arXiv subjects

Xuefei Wang

Publications and source records attributed to Xuefei Wang.

At least 19 recordsLinked to original sources

arXiv2026

Frequency-aware forecasting for short-term typhoon gust prediction

Accurate gust forecasting under typhoon conditions remains challenging due to the highly non-stationary and multi-scale characteristics of extreme wind fluctuations. Existing deep learning models often struggle to simultaneously capture long-term trends and rapid local variations, resulting in degraded performance during extreme events. We propose WDANet, a frequency-aware forecasting framework that integrates stationary wavelet decomposition, a Feature-wise Linear Modulation (FiLM) strategy, and a dual-branch encoder-decoder architecture, enabling separate modeling of trend and fluctuation components. Taking the offshore regions of the Western Pacific in China as an example, we conduct fine-grid wind gust prediction research. The results demonstrate that WDANet shows advantages for short lead times under the experimental setting across a 24-h forecasting horizon and achieves higher prediction accuracy than ECMWF-HRES within the first 6 h. During extreme wind events, WDANet more accurately captures gust peaks and attains the best RMSE and MAE performance. These results highlight its potential for offshore wind power operation, disaster warning, and risk mitigation.

cs.LG↗

arXiv2026

SGAD: A State-Guided Adaptive Decision Framework for Robust EEG-Based Auditory Attention Switch Decoding

Achieving robust EEG-based auditory attention switch decoding (AASD) is crucial for intelligent hearing aids. However, its application is limited as EEG non-stationarity complicates sequential decision-making, and insufficient control of potential confounding factors may overestimate performance. Therefore, we propose a state-guided adaptive decision (SGAD) framework that infers attention transition states via causal state detection and dynamically modulates temporal smoothing through state-guided adaptive gating. We further introduce six hierarchical evaluation protocols to assess generalization across audio, speaker, and subject dimensions. Experimental results show that SGAD improves decoding accuracy and stability while maintaining low response latency across evaluation scenarios. Performance variations across protocols further suggest data partition-related biases. Together, these findings advance robust AASD for neuro-steered hearing applications.

eess.SP↗

arXiv2026

SAGE: Switch-Aware EEG-Guided Soft Gating for Target Speaker Extraction with In-Trial Switching

EEG-guided target speaker extraction is challenging under in-trial auditory attention switching, where neural noise and intrinsic latency can delay or destabilize attention tracking. Conventional methods struggle with dynamic switches and often cause discontinuities at switching points. Therefore, we propose SAGE, a switch-aware EEG-guided soft gating framework that treats in-trial switching as dynamic selection. SAGE generates two candidate speech streams with a robust separator and uses an EEG-guided switch-aware gating module to produce smooth fusion weights and suppress transition artifacts. We further integrate latency-compensated alignment and an uncertainty-driven conservative strategy to handle latency discrepancies and fluctuating EEG reliability. SAGE outperforms baselines, achieving 8.67 dB SI-SDR and 88.24% STOI while reducing average switching latency to 2.04 s. By coupling neural decoding with speech separation, it enables robust target extraction in dynamic scenarios.

eess.SP↗

arXiv2026

Learning to Adapt Cross-Domain Preferences via Meta-LoRA for LLM Personalization

Cross-domain zero- or few-shot personalization aims to generate user-preferred responses in unseen conversational domains from only a handful of target-domain interactions. Existing adaptation methods struggle to calibrate update magnitude under sparse evidence and thus overfit, whereas history-transfer methods often entangle user preferences with source-domain artifacts, yielding unreliable personalization priors and negative transfer. To calibrate adaptation to evidence quality, we propose PAC-Bayes-regularized Meta-LoRA, which uses a meta-learned LoRA initialization as both the adaptation start and prior center, while adjusting update strength according to support-set size and predictive uncertainty. This limits overfitting under sparse or ambiguous evidence while permitting stronger personalization as evidence grows. Controlled adaptation alone does not determine which preferences should transfer across domains or how they should be expressed. We therefore functionally decompose personalization priors into user and domain components, using a human-readable prompt for stable preferences and topology-preserving soft tokens for domain-specific hidden-space conditioning. Experiments across multiple benchmarks and personalization tasks show consistent gains over strong baselines. On HiCUPID, our method reduces cross-domain win-rate degradation by 47.9% relative to the best competing baseline and improves win rate by 110.2% under unseen-user cold start.

cs.AI↗

arXiv2026

Out of Sight: Compression-Aware Content Protection against Agentic Crawlers

The rise of LLM-based agents with reasoning, summarization, and memory capabilities has created a new threat surface for online content that conventional defenses fail to address. Existing defenses like access controls can be circumvented by agents mimicking ordinary browsers, and injection-based defenses often degrade human readability. In this paper, we revisit the agent pipeline and identify context compression, which agents routinely invoke to fit context budgets, as a critical yet overlooked defense layer. We propose CAPE, a framework that protects high-value textual content by injecting invisible perturbations without changing its human-visible surface form, thereby inducing severe information loss during agent compression. CAPE extracts disruptive seed perturbations from an accessible surrogate compressor, then adapts them to query-only target compressors through prior-guided evolution and preference-calibrated candidate prioritization, achieving effective protection under a low query budget. Experiments on three content types and four compression settings show that CAPE improves information loss by up to 75.8% over the strongest baseline while keeping protected content visually indistinguishable from originals. CAPE also transfers to real-world settings, including the LangGraph agent workflow and GitHub Copilot, highlighting its generality and practical value. This paper aims to reveal context compression as a new defense layer, promoting content protection research in the agent era.

cs.CR↗

arXiv2026

MasFACT: Continual Multi-Agent Topology Learning via Geometry-Aware Posterior Transfer

Multi-agent systems (MAS) powered by large language models (LLMs) have emerged as a powerful paradigm for complex problem solving, where performance critically depends on the underlying inter-agent communication topology. However, existing topology generation methods mainly optimize for isolated tasks, while real-world deployments involve streams of evolving tasks, requiring previously effective collaboration patterns to be retained and reused rather than rediscovered or overwritten. We identify a previously underexplored failure mode, \emph{topology forgetting}, in which adapting to new tasks shifts the topology generator away from communication structures required by earlier tasks. This issue stems from cross-task misalignment in both agent-level functional semantics and relational communication structures. To address this challenge, we propose \textbf{\textsc{MasFACT}}, a geometry-aware posterior transfer framework that preserves and reuses historical collaboration knowledge as transferable topology priors. We transfer these priors across task-specific agent spaces through Fused Gromov-Wasserstein optimal transport and perform PAC-Bayes-guided conservative posterior adaptation to balance task-specific plasticity with structural stability. Experiments across class-, domain-, and task-level continual settings demonstrate that \textsc{MasFACT} consistently improves average accuracy while reducing topology forgetting compared to strong topology generation and replay-based baselines, and can be seamlessly integrated with different MAS topology generators.

cs.LG↗

arXiv2026

HiFloat4 Format for Language Model Pre-training on Ascend NPUs

Large foundation models have become central to modern machine learning, with performance scaling predictably with model size and data. However, training and deploying such models incur substantial computational and memory costs, motivating the development of low-precision training techniques. Recent work has demonstrated that 4-bit floating-point (FP4) formats--such as MXFP4 and NVFP4--can be successfully applied to linear GEMM operations in large language models (LLMs), achieving up to 4x improvements in compute throughput and memory efficiency compared to higher-precision baselines. In this work, we investigate the recently proposed HiFloat4 FP4 format for Huawei Ascend NPUs and systematically compare it with MXFP4 in large-scale training settings. All experiments are conducted on Ascend NPU clusters, with linear and expert GEMM operations performed entirely in FP4 precision. We evaluate both dense architectures (e.g., Pangu and LLaMA-style models) and mixture-of-experts (MoE) models, where both standard linear layers and expert-specific GEMMs operate in FP4. Furthermore, we explore stabilization techniques tailored to FP4 training that significantly reduce numerical degradation, maintaining relative error within 1% of full-precision baselines while preserving the efficiency benefits of 4-bit computation. Our results provide a comprehensive empirical study of FP4 training on NPUs and highlight the practical trade-offs between FP4 formats in large-scale dense and MoE models.

cs.LG↗

arXiv2025

Body Management Information Practices on a Female-dominant Platform

With growing awareness of long-term health and wellness, everyday body management has become a widespread practice. Social media platforms and health-related applications offer abundant information for those pursuing healthier lifestyles and more positive body images. While prior Human-Computer Interaction research has focused extensively on technology-mediated health interventions, the user-initiated practices of browsing and evaluating body management information remain underexplored. In this paper, we study a female-dominant social media platform in China to examine how users seek such information and how it shapes their lifestyle choices. Through semi-structured interviews with 18 users, we identify factors including consumerism, poster popularity, and perceived authenticity that influence decision-making, alongside challenges such as discerning reliable methods and managing body anxiety triggered by social media. We contribute insights into how content and media formats interact to shape users' information evaluation, and we outline design implications for supporting more reliable and healthy engagements with body management information.

cs.HC↗

arXiv2025

Cluster Sliding Ferroelectricity in Trilayer Quasi-Hexagonal C$_{60}$

Electric polarization typically originates from non-centrosymmetric charge distributions in compounds. In elemental crystalline materials, chemical bonds between atoms of the same element favor symmetrically distributed electron charges and centrosymmetric structures, making elemental ferroelectrics rare. Compared to atoms, elemental clusters are intrinsically less symmetric and can have various preferred orientations when they are assembled to form crystals. Consequently, the assembly of clusters with different orientations tends to break the inversion symmetry. By exploiting this concept, we show that sliding ferroelectricity naturally emerges in trilayer quasi-hexagonal phase (qHP) C$_{60}$, a cluster-assembled carbon allotrope recently synthesized. Compared to many metallic or semi-metallic elemental ferroelectrics, trilayer qHP C$_{60}$'s have sizable band gaps and several ferroelectric structures, which are distinguishable by measuring their second-harmonic generation (SHG) responses. Some of these phases show both switchable out-of-plane and in-plane polarizations on the order of 0.2 pC/m. The out-of-plane and in-plane polarizations can be switched independently and enable an easy-to-implement construction of Van der Waals homostructures with ferroelectrically switchable chirality.

cond-mat.mtrl-sci↗

arXiv2025

Robust altermagnetism and compensated ferrimagnetism in MnPX$_3$-based (X = S or Se) heterostructures

The recent research interests in the non-relativistic spin splitting of electronic band structures have led to the exploration of altermagnets and other compensated magnets. Here, we show that various types of non-relativistic spin splitting can be robustly induced by constructing Van der Waals heterostructures consisting of materials with intra-plane anti-ferromagnetic orders and suitable substrates. Using MnPX$_3$ (X = S or Se) as an example, which has a Néel magnetic order, we demonstrate that altermagnetic spin splitting can arise in the AA-stacking MnPX$_3$/MPX$_3$ (M = Cd, Mg, or Zn) heterostructures. For the AB-stacking heterostructures that are semiconducting, ferrimagnetic-type spin splitting emerges, and the fully compensated magnetization is protected by the Luttinger theorem. By combining with a Van der Waals ferroelectric substrate like CuInP$_2$S$_6$, MnPX$_3$-based heterostructures can show tunable spin splitting and spin-related properties that depend on the electronic band structures and ferroelectric polarization, which can be non-volatilely reversed by applying an out-of-plane electric field. Our study provides a route to induce tunable non-relativistic spin splitting in experimentally synthesizable two-dimensional magnets.

cond-mat.mtrl-sci↗

arXiv2024

WeSep: A Scalable and Flexible Toolkit Towards Generalizable Target Speaker Extraction

Target speaker extraction (TSE) focuses on isolating the speech of a specific target speaker from overlapped multi-talker speech, which is a typical setup in the cocktail party problem. In recent years, TSE draws increasing attention due to its potential for various applications such as user-customized interfaces and hearing aids, or as a crutial front-end processing technologies for subsequential tasks such as speech recognition and speaker recongtion. However, there are currently few open-source toolkits or available pre-trained models for off-the-shelf usage. In this work, we introduce WeSep, a toolkit designed for research and practical applications in TSE. WeSep is featured with flexible target speaker modeling, scalable data management, effective on-the-fly data simulation, structured recipes and deployment support. The toolkit is publicly avaliable at \url{https://github.com/wenet-e2e/WeSep.}

eess.AS↗

arXiv2023

Multi-pass Training and Cross-information Fusion for Low-resource End-to-end Accented Speech Recognition

Low-resource accented speech recognition is one of the important challenges faced by current ASR technology in practical applications. In this study, we propose a Conformer-based architecture, called Aformer, to leverage both the acoustic information from large non-accented and limited accented training data. Specifically, a general encoder and an accent encoder are designed in the Aformer to extract complementary acoustic information. Moreover, we propose to train the Aformer in a multi-pass manner, and investigate three cross-information fusion methods to effectively combine the information from both general and accent encoders. All experiments are conducted on both the accented English and Mandarin ASR tasks. Results show that our proposed methods outperform the strong Conformer baseline by relative 10.2% to 24.5% word/character error rate reduction on six in-domain and out-of-domain accented test sets.

cs.SD↗

arXiv2023

Rich structural polymorphism of monolayer C60 from cluster rotation

The recent experimental fabrication of monolayer and few-layer C60 polymers paves the way for synthesizing two-dimensional cluster-assembled materials. Compared to atoms with the SO(3) symmetry, clusters as superatoms (e.g., C60) have an additional rotational degree of freedom, greatly enriching the phase spaces of superatom-assembled materials. Using first-principles calculations, we find the energy barriers of cluster rotation in quasi-tetragonal monolayer C60 structures are rather low (about 10 meV/atom). The small rotational energy barriers lead to a series of tetragonal C60 polymorphs with energies that are close to the experimental quasi-tetragonal (expt-qT) phase. Similarly, several dynamically stable quasi-hexagonal monolayer C60 structures are found to have energies within 7 meV/atom above the experimental quasi-hexagonal phase. Our calculations demonstrate photo-excited electron-hole pairs and electrostatic doping of electrons can effectively modulate the relative energies of quasi-tetragonal C60 polymorphs. Particularly, the unstable monolayer expt-qT phase becomes dynamically stable when it is electrostatically doped with electrons. In contrast, the relative energies between different quasi-hexagonal polymorphs are insensitive to electrostatic doping of electrons.

cond-mat.mtrl-sci↗

arXiv2021

Highly efficient BiVO4 single-crystal nanosheets with dual modification: phosphorus doping and selective Ag modification

BiVO4, a visible-light response photocatalyst, has shown tremendous potential because of abundant raw material sources, good stability and low cost. There exist some limitations for further applicaitions due to poor capability to separate electron-hole pairs. In fact, a single-component modification strategy is barely adequate to obtain highy efficient photocatalytic performance. In this work, P substituted some of the V atoms from VO4 oxoanions, namely P was doped into the V sites in the host lattice of BiVO4 by a hydrothermal route. Meanwhile, Ag as an attractive and efficient electron-cocatalyst was selectively modified on the (010) facet of BiVO4 nanosheets via facile photo-deposition. As a result, the obtained dually modified BiVO4 sheets exhibited enhanced photocatalytic degradation property of methylene blue (MB). In detail, photocatalytic rate constant (k) was 2.285 min-1g-1, which was 2.78 times higher than pristine BiVO4 nanosheets. Actually, P-doping favored the formation of O vacancies, led to more charge carriers, and facilitated photocatalytic reaction. On the other hand, metallic Ag loaded on (010) facet effectively transferred photogenerated electrons, which consequently helped electron-hole pairs separation. The present work may enlighten new thoughts for smart design and controllable synthesis of highly efficient photocatalytic materials.

physics.app-ph↗

arXiv2017

Pressure-induced metallization and superconducting phase in ReS2

Among the family of TMDs, ReS2 takes a special position, which crystalizes in a unique distorted low-symmetry structure at ambient conditions. The interlayer interaction in ReS2 is rather weak, thus its bulk properties are similar to that of monolayer. However, how does compression change its structure and electronic properties is unknown so far. Here using ab initio crystal structure searching techniques, we explore the high-pressure phase transitions of ReS2 extensively and predict two new high-pressure phases. The ambient pressure phase transforms to a "distorted-1T" structure at very low pressure and then to a tetragonal I41/amd structure at around 90 GPa. The "distorted-1T" structure undergoes a semiconductor-metal transition (SMT) at around 70 GPa with a band overlap mechanism. Electron-phonon calculations suggest that the I41/amd structure is superconducting and has a critical superconducting temperature of about 2 K at 100 GPa. We further perform high-pressure electrical resistance measurements up to 102 GPa. Our experiments confirm the SMT and the superconducting phase transition of ReS2 under high pressure. These experimental results are in good agreement with our theoretical predictions.

cond-mat.mtrl-sci↗

arXiv2017

Topological Dirac line nodes and superconductivity coexist in SnSe at high pressure

Due to fundamental interest and potential applications in quantum computation, tremendous efforts have been invested to study topological superconductivity. However, bulk topological superconductivity seems to be difficult to realize and its mechanism is still elusive. Several possible routes to induce topological superconductivity have been proposed, including proximity efforts, doping or pressurizing a topological insulator or semimetal. Among them, the pressurizing is considered to be a clean way to tune the electronic structures. Here we report the discovery of a pressure-induced topological and superconducting phase of SnSe, a material which is highly focused recently due to its superior thermoelectric properties. In situ high-pressure electrical transport and synchrotron X-ray diffraction measurements show that the superconductivity emerges along with the formation of a CsCl-type structural symmetry of SnSe above around 27 GPa, with a maximum critical temperature of 3.2 K at 39 GPa. Based on ab initio calculations, this CsCl-type SnSe is predicted to be a Dirac line nodes (DLN) semimetal in the absence of spin-orbit coupling, whose DLN states are protected by the coexistence of time-reversal and inversion symmetries. These results make CsCl-type SnSe an interesting model platform with simple crystal symmetry to study the interplay of topological physics and superconductivity.

cond-mat.mtrl-sci↗

arXiv2016

Superconductivity Induced by High Pressure in Weyl Semimetal TaP

Weyl semimetal defines a material with three dimensional Dirac cones which appear in pair due to the breaking of spatial inversion or time reversal symmetry. Superconductivity is the state of quantum condensation of paired electrons. Turning a Weyl semimetal into superconducting state is very important in having some unprecedented discoveries. In this work, by doing resistive measurements on a recently recognized Weyl semimetal TaP under pressure up to about 100 GPa, we observe superconductivity at about 70 GPa. The superconductivity retains when the pressure is released. The systematic evolutions of resistivity and magnetoresistance with pressure are well interpreted by the relative shift between the chemical potential and paired Weyl points. Calculations based on the density functional theory also illustrate the structure transition at about 70GPa, the phase at higher pressure may host superconductivity. Our discovery of superconductivity in TaP by pressure will stimulate further study on superconductivity in Weyl semimetals.

cond-mat.supr-con↗

arXiv2016

Pressure induced re-emergence of superconductivity in superconducting topological insulator Sr0.065Bi2Se3

The recent-discovered Sr$_x$Bi$_2$Se$_3$ superconductor provides an alternative and ideal material base for investigating possible topological superconductivity. Here, we report that in Sr$_{0.065}$Bi$_{2}$Se$_3$, the ambient superconducting phase is gradually depressed upon the application of external pressure. At high pressure, a second superconducting phase emerges at above 6 GPa, with a maximum $T_c$ value of $\sim$8.3 K. The joint investigations of the high-pressure synchrotron x-ray diffraction and electrical transport properties reveal that the re-emergence of superconductivity in Sr$_{0.065}$Bi$_{2}$Se$_3$ is closely related to the structural phase transition from ambient rhombohedral phase to high-pressure monoclinic phase around 6 GPa, and further to another high-pressure tetragonal phase above 25 GPa.

cond-mat.supr-con↗