Searcharxiv⌕ Search

arXiv subjects

Ping He

Publications and source records attributed to Ping He.

At least 37 records · Page 2Linked to original sources
arXiv2026

FraudShield: Knowledge Graph Empowered Defense for LLMs against Fraud Attacks

Large language models (LLMs) have been widely integrated into critical automated workflows, including contract review and job application processes. However, LLMs are susceptible to manipulation by fraudulent information, which can lead to harmful outcomes. Although advanced defense methods have been developed to address this issue, they often exhibit limitations in effectiveness, interpretability, and generalizability, particularly when applied to LLM-based applications. To address these challenges, we introduce FraudShield, a novel framework designed to protect LLMs from fraudulent content by leveraging a comprehensive analysis of fraud tactics. Specifically, FraudShield constructs and refines a fraud tactic-keyword knowledge graph to capture high-confidence associations between suspicious text and fraud techniques. The structured knowledge graph augments the original input by highlighting keywords and providing supporting evidence, guiding the LLM toward more secure responses. Extensive experiments show that FraudShield consistently outperforms state-of-the-art defenses across four mainstream LLMs and five representative fraud types, while also offering interpretable clues for the model's generations.

cs.CR↗
arXiv2026

Assimilating rough features: A data-driven framework to infer rough wall properties from sparse experimental data

Surface roughness influences turbulent boundary layers (TBLs) primarily through the roughness function $ΔU^+$ and the equivalent sand-grain roughness height \(k_s\). Direct determination of \(k_s\) typically requires detailed velocity and wall-shear stress measurements, which are often impractical. As an alternative, this study presents a data assimilation framework that modifies a smooth-wall Reynolds-Averaged Navier-Stokes (RANS) baseline to match sparse rough-wall particle image velocimetry (PIV) data in the fully rough regime. Through this approach, secondary variables such as the friction velocity, \(u_τ\), and \(k_s\) can be inferred from the assimilated flow fields. The assimilated TBL reproduces experimental velocity profiles within 1\% and predicts friction velocity within 1-6\% of the experimental measurements. Furthermore, the \(k_s\) values inferred from the assimilation also match the experimental data up to 1\%. These results demonstrate the potential of data assimilation as a cost-effective alternative to high-fidelity methods and support the generalisation of the framework to model streamwise-varying roughness by treating \(k_s\) as a function of fetch length.

physics.flu-dyn↗
arXiv2026

HogVul: Black-box Adversarial Code Generation Framework Against LM-based Vulnerability Detectors

Recent advances in software vulnerability detection have been driven by Language Model (LM)-based approaches. However, these models remain vulnerable to adversarial attacks that exploit lexical and syntax perturbations, allowing critical flaws to evade detection. Existing black-box attacks on LM-based vulnerability detectors primarily rely on isolated perturbation strategies, limiting their ability to efficiently explore the adversarial code space for optimal perturbations. To bridge this gap, we propose HogVul, a black-box adversarial code generation framework that integrates both lexical and syntax perturbations under a unified dual-channel optimization strategy driven by Particle Swarm Optimization (PSO). By systematically coordinating two-level perturbations, HogVul effectively expands the search space for adversarial examples, enhancing the attack efficacy. Extensive experiments on four benchmark datasets demonstrate that HogVul achieves an average attack success rate improvement of 26.05\% over state-of-the-art baseline methods. These findings highlight the potential of hybrid optimization strategies in exposing model vulnerabilities.

cs.CR↗
arXiv2026

The dynamical and thermodynamic effects of turbulence on the cosmic baryonic fluid

Both simulations and observations indicate that the so-called missing baryons reside in the intergalactic medium known as the warm-hot intergalactic medium (WHIM). In this study we employed the IllustrisTNG50-1 simulation to demonstrate that knowledge of the turbulence in the cosmic baryonic fluid is crucial for correctly understanding both the spatial distribution and the physical origins of the missing baryons in the Universe. First, we find that dynamical effects cause the gas to be detained in low-density and intermediate-density regions, resulting in high baryon fractions, and prevent the convergence of the gas in high-density regions, leading to low baryon fractions. Second, turbulent energy is converted into thermal energy, and the injection and dissipation of turbulent energy have essentially reached a balance from $z=1$ to $0$. This indicates that the cosmic fluid is in a steady state within this redshift range. Due to turbulent heating, as the redshift decreases, an increasing amount of warm gas is heated and converted into the WHIM, and some even into hot gas. We find that, compared with turbulence in the cosmic fluid, shocks are unimportant in intermediate-density regions and even negligible in high-density regions, both dynamically and thermodynamically. This finding accounts for the origin of the WHIM in terms of both dynamics and thermodynamics, calls into question the traditional view of shock-heating, and highlights the importance of turbulence in shaping the large-scale structure of the Universe, particularly in the evolution of galaxies and galaxy clusters. In addition to TNG50-1, we validated our key findings with TNG50-2, TNG100-1, WIGEON, and EAGLE simulations, demonstrating that the spatial resolution, box size, and sub-grid-physics variations do not affect our main conclusions.

astro-ph.CO↗
arXiv2025

Symmetric Dicke States as Optimal Probes for Wave-Like Dark Matter

We identify symmetric Dicke states as the optimal quantum probes for distributed sensing of wave-like dark-matter fields. Within an ensemble-averaged quantum-metrological framework that incorporates the field's random phases and finite coherence, they maximize the Fisher information for short-baseline arrays with $N_d$ sensors and realize a robust $N_d^2$ enhancement. They also retain this collective advantage under amplitude-damping noise, whereas GHZ-type probes are highly fragile and rapidly lose their sensitivity once such noise is included. For two sensors at separations comparable to the dark-matter coherence length, the optimal entangled state acquires an additional spatial-correlation phase and outperforms both Dicke and independent probes. Our framework applies broadly to stochastic bosonic fields, including gravitational waves, and can be implemented with superconducting qubits, atomic ensembles, and NV centers.

hep-ph↗
arXiv2025

Long memory score-driven models as approximations for rough Ornstein-Uhlenbeck processes

This paper investigates the continuous-time limit of score-driven models with long memory. By extending score-driven models to incorporate infinite-lag structures with coefficients exhibiting heavy-tailed decay, we establish their weak convergence, under appropriate scaling, to fractional Ornstein-Uhlenbeck processes with Hurst parameter $H < 1/2$. When score-driven models are used to characterize the dynamics of volatility, they serve as discrete-time approximations for rough volatility. We present several examples, including EGARCH($\infty$) whose limits give rise to a new class of rough volatility models. Building on this framework, we carry out numerical simulations and option pricing analyses, offering new tools for rough volatility modeling and simulation.

math.PR↗
arXiv2025

LUNA: Linear Universal Neural Attention with Generalization Guarantees

Scaling attention faces a critical bottleneck: the $\mathcal{O}(n^2)$ quadratic computational cost of softmax attention, which limits its application in long-sequence domains. While linear attention mechanisms reduce this cost to $\mathcal{O}(n)$, they typically rely on fixed random feature maps, such as random Fourier features or hand-crafted functions. This reliance on static, data-agnostic kernels creates a fundamental trade-off, forcing practitioners to sacrifice significant model accuracy for computational efficiency. We introduce \textsc{LUNA}, a kernelized linear attention mechanism that eliminates this trade-off, retaining linear cost while matching and surpassing the accuracy of quadratic attention. \textsc{LUNA} is built on the key insight that the kernel feature map itself should be learned rather than fixed a priori. By parameterizing the kernel, \textsc{LUNA} learns a feature basis tailored to the specific data and task, overcoming the expressive limitations of fixed-feature methods. \textsc{Luna} implements this with a learnable feature map that induces a positive-definite kernel and admits a streaming form, yielding linear time and memory scaling in the sequence length. Empirical evaluations validate our approach across diverse settings. On the Long Range Arena (LRA), \textsc{Luna} achieves state-of-the-art average accuracy among efficient Transformers under compute parity, using the same parameter count, training steps, and approximate FLOPs. \textsc{Luna} also excels at post-hoc conversion: replacing softmax in fine-tuned BERT and ViT-B/16 checkpoints and briefly fine-tuning recovers most of the original performance, substantially outperforming fixed linearizations.

cs.LG↗
arXiv2025

Field Inversion Machine Learning for Time-Resolved Unsteady Flows in Airfoil Dynamic Stall

While many existing machine learning studies have focused on augmenting Reynolds averaged Navier Stokes (RANS) turbulence models for steady or time averaged unsteady flows, this paper takes a first step toward extending such augmentation to time resolved unsteady flows. An unsteady field inversion and machine learning (FIML) method is developed, in which a temporally evolving correction field (beta) is incorporated into the production term of a RANS turbulence model. The inverse problem is solved by optimizing the spatial temporal distribution of beta to minimize the regularized prediction errors. The resulting optimized beta field is then used to train a multi layer neural network that learns the time dependent relationship between local flow features and beta. The approach is demonstrated using the unsteady flow over a NACA0012 airfoil undergoing dynamic stall. Results show that the unsteady FIML model, trained using only the time series of drag data at a given pitch rate, can accurately reproduce the spatial temporal evolution of reference drag, lift, pitching moment, surface pressure, and velocity fields at both identical and different pitch rates. The unsteady FIML is integrated into the open source DAFoam framework, enabling a pathway toward developing accurate and generalizable RANS turbulence models for time resolved unsteady flows.

physics.flu-dyn↗
arXiv2025

Better Datasets Start From RefineLab: Automatic Optimization for High-Quality Dataset Refinement

High-quality Question-Answer (QA) datasets are foundational for reliable Large Language Model (LLM) evaluation, yet even expert-crafted datasets exhibit persistent gaps in domain coverage, misaligned difficulty distributions, and factual inconsistencies. The recent surge in generative model-powered datasets has compounded these quality challenges. In this work, we introduce RefineLab, the first LLM-driven framework that automatically refines raw QA textual data into high-quality datasets under a controllable token-budget constraint. RefineLab takes a set of target quality attributes (such as coverage and difficulty balance) as refinement objectives, and performs selective edits within a predefined token budget to ensure practicality and efficiency. In essence, RefineLab addresses a constrained optimization problem: improving the quality of QA samples as much as possible while respecting resource limitations. With a set of available refinement operations (e.g., rephrasing, distractor replacement), RefineLab takes as input the original dataset, a specified set of target quality dimensions, and a token budget, and determines which refinement operations should be applied to each QA sample. This process is guided by an assignment module that selects optimal refinement strategies to maximize overall dataset quality while adhering to the budget constraint. Experiments demonstrate that RefineLab consistently narrows divergence from expert datasets across coverage, difficulty alignment, factual fidelity, and distractor quality. RefineLab pioneers a scalable, customizable path to reproducible dataset design, with broad implications for LLM evaluation.

cs.CL↗
arXiv2025

Serre functor and torsion pairs

Given a torsion pair $(\mathcal{T},\mathcal{F})$ in an abelian category $\mathcal{A}$ and its Happel-Reiten-Smalø tilt $\mathcal{B}$, the equivalence of the realization functor $D^b({\mathcal B})\to D^b({\mathcal A})$ is determined by some properties of the torsion pair [9]. We call $(\mathcal{T},\mathcal{F})$ satisfying such a property effaceable. If $\mathcal{A}$ is an Ext-finite abelian category with Serre duality, we prove that $(\mathcal{T},\mathcal{F})$ is effaceable implies that $\mathcal{U}_{\mathcal T}$ is closed under Serre functor. Conversely, when $\mathcal A$ is the module category of a finite-dimensional hereditary algebra, we prove that the torsion pair $(\mathcal{T},\mathcal{F})$ is effaceable if and only if $\mathcal{U}_\mathcal{T}$ is closed under the Serre functor via a recollement of $D^b({\mathcal A})$.

math.RT↗
arXiv2025

Automatic Red Teaming LLM-based Agents with Model Context Protocol Tools

The remarkable capability of large language models (LLMs) has led to the wide application of LLM-based agents in various domains. To standardize interactions between LLM-based agents and their environments, model context protocol (MCP) tools have become the de facto standard and are now widely integrated into these agents. However, the incorporation of MCP tools introduces the risk of tool poisoning attacks, which can manipulate the behavior of LLM-based agents. Although previous studies have identified such vulnerabilities, their red teaming approaches have largely remained at the proof-of-concept stage, leaving the automatic and systematic red teaming of LLM-based agents under the MCP tool poisoning paradigm an open question. To bridge this gap, we propose AutoMalTool, an automated red teaming framework for LLM-based agents by generating malicious MCP tools. Our extensive evaluation shows that AutoMalTool effectively generates malicious MCP tools capable of manipulating the behavior of mainstream LLM-based agents while evading current detection mechanisms, thereby revealing new security risks in these agents.

cs.CR↗
arXiv2025

CLMTracing: Black-box User-level Watermarking for Code Language Model Tracing

With the widespread adoption of open-source code language models (code LMs), intellectual property (IP) protection has become an increasingly critical concern. While current watermarking techniques have the potential to identify the code LM to protect its IP, they have limitations when facing the more practical and complex demand, i.e., offering the individual user-level tracing in the black-box setting. This work presents CLMTracing, a black-box code LM watermarking framework employing the rule-based watermarks and utility-preserving injection method for user-level model tracing. CLMTracing further incorporates a parameter selection algorithm sensitive to the robust watermark and adversarial training to enhance the robustness against watermark removal attacks. Comprehensive evaluations demonstrate CLMTracing is effective across multiple state-of-the-art (SOTA) code LMs, showing significant harmless improvements compared to existing SOTA baselines and strong robustness against various removal attacks.

cs.PL↗
arXiv2025

Precision calculation of the EFT likelihood with primordial non-Gaussianities

We perform a precision calculation of the effective field theory (EFT) conditional likelihood for large-scale structure (LSS) using the saddle-point expansion method in the presence of primordial non-Gaussianities (PNG). The precision is manifested at two levels: one corresponding to the consideration of higher-order noise terms, and the other to the inclusion of contributions around the saddle points. In computing the latter, we encounter the same issue of the negative modes as in the context of false vacuum decay, which necessitates deforming the original integration contour into a combination of the steepest descent contours to ensure a convergent and real result. We demonstrate through detailed calculations that, upon incorporating leading-order PNG, both types of extensions introduce irreducible field-dependent contributions to the conditional likelihood. This insight motivates the systematic inclusion of additional effective terms within the forward modeling framework. Our work facilitates Bayesian forward modeling under non-Gaussian initial conditions, thereby enabling more stringent constraints on the parameters describing PNG.

astro-ph.CO↗
arXiv2025

ADAptation: Reconstruction-based Unsupervised Active Learning for Breast Ultrasound Diagnosis

Deep learning-based diagnostic models often suffer performance drops due to distribution shifts between training (source) and test (target) domains. Collecting and labeling sufficient target domain data for model retraining represents an optimal solution, yet is limited by time and scarce resources. Active learning (AL) offers an efficient approach to reduce annotation costs while maintaining performance, but struggles to handle the challenge posed by distribution variations across different datasets. In this study, we propose a novel unsupervised Active learning framework for Domain Adaptation, named ADAptation, which efficiently selects informative samples from multi-domain data pools under limited annotation budget. As a fundamental step, our method first utilizes the distribution homogenization capabilities of diffusion models to bridge cross-dataset gaps by translating target images into source-domain style. We then introduce two key innovations: (a) a hypersphere-constrained contrastive learning network for compact feature clustering, and (b) a dual-scoring mechanism that quantifies and balances sample uncertainty and representativeness. Extensive experiments on four breast ultrasound datasets (three public and one in-house/multi-center) across five common deep classifiers demonstrate that our method surpasses existing strong AL-based competitors, validating its effectiveness and generalization for clinical domain adaptation. The code is available at the anonymized link: https://github.com/miccai25-966/ADAptation.

cs.CV↗
arXiv2025

Statistical Inference for Cumulative INAR($\infty$) Processes via Least-Squares

This paper investigates the cumulative Integer-Valued Autoregressive model of infinite order, denoted as INAR($\infty$), a class of processes crucial for modeling count time series and equivalent to discrete-time Hawkes processes. We propose a computationally efficient conditional least-squares (CLS) estimator to address the challenge of parameter inference in this infinite-dimensional setting. We establish the key theoretical properties of the estimator, including its consistency and asymptotic normality. A central contribution is the rigorous treatment of its large-sample distribution in a framework where the parameter dimension grows with the sample size, for which we derive the corresponding sandwich-form covariance matrix. The theoretical results are substantiated through comprehensive Monte Carlo simulations. These experiments demonstrate that the estimator's accuracy and stability systematically improve as the sample size increases, confirming its consistency. Furthermore, we show that the estimator's finite-sample distribution is well-approximated by a normal distribution, and this approximation becomes more robust with larger samples. Our work provides a complete and practical framework for statistical inference in cumulative INAR($\infty$) models. The code to reproduce the numerical experiments is publicly available at https://github.com/gagawjbytw/INAR_estimation.

math.ST↗
arXiv2025

The continuous-time limit of quasi score-driven volatility models

This paper explores the continuous-time limit of a class of Quasi Score-Driven (QSD) models that characterize volatility. As the sampling frequency increases and the time interval tends to zero, the model weakly converges to a continuous-time stochastic volatility model where the two Brownian motions are correlated, thereby capturing the leverage effect in the market. Subsequently, we identify that a necessary condition for non-degenerate correlation is that the distribution of driving innovations differs from that of computing score, and at least one being asymmetric. We then illustrate this with two typical examples. As an application, the QSD model is used as an approximation for correlated stochastic volatility diffusions and quasi maximum likelihood estimation is performed. Simulation results confirm the method's effectiveness, particularly in estimating the correlation coefficient.

math.PR↗
arXiv2025

Closed-form solutions for VIX derivatives in a Legendre empirical model

In this paper, we introduce a data-driven, single-parameter Markov diffusion model for the VIX. The volatility factor evolves in $(-1,1)$ with a uniform invariant distribution ensured by Legendre polynomials, mapped to the empirical distribution. We derive analytical series solutions for VIX futures and options using separation of variables to solve the Feynman-Kac PDE. Compared to the 3/2 model, our approach offers equal or superior accuracy and flexibility, providing an efficient, robust alternative for VIX pricing and risk management. Code and data are available at github.com/gagawjbytw/empirical-VIX.

q-fin.PR↗