arXiv subjects

Shuyu Cheng

Publications and source records attributed to Shuyu Cheng.

At least 19 recordsLinked to original sources

arXiv2025

Improving the Convergence Rate of Ray Search Optimization for Query-Efficient Hard-Label Attacks

In hard-label black-box adversarial attacks, where only the top-1 predicted label is accessible, the prohibitive query complexity poses a major obstacle to practical deployment. In this paper, we focus on optimizing a representative class of attacks that search for the optimal ray direction yielding the minimum $\ell_2$-norm perturbation required to move a benign image into the adversarial region. Inspired by Nesterov's Accelerated Gradient (NAG), we propose a momentum-based algorithm, ARS-OPT, which proactively estimates the gradient with respect to a future ray direction inferred from accumulated momentum. We provide a theoretical analysis of its convergence behavior, showing that ARS-OPT enables more accurate directional updates and achieves faster, more stable optimization. To further accelerate convergence, we incorporate surrogate-model priors into ARS-OPT's gradient estimation, resulting in PARS-OPT with enhanced performance. The superiority of our approach is supported by theoretical guarantees under standard assumptions. Extensive experiments on ImageNet and CIFAR-10 demonstrate that our method surpasses 13 state-of-the-art approaches in query efficiency.

cs.LG↗

arXiv2025

Boosting Ray Search Procedure of Hard-label Attacks with Transfer-based Priors

One of the most practical and challenging types of black-box adversarial attacks is the hard-label attack, where only the top-1 predicted label is available. One effective approach is to search for the optimal ray direction from the benign image that minimizes the $\ell_p$-norm distance to the adversarial region. The unique advantage of this approach is that it transforms the hard-label attack into a continuous optimization problem. The objective function value is the ray's radius, which can be obtained via binary search at a high query cost. Existing methods use a "sign trick" in gradient estimation to reduce the number of queries. In this paper, we theoretically analyze the quality of this gradient estimation and propose a novel prior-guided approach to improve ray search efficiency both theoretically and empirically. Specifically, we utilize the transfer-based priors from surrogate models, and our gradient estimators appropriately integrate them by approximating the projection of the true gradient onto the subspace spanned by these priors and random directions, in a query-efficient manner. We theoretically derive the expected cosine similarities between the obtained gradient estimators and the true gradient, and demonstrate the improvement achieved by incorporating priors. Extensive experiments on the ImageNet and CIFAR-10 datasets show that our approach significantly outperforms 11 state-of-the-art methods in terms of query efficiency.

cs.CV↗

arXiv2025

The surface binding and energy issues in rational design of the separation membrane of Li||S batteries

Lithium-sulfur batteries (LSBs) represent one of the most promising next-generation energy storage technologies, offering exceptionally high energy densities. However, their widespread adoption remains hindered by challenges such as sluggish conversion reactions and the dissolution of lithium polysulfides, which lead to poor cycling stability and reduced performance. While significant efforts have been made to address these limitations, the energy storage capabilities of LSBs in practical devices remain far from achieving their full potential. This report delves into recent advancements in the rational design of separation membranes for LSBs, focusing on addressing fundamental issues related to surface binding and surface energy interactions within materials science. By examining the functionalization and optimization of separation membranes, we aim to highlight strategies that can guide the development of more robust and efficient LSBs, bringing them closer to practical implementation.

cond-mat.mtrl-sci↗

arXiv2024

Efficient Black-box Adversarial Attacks via Bayesian Optimization Guided by a Function Prior

This paper studies the challenging black-box adversarial attack that aims to generate adversarial examples against a black-box model by only using output feedback of the model to input queries. Some previous methods improve the query efficiency by incorporating the gradient of a surrogate white-box model into query-based attacks due to the adversarial transferability. However, the localized gradient is not informative enough, making these methods still query-intensive. In this paper, we propose a Prior-guided Bayesian Optimization (P-BO) algorithm that leverages the surrogate model as a global function prior in black-box adversarial attacks. As the surrogate model contains rich prior information of the black-box one, P-BO models the attack objective with a Gaussian process whose mean function is initialized as the surrogate model's loss. Our theoretical analysis on the regret bound indicates that the performance of P-BO may be affected by a bad prior. Therefore, we further propose an adaptive integration strategy to automatically adjust a coefficient on the function prior by minimizing the regret bound. Extensive experiments on image classifiers and large vision-language models demonstrate the superiority of the proposed algorithm in reducing queries and improving attack success rates compared with the state-of-the-art black-box attacks. Code is available at https://github.com/yibo-miao/PBO-Attack.

cs.LG↗

arXiv2024

Revealing the EuCd_{2}As_{2} Semiconducting Band Gap via n-type La-Doping

EuCd_{2}As_{2} has attracted considerable interest as one of the few magnetic Weyl semimetal candidate materials, although recently there have been emerging reports that claim it to have a semiconducting electronic structure. To resolve this debate, we established the growth of n-type EuCd_{2}As_{2} crystals, to directly visualize the nature of the conduction band using angle resolve photoemission spectroscopy (ARPES). We show that La-doping leads to n-type transport signatures in both the thermopower and Hall effect measurements, in crystals with doping levels at 2 - 6 x 10^{17} e^{-} cm^{-3}. Both p-type and n-type doped samples exhibit antiferromagnetic ordering at 9 K. ARPES experiments at 6 K clearly show the presence of the conduction band minimum at 0.8 eV above the valence band maximum, which is further corroborated by the observation of a 0.71 - 0.72 eV band gap in room temperature diffuse reflectance absorbance measurements. Together these findings unambiguously show that EuCd_{2}As_{2} is indeed a semiconductor with a substantial band gap and not a topological semimetal.

cond-mat.mtrl-sci↗

arXiv2024

Momentum-space Observation of Optically Excited Non-Thermal Electrons in Graphene with Persistent Pseudospin Polarization

The unique optical properties of graphene, with broadband absorption and ultrafast response, make it a critical component of optoelectronic and spintronic devices. Using time-resolved momentum microscopy with high data rate and high dynamic range, we report momentum-space measurements of electrons promoted to the graphene conduction band with visible light, and their subsequent relaxation. We observe a pronounced non-thermal distribution of nascent photoexcited electrons with lattice pseudospin polarization in remarkable agreement with results of simple tight-binding theory. By varying the excitation fluence, we vary the relative importance of electron-electron vs. electron-phonon scattering in the relaxation of the initial distribution. Increasing the excitation fluence results in increased noncollinear electron-electron scattering and reduced pseudospin polarization, although up-scattered electrons retain a degree of polarization. These detailed momentum-resolved electron dynamics in graphene demonstrate the capabilities of high-performance time-resolved momentum microscopy in the study of 2D materials and can inform the design of graphene devices.

cond-mat.mes-hall↗

arXiv2024

Atomic Layer Molecular Beam Epitaxy of Kagome Magnet RMn$_6$Sn$_6$ (R = Er, Tb) Thin Films

Kagome lattices have garnered substantial interest because their band structure consists of topological flat bands and Dirac cones. The RMn$_6$Sn$_6$ (R = rare earth) compounds are particularly interesting because of the existence of large intrinsic anomalous Hall effect (AHE) which originates from the gapped Dirac cones near the Fermi level. This makes RMn$_6$Sn$_6$ an outstanding candidate for realizing the high-temperature quantum anomalous Hall effect. The growth of RMn$_6$Sn$_6$ thin films is beneficial for both fundamental research and potential applications. However, most of the studies on RMn$_6$Sn$_6$ have focused on bulk crystals so far, and the synthesis of RMn$_6$Sn$_6$ thin films has not been reported so far. Here we report the atomic layer molecular beam epitaxy growth, structural and magnetic characterizations, and transport properties of ErMn$_6$Sn$_6$ and TbMn$_6$Sn$_6$ thin films. It is especially noteworthy that TbMn$_6$Sn$_6$ thin films have out-of-plane magnetic anisotropy, which is important for realizing the quantum anomalous Hall effect. Our work paves the avenue toward the control of the AHE using devices patterned from RMn$_6$Sn$_6$ thin films.

cond-mat.mtrl-sci↗

arXiv2023

Tuning the Curie temperature of a 2D magnet/topological insulator heterostructure to above room temperature by epitaxial growth

Heterostructures of two-dimensional (2D) van der Waals (vdW) magnets and topological insulators (TI) are of substantial interest as candidate materials for efficient spin-torque switching, quantum anomalous Hall effect, and chiral spin textures. However, since many of the vdW magnets have Curie temperatures below room temperature, we want to understand how materials can be modified to stabilize their magnetic ordering to higher temperatures. In this work, we utilize molecular beam epitaxy to systematically tune the Curie temperature ($T_C$) in thin film Fe$_3$GeTe$_2$/Bi$_2$Te$_3$ from bulk-like values ($\sim$220 K) to above room temperature by increasing the growth temperature from 300 $^\circ$C to 375 $^\circ$C. For samples grown at 375 $^\circ$C, cross-sectional scanning transmission electron microscopy (STEM) reveals the spontaneous formation of different Fe$_m$Ge$_n$Te$_2$ compositions (e.g. Fe$_5$Ge$_2$Te$_2$ and Fe$_7$Ge$_6$Te$_2$) as well as intercalation in the vdW gaps, which are possible origins of the enhanced Curie temperature. This observation paves the way for developing various Fe$_m$Ge$_n$Te$_2$/TI heterostructures with novel properties.

cond-mat.mtrl-sci↗

arXiv2023

Magnon Diffusion Length and Longitudinal Spin Seebeck Effect in Vanadium Tetracyanoethylene (V[TCNE]$_x$, $x \sim 2$)

Spintronic, spin caloritronic, and magnonic phenomena arise from complex interactions between charge, spin, and structural degrees of freedom that are challenging to model and even more difficult to predict. This situation is compounded by the relative scarcity of magnetically-ordered materials with relevant functionality, leaving the field strongly constrained to work with a handful of well-studied systems that do not encompass the full phase space of phenomenology predicted by fundamental theory. Here we present an important advance in this coupled theory-experiment challenge, wherein we extend existing theories of the spin Seebeck effect (SSE) to explicitly include the temperature-dependence of magnon non-conserving processes. This expanded theory quantitatively describes the low-temperature behavior of SSE signals previously measured in the mainstay material yttrium iron garnet (YIG) and predicts a new regime for magnonic and spintronic materials that have low saturation magnetization, $M_S$, and ultra-low damping. Finally, we validate this prediction by directly observing the spin Seebeck resistance (SSR) in the molecule-based ferrimagnetic semiconductor vanadium tetracyanoethylene (V[TCNE]$_x$, $x \sim 2$). These results validate the expanded theory, yielding SSR signals comparable in magnitude to YIG and extracted magnon diffusion length ($λ_m>1$ $μ$ m) and magnon lifetime for V[TCNE]$_x$ ($τ_{th}\approx 1-10$ $μ$ s) exceeding YIG ($τ_{th}\sim 10$ ns). Surprisingly, these properties persist to room temperature despite relatively low spin wave stiffness (exchange). This identification of a new regime for highly efficient SSE-active materials opens the door to a new class of magnetic materials for spintronic and magnonic applications.

cond-mat.mtrl-sci↗

arXiv2023

Epitaxial Kagome Thin Films as a Platform for Topological Flat Bands

Systems with flat bands are ideal for studying strongly correlated electronic states and related phenomena. Among them, kagome-structured metals such as CoSn have been recognized as promising candidates due to the proximity between the flat bands and the Fermi level. A key next step will be to realize epitaxial kagome thin films with flat bands to enable tuning of the flat bands across the Fermi level via electrostatic gating or strain. Here we report the band structures of epitaxial CoSn thin films grown directly on insulating substrates. Flat bands are observed using synchrotron-based angle-resolved photoemission spectroscopy (ARPES). The band structure is consistent with density functional theory (DFT) calculations, and the transport properties are quantitatively explained by the band structure and semiclassical transport theory. Our work paves the way to realize flat band-induced phenomena through fine-tuning of flat bands in kagome materials.

cond-mat.mtrl-sci↗

arXiv2023

Observation of Skyrmion Bubbles in Multilayer [Pt/Co/Cu]n using spin-polarized STM

Magnetic multilayers are a promising platform for storage and logic devices based on skyrmion spin textures, due to the large materials phase space for tuning properties. Epitaxial superlattice structures of [Pt/Co/Cu]n thin films were grown by molecular beam epitaxy at room temperature. Spin-polarized scanning tunneling microscopy (SP-STM) of these samples was used to probe the connection between surface structure and skyrmion morphology with nanoscale spatial resolution. Irregular-shaped skyrmion bubbles were observed, with effective diameters from 20-200 nm that are much larger than the nanoscale grain structure of the surface topography. Nucleation, annihilation, and motion of skyrmion bubbles could be driven using the stray field of the ferromagnetic tip in repeated imaging, and spin-polarized current/voltage pulses. Our detailed comparison of STM topography and differential conductance images shows that there are no surface defects or inhomogeneities at length scales that could account for the range in skyrmion bubble size or shape observed in the measurements.

cond-mat.mtrl-sci↗

arXiv2023

An Atomically Tailored Chiral Magnet with Small Skyrmions at Room Temperature

Creating materials that do not exist in nature can lead to breakthroughs in science and technology. Magnetic skyrmions are topological excitations that have attracted great attention recently for their potential applications in low power, ultrahigh density memory. A major challenge has been to find materials that meet the dual requirement of small skyrmions stable at room temperature. Here we meet both these goals by developing epitaxial FeGe films with excess Fe using atomic layer molecular beam epitaxy (MBE) far from thermal equilibrium. Our novel atomic layer design permits the incorporation of 20% excess Fe while maintaining a non-centrosymmetric crystal structure supported by theoretical calculations and necessary for stabilizing skyrmions. We show that the Curie temperature is well above room temperature, and that the skyrmions probed by topological Hall effect have sizes down to 15 nm as imaged by Lorentz transmission electron microscopy (LTEM) and magnetic force microscopy (MFM). Our results illustrate new avenues for creating artificial materials tailored at the atomic scale that can impact nanotechnology.

cond-mat.mtrl-sci↗

arXiv2023

Room-Temperature Magnetic Skyrmions in Pt/Co/Cu Multilayers

Magnetic skyrmions are promising for next-generation information storage and processing owing to their potential advantages in data storage density, robustness, and energy efficiency. The magnetic multilayers consisting of Pt, Co, and a third metal element $X$ provide an ideal platform to study the skyrmions due to their highly tunable magnetic properties. Here, we report the observation of room-temperature bubble-like N\'eel skyrmions in epitaxial Pt/Co/Cu multilayers in samples with multidomain states in zero field. The magneto-optic Kerr effect (MOKE) and superconducting quantum interference device (SQUID) magnetometry are applied to investigate the shapes of the hysteresis loops, the magnetic anisotropy, and the saturation magnetization. By tuning the Co thickness and the number of periods, we achieve perpendicular and in-plane magnetized states and multidomain states that are identified by a wasp-waisted hysteresis loop. Skyrmions are directly imaged by magnetic force microscopy (MFM) and Lorentz transmission electron microscopy (LTEM). The development of room-temperature skyrmions in Pt/Co/Cu multilayers may lead to advances in skyrmion-related research and applications.

cond-mat.mtrl-sci↗

arXiv2022

Atomic Layer Epitaxy of Kagome Magnet Fe${_3}$Sn${_2}$ and Sn-modulated Heterostructures

Magnetic materials with kagome crystal structure exhibit rich physics such as frustrated magnetism, skyrmion formation, topological flat bands, and Dirac/Weyl points. Until recently, most studies on kagome magnets have been performed on bulk crystals or polycrystalline films. Here we report the atomic layer molecular beam epitaxy synthesis of high-quality thin films of topological kagome magnet Fe${_3}$Sn${_2}$. Structural and magnetic characterization of Fe${_3}$Sn${_2}$ on epitaxial Pt(111) identifies highly ordered films with c-plane orientation and an in-plane magnetic easy axis. Studies of the local magnetic structure by anomalous Nernst effect imaging reveals in-plane oriented micrometer size domains. Superlattice structures consisting of Fe${_3}$Sn${_2}$ and Fe${_3}$Sn are also synthesized by atomic layer molecular beam epitaxy, demonstrating the ability to modulate the sample structure at the atomic level. The realization of high-quality films by atomic layer molecular beam epitaxy opens the door to explore the rich physics of this system and investigate novel spintronic phenomena by interfacing Fe${_3}$Sn${_2}$ with other materials.

cond-mat.mtrl-sci↗

arXiv2022

Query-Efficient Black-box Adversarial Attacks Guided by a Transfer-based Prior

Adversarial attacks have been extensively studied in recent years since they can identify the vulnerability of deep learning models before deployed. In this paper, we consider the black-box adversarial setting, where the adversary needs to craft adversarial examples without access to the gradients of a target model. Previous methods attempted to approximate the true gradient either by using the transfer gradient of a surrogate white-box model or based on the feedback of model queries. However, the existing methods inevitably suffer from low attack success rates or poor query efficiency since it is difficult to estimate the gradient in a high-dimensional input space with limited information. To address these problems and improve black-box attacks, we propose two prior-guided random gradient-free (PRGF) algorithms based on biased sampling and gradient averaging, respectively. Our methods can take the advantage of a transfer-based prior given by the gradient of a surrogate model and the query information simultaneously. Through theoretical analyses, the transfer-based prior is appropriately integrated with model queries by an optimal coefficient in each method. Extensive experiments demonstrate that, in comparison with the alternative state-of-the-arts, both of our methods require much fewer queries to attack black-box models with higher success rates.

cs.LG↗

arXiv2022

Momentum-Resolved Exciton Coupling and Valley Polarization Dynamics in Monolayer WS$_2$

Coupling between exciton states across the Brillouin zone in monolayer transition metal dichalcogenides can lead to ultrafast valley depolarization. Using time- and angle-resolved photoemission, we present momentum- and energy-resolved measurements of exciton coupling in monolayer WS$_2$. By comparing full 4D ($k_x, k_y, E, t$) data sets after both linearly and circularly polarized excitation, we are able to disentangle intervalley and intravalley exciton coupling dynamics. Recording in the exciton binding energy basis instead of excitation energy, we observe strong mixing between the B$_{1s}$ exciton and A$_{n>1}$ states. The photoelectron energy and momentum distributions observed from excitons populated via intervalley coupling (e.g. K$^-$ $\rightarrow$ K$^+$) indicate that the dominant valley depolarization mechanism conserves the exciton binding energy and center-of-mass momentum, consistent with intervalley Coulomb exchange. On longer timescales, exciton relaxation is accompanied by contraction of the momentum space distribution.

cond-mat.mtrl-sci↗

arXiv2021

On the Convergence of Prior-Guided Zeroth-Order Optimization Algorithms

Zeroth-order (ZO) optimization is widely used to handle challenging tasks, such as query-based black-box adversarial attacks and reinforcement learning. Various attempts have been made to integrate prior information into the gradient estimation procedure based on finite differences, with promising empirical results. However, their convergence properties are not well understood. This paper makes an attempt to fill up this gap by analyzing the convergence of prior-guided ZO algorithms under a greedy descent framework with various gradient estimators. We provide a convergence guarantee for the prior-guided random gradient-free (PRGF) algorithms. Moreover, to further accelerate over greedy descent methods, we present a new accelerated random search (ARS) algorithm that incorporates prior information, together with a convergence analysis. Finally, our theoretical results are confirmed by experiments on several numerical benchmarks as well as adversarial attacks.

stat.ML↗

arXiv2021

Epitaxial Growth and Domain Structure Imaging of Kagome Magnet Fe$_3$Sn$_2$

Magnetic materials with kagome crystal structure exhibit rich physics such as frustrated magnetism, skyrmion formation, topological flat bands, and Dirac/Weyl points. Until recently, most studies on kagome magnets have been performed on bulk crystals or polycrystalline films. Here we report the synthesis of high-quality epitaxial films of topological kagome magnet Fe$_3$Sn$_2$ by atomic layer molecular beam epitaxy. Structural and magnetic characterization of Fe$_3$Sn$_2$ on epitaxial Pt(111) identifies highly ordered films with c-plane orientation and an in-plane magnetic easy axis. Studies of the local magnetic structure by anomalous Nernst effect imaging reveals in-plane oriented micrometer size domains. The realization of high-quality films by atomic layer molecular beam epitaxy opens the door to explore the rich physics of this system and investigate novel spintronic phenomena by interfacing Fe$_3$Sn$_2$ with other materials.

cond-mat.mtrl-sci↗