantum algorithms, make the standard cryptography based on RSA or ECDSA over FL or Blockchain vulnerable. The implementation of post-quantum cryptography (PQC) over such systems is poorly understood as PQC is still in its standardization phase. In this work, we propose a hybrid approach to employ PQC over blockchain-based FL (BFL), where we combine a stateless signature scheme like Dilithium (or Falcon) with a stateful hash-based signature scheme like the extended Merkle Signature Scheme (XMSS). We propose a linear-based formulaic approach to device role selection mechanisms based on multiple factors to address the performance aspect. Our holistic approach of utilizing a verifiable random function (VRF) to assist in the blockchain consensus mechanism shows the practicality of the proposed approaches. The proposed method and extensive experimental results contribute to enhancing the security and performance aspects of BFL systems.

cs.CR↗
arXiv2023

Improving Federated Aggregation with Deep Unfolding Networks

The performance of Federated learning (FL) is negatively affected by device differences and statistical characteristics between participating clients. To address this issue, we introduce a deep unfolding network (DUN)-based technique that learns adaptive weights that unbiasedly ameliorate the adverse impacts of heterogeneity. The proposed method demonstrates impressive accuracy and quality-aware aggregation. Furthermore, it evaluated the best-weighted normalization approach to define less computational power on the aggregation method. The numerical experiments in this study demonstrate the effectiveness of this approach and provide insights into the interpretability of the unbiased weights learned. By incorporating unbiased weights into the model, the proposed approach effectively addresses quality-aware aggregation under the heterogeneity of the participating clients and the FL environment. Codes and details are \href{https://github.com/shanikairoshi/Improved_DUN_basedFL_Aggregation}{here}.

cs.LG↗
arXiv2023

Quantum Federated Learning: Analysis, Design and Implementation Challenges

Quantum Federated Learning (QFL) has gained significant attention due to quantum computing and machine learning advancements. As the demand for QFL continues to surge, there is a pressing need to comprehend its intricacies in distributed environments. This paper aims to provide a comprehensive overview of the current state of QFL, addressing a crucial knowledge gap in the existing literature. We develop ideas for new QFL frameworks, explore diverse use cases of applications, and consider the critical factors influencing their design. The technical contributions and limitations of various QFL research projects are examined while presenting future research directions and open questions for further exploration.

quant-ph↗
arXiv2023

PLay: Parametrically Conditioned Layout Generation using Latent Diffusion

Layout design is an important task in various design fields, including user interface, document, and graphic design. As this task requires tedious manual effort by designers, prior works have attempted to automate this process using generative models, but commonly fell short of providing intuitive user controls and achieving design objectives. In this paper, we build a conditional latent diffusion model, PLay, that generates parametrically conditioned layouts in vector graphic space from user-specified guidelines, which are commonly used by designers for representing their design intents in current practices. Our method outperforms prior works across three datasets on metrics including FID and FD-VG, and in user study. Moreover, it brings a novel and interactive experience to professional layout design processes.

cs.LG↗
arXiv2023

Decentralized Quantum Federated Learning for Metaverse: Analysis, Design and Implementation

With the emerging developments of the Metaverse, a virtual world where people can interact, socialize, play, and conduct their business, it has become critical to ensure that the underlying systems are transparent, secure, and trustworthy. To this end, we develop a decentralized and trustworthy quantum federated learning (QFL) framework. The proposed QFL leverages the power of blockchain to create a secure and transparent system that is robust against cyberattacks and fraud. In addition, the decentralized QFL system addresses the risks associated with a centralized server-based approach. With extensive experiments and analysis, we evaluate classical federated learning (CFL) and QFL in a distributed setting and demonstrate the practicality and benefits of the proposed design. Our theoretical analysis and discussions develop a genuinely decentralized financial system essential for the Metaverse. Furthermore, we present the application of blockchain-based QFL in a hybrid metaverse powered by a metaverse observer and world model. Our implementation details and code are publicly available 1.

cs.LG↗
arXiv2023

Insight-HXMT Measurements of the Diffuse X-ray Background

We present an X-ray spectrum of the diffuse X-ray background (DXRB) between 1.5 and 120 keV, as measured with the Low-Energy Detector (LE) and the High-Energy Detector (HE) aboard the Insight-HXMT satellite, based on 'blank-sky' observations. LE covers a nominal energy range of 1-15 keV and HE 20-250 keV, but calibration issues and data quality narrowed the energy range for this work. The LE background was directly measured with `blind' detector modules, while the HE background was derived from Earth-occultation data. With the LE data alone, the measured DXRB spectrum can be well described by a power law; fitting the LE and HE data jointly, however, a spectral cut-off must be introduced in the model to account for the measurements above 30 keV. Modelling the combined spectrum with a cut-off power law, the best-fit photon index is 1.40, normalisation $9.57$~$\rm ph~cm^{-2}~s^{-1}~keV^{-1}~sr^{-1} $ (at 1 keV), and cut-off energy 55 keV, after correcting for the effects of the Earth albedo and atmospheric emission (which are significant in the HE band). Based on the best-fit cut-off power law, we derived the spectral energy distribution (SED) of the DXRB. The shape of the SED is in general agreement with the published measurements, but the overall normalization is lower by varying amounts, except for the HEAO-1 result, with which our result is in good agreement.

astro-ph.HE↗
arXiv2023

Influence of the deviation of the matter power spectrum at small scales on the global 21-cm signal at cosmic dawn

The matter power spectrum has been strongly constrained by astronomical measurements at large scales, but only weakly at small scales. Compared with the standard scenario, the deviation of the matter power spectrum at small scales has influence on the cosmological structure formation, e.g., the comoving number density of dark matter halos. The thermal history of the intergalactic medium (IGM) can be changed if dark matter is made of weakly interacting massive particles and can annihilate into standard model particles. The changes of the evolution of IGM could leave imprints on the relevant astronomical observations. Taking into account the dark matter annihilation, we investigate the impact of the deviation of matter power spectrum at small scales on the global 21-cm signal. In view of the measurements of the global 21-cm signal by the EDGES experiment, we explore the allowed parameter space of $m_s$, which describes the degree of deviation, by requiring the differential brightness temperature of the global 21-cm signal $δT_{21} \le -50~\rm mK$ at redshift $z=17$.

astro-ph.CO↗
arXiv2023

LibAUC: A Deep Learning Library for X-Risk Optimization

This paper introduces the award-winning deep learning (DL) library called LibAUC for implementing state-of-the-art algorithms towards optimizing a family of risk functions named X-risks. X-risks refer to a family of compositional functions in which the loss function of each data point is defined in a way that contrasts the data point with a large number of others. They have broad applications in AI for solving classical and emerging problems, including but not limited to classification for imbalanced data (CID), learning to rank (LTR), and contrastive learning of representations (CLR). The motivation of developing LibAUC is to address the convergence issues of existing libraries for solving these problems. In particular, existing libraries may not converge or require very large mini-batch sizes in order to attain good performance for these problems, due to the usage of the standard mini-batch technique in the empirical risk minimization (ERM) framework. Our library is for deep X-risk optimization (DXO) that has achieved great success in solving a variety of tasks for CID, LTR and CLR. The contributions of this paper include: (1) It introduces a new mini-batch based pipeline for implementing DXO algorithms, which differs from existing DL pipeline in the design of controlled data samplers and dynamic mini-batch losses; (2) It provides extensive benchmarking experiments for ablation studies and comparison with existing libraries. The LibAUC library features scalable performance for millions of items to be contrasted, faster and better convergence than existing libraries for optimizing X-risks, seamless PyTorch deployment and versatile APIs for various loss optimization. Our library is available to the open source community at https://github.com/Optimization-AI/LibAUC, to facilitate further academic research and industrial applications.

cs.LG↗
arXiv2023

Nonlinear optomechanical resonance entering a self-organized energy transfer pattern

The energy transfer between different subsystems or different vibration modes is always one of the most interested problems in the study of the resonance phenomena in coupled nonlinear dynamical systems. With an optomechanical system operating in the regime of unresolved sideband, where its mechanical frequency is lower than the cavity field damping rate, we illustrate the existence of a special nonlinear resonance phenomenon. This type of previously unknown resonance manifests an organized pattern of the coupled cavity field and mechanical oscillation, so that the cavity field precisely pushes the mechanical oscillator within an appropriate small time window in each mechanical oscillation period and the mechanical energy will increase by a jump of almost fixed amount after each oscillation cycle. The scenario is realized at a resonance point where the frequency difference of two driving fields matches the mechanical frequency of the system, and this condition of drive-frequency match is found to trigger a mechanism to lock the two subsystems of an unresolved-sideband optomechanical system into a highly ordered energy transfer as the above mentioned. Due to a significantly enhanced nonlinearity in the vicinity of the resonance point, optical frequency combs can be generated under pump powers of thousand times lower, as compared to the use of a single-tone driving field for the purpose. An unresolved sideband system under the drives without satisfying the resonance condition also demonstrates other interesting dynamical behaviors. Most of all, by providing a realistic picture for the nonlinear optomechanical dynamics in unresolved sideband regime, our study points to a direction to observe novel dynamical phenomena and realize other applications with the systems of less technical restrictions.

physics.optics↗