arXiv subjects

Xiaoyan Li

Publications and source records attributed to Xiaoyan Li.

At least 19 recordsLinked to original sources

arXiv2026

Universal Defenses for Tool-Integrated LLM Agents Against Adversarial Attacks

Large Language Model (LLM) agents have demonstrated impressive capabilities across a variety of domains, particularly when integrated with external tools for multi-step task completion. However, they are increasingly vulnerable to adversarial attacks, including direct prompt injection, indirect prompt injection, memory poisoning, and backdoor attacks, which exploit the model's openness to prompt injection and tool manipulation. In this work, we explore practical and generalizable defense strategies within a unified framework across these four attack types. We introduce two universal tool-based defenses: Attacker Tool Filtering, which uses anomaly detection (e.g., Isolation Forest) to identify and remove suspicious tools, and Normal Tool Recalling, a white-box method that restores the agent's original toolset prior to planning. Additionally, we incorporate prompt-based defenses: Chain-of-Thought prompting and self-reflection techniques to enhance reasoning and task paraphrasing to mitigate attacks. Experimental results across both four open-source LLMs (Gemma2-9B, Qwen2-7B, LLaMA3-8B, and LLaMA3.1-8B) and three proprietary LLMs (GPT-3.5, GPT-4, and GPT-5) show that our methods significantly reduce the Attack Success Rates (ASR), achieving 0% ASR in many settings, while preserving or even improving the original task success rate. These findings highlight the promise of simple, modular, multi-layered defenses for strengthening the security and robustness of tool-integrated LLM agents. The code is available at https://github.com/Xiaoyan-Lisa/Defenses-for-Tool-Integrated-LLM-Agents-Against-Adversarial-Attacks.

cs.CR↗

arXiv2026

Interpretable Fundus Image Classification via Ring-Based Retinal Vasculature Features

Retinal fundus photography is widely used for screening and monitoring ocular diseases, but many modern classification pipelines rely on deep latent representations and provide limited interpretability. This study develops an interpretable fundus image classification framework based on a ring-structured representation of the retinal vasculature centered on the optic disc. The method quantifies vessel geometry, color appearance, oxygenation-related vascular appearance, and vessel--background entropy within concentric retinal regions. These physiologically motivated descriptors are derived from vessel masks, image intensities, and optical-density measurements and aggregated across rings to capture spatial variation in vascular properties. Using only quantitative vascular descriptors, the proposed method achieved strong classification performance across three public fundus datasets. On HRF, it achieved 91.1\% accuracy using automatically generated vessel masks, matching RETFound, a vision transformer pretrained on large-scale retinal fundus image data, under the same evaluation setting. Additional analyses suggest that pretrained image models are sensitive to acquisition-related spatial cues, including fundus scale and retinal position within the field of view, as well as broader non-vessel image characteristics. This framework may support interpretable disease classification, quantitative retinal phenotyping, and retinal biomarker discovery without requiring large task-specific training datasets.

cs.CV↗

arXiv2026

Radiology-Report Semantic Modelling and Host-Response Laboratory Biomarkers for Multimodal Survival Prediction in Lung Cancer

TNM staging is essential for lung cancer management, but patients within the same anatomic stage often show heterogeneous survival outcomes. We developed a multimodal adaptive risk score (AMRS) that integrates radiology-report semantics with routinely available clinical laboratory biomarkers. In a retrospective two-center cohort, 1129 patients diagnosed between December 2017 and February 2026 were screened; 574 patients were included after exclusion for short follow-up or missing imaging reports and were split into training (n = 459) and test (n = 115) cohorts. Radiology reports were encoded with a domain-adapted MC-BERT branch to capture imaging-derived semantic information, while clinical and laboratory variables were modeled after Mahalanobis-distance-based imputation using random survival forests. Weighted risk fusion generated the final patient-level score. AMRS achieved C-index values of 0.920 in training and 0.849 in testing, and separated survival trajectories across clinical subgroups and TNM-related strata. SHAP analysis identified hematologic, inflammatory, coagulation, nutritional, tumor-marker, organ-function, and age-related contributors. AMRS may complement TNM staging in imaging-centered oncology workflows, but prospective validation, calibration, ablation testing, and clinical-utility assessment are required before deployment.

physics.med-ph↗

arXiv2026

RPC-Bench: A Fine-grained Benchmark for Research Paper Comprehension

Understanding research papers remains challenging for foundation models due to specialized scientific discourse and complex figures and tables, yet existing benchmarks offer limited fine-grained evaluation at scale. To address this gap, we introduce RPC-Bench, a large-scale question-answering benchmark built from review-rebuttal exchanges of high-quality computer science papers, containing 15K human-verified QA pairs. We design a fine-grained taxonomy aligned with the scientific research flow to assess models' ability to understand and answer why, what, and how questions in scholarly contexts. We also define an elaborate LLM-human interaction annotation framework to support large-scale labeling and quality control. Following the LLM-as-a-Judge paradigm, we develop a scalable framework that evaluates models on correctness-completeness and conciseness, with high agreement to human judgment. Experiments reveal that even the strongest models (GPT-5) achieve only 68.2% correctness-completeness, dropping to 37.46% after conciseness adjustment, highlighting substantial gaps in precise academic paper understanding. Our code and data are available at https://rpc-bench.github.io/.

cs.CL↗

arXiv2026

TagaVLM: Topology-Aware Global Action Reasoning for Vision-Language Navigation

Vision-Language Navigation (VLN) presents a unique challenge for Large Vision-Language Models (VLMs) due to their inherent architectural mismatch: VLMs are primarily pretrained on static, disembodied vision-language tasks, which fundamentally clash with the dynamic, embodied, and spatially-structured nature of navigation. Existing large-model-based methods often resort to converting rich visual and spatial information into text, forcing models to implicitly infer complex visual-topological relationships or limiting their global action capabilities. To bridge this gap, we propose TagaVLM (Topology-Aware Global Action reasoning), an end-to-end framework that explicitly injects topological structures into the VLM backbone. To introduce topological edge information, Spatial Topology Aware Residual Attention (STAR-Att) directly integrates it into the VLM's self-attention mechanism, enabling intrinsic spatial reasoning while preserving pretrained knowledge. To enhance topological node information, an Interleaved Navigation Prompt strengthens node-level visual-text alignment. Finally, with the embedded topological graph, the model is capable of global action reasoning, allowing for robust path correction. On the R2R benchmark, TagaVLM achieves state-of-the-art performance among large-model-based methods, with a Success Rate (SR) of 51.09% and SPL of 47.18 in unseen environments, outperforming prior work by 3.39% in SR and 9.08 in SPL. This demonstrates that, for embodied spatial reasoning, targeted enhancements on smaller open-source VLMs can be more effective than brute-force model scaling. The code can be found on our project page: https://apex-bjut.github.io/Taga-VLM

cs.CV↗

arXiv2026

SegTTA: Training-Free Test-Time Augmentation for Zero-Shot Medical Imaging Segmentation

Increasingly advanced data augmentation techniques have greatly aided clinical medical research, increasing data diversity and improving model generalization capabilities. Although most current basic models exhibit strong generalization abilities, image quality varies due to differences in equipment and operators. To address these challenges, we present SegTTA, a framework that improves medical image segmentation without model retraining by combining four augmentations (Gamma correction, Contrast enhancement, Gaussian blur, Gaussian noise) with weighted voting across multiple MedSAM2 checkpoints. Experiments demonstrate consistent improvements across three diverse datasets: healthy uterus segmentation, uterine myoma detection, and multi class hepatic structure segmentation. Ablation studies reveal that large organs benefit from intensity augmentations while small lesions require noise augmentations. The voting threshold controls the coverage precision trade off, enabling task specific optimization for different clinical requirements. Ultimately, on a multiclass hepatic vessel dataset, compared to MedSAM2 baselines, our method achieves an increase of 1.6 in mIoU and 1.9 in aIoU, along with a reduction of approximately 2.0 in HD95. Code will be available at https://github.com/AIGeeksGroup/SegTTA.

cs.CV↗

arXiv2026

Optically Activated Superconductivity in MgB2 via Electroluminescent GaP Inhomogeneous Phase

Experimental results demonstrate a viable strategy for tuning the superconducting properties of MgB2 through the incorporation of an electroluminescent inhomogeneous phase, revealing an interfacial light-phonon-electron synergistic mechanism that enhances superconductivity in conventional phonon-mediated systems. By introducing GaP electroluminescent inhomogeneous phases into MgB2 and activating their emission in situ through the application of a bias current during measurements, it is experimentally observed that the localized optical field and electromagnetic near field generated at the interface can effectively couple with the E2g phonon mode of the Mg-B layers, thereby significantly enhancing the electron-phonon interaction. As the emission intensity of the inhomogeneous phase increases, the interface light-field-driven mechanism markedly enhances the electron-phonon coupling constant lambda and leads to a gradual increase in the superconducting transition temperature Tc (with a maximum enhancement of approximately 1.4 K), enabling a tunable enhancement of the superconducting pairing channel in MgB2 without altering its primary chemical composition. In addition, the nanoscale dispersed distribution of the GaP inhomogeneous phase is expected to induce fine-scale defects that act as effective pinning centers and promote densification, resulting in an increase of the critical current density by approximately 69% at 20 K in the self-field and an enhancement of Hirr by about 31.5%. These results indicate that the electroluminescent inhomogeneous phase can synergistically enhance the superconducting performance of MgB2 through two mechanisms: "in situ near-field-enhanced pairing" and "structural pinning-assisted flux optimization", thereby providing a new design strategy for constructing superconducting material systems that can be activated by internal optical fields.

cond-mat.supr-con↗

arXiv2025

Synthesis of State-Attack Strategies for Anonymity and Opacity Violation in Discrete Event Systems

Attacks, including the manipulation of sensor readings and the modification of actuator commands, pose a significant challenge to the security and privacy of automated systems. This paper considers discrete event systems that can be modeled with nondeterministic finite state automata that are susceptible to state attacks. A state attack allows an intruder to learn whether or not the current state of a system falls into certain subsets of states. The intruder has a limited total number of state attacks at its disposal, but can launch state attacks at arbitrary instants of its choosing. We are interested on violations of current-state anonymity (resp. opacity), i.e., situations where the intruder, based on the sequence of observations generated by the system and the outcome of any performed state attacks, can ascertain the exact current state of the system (resp. that the current state of the system definitely resides in a subset of secret states). When the system violates current-state anonymity (resp. opacity) under a bounded number of state attacks, a subsequent question is whether the intruder can design an attack strategy such that anonymity-violating (resp. opacity-violating) situations will always be reached. In this latter case, we also design an attack strategy that guarantees that the system will reach a violating situation regardless of system actions. We provide pertinent complexity analysis of the corresponding verification algorithms and examples to illustrate the proposed methods.

cs.FL↗

arXiv2025

Advancing Time-Resolved Spectroscopies with Custom Scanning Units and Event-Based Electron Detection

Direct electron detection is revolutionizing electron microscopy by offering lower noise, reduced point-spread function, and increased quantum efficiency. Among these advancements, the Timepix3 hybrid-pixel direct electron detector stands out for its unique ability to output temporal information about individual hits within its pixel array. Its event-based architecture enables data-driven detection, where individual events are immediately read out from the chip. While recent studies have demonstrated the potential of event-based detectors in various applications in the context of continuous-gun scanning transmission electron microscopes (STEM), the use of such detectors with standard scanning units remains underdeveloped. In this work, we present a custom-designed, Timepix3-compatible scanning unit specifically developed to leverage the advantages of event-based detection. We explore its performance in enabling spatially and temporally resolved experiments, as well as its seamless interfacing with other time-resolved instruments, such as pulsed lasers and electron beam blankers. Additionally, we examine the prospects for achieving enhanced temporal resolution in time-resolved experiments using continuous-gun electron microscopes, identifying key challenges and proposing solutions to improve performance. This combination of custom hardware with advanced detection technology promises to expand the capabilities of electron microscopy in both fundamental research and practical applications.

physics.ins-det↗

arXiv2025

Compositional System Dynamics: The Higher Mathematics Underlying System Dynamics Diagrams & Practice

This work establishes a robust mathematical foundation for compositional System Dynamics modeling, leveraging category theory to formalize and enhance the representation, analysis, and composition of system models. Here, System Dynamics diagrams, such as stock & flow diagrams, system structure diagrams, and causal loop diagrams, are formulated as categorical constructs, enabling scalable, transparent, and systematic reasoning. By encoding these diagrams as data using attributed C-sets and utilizing advanced categorical tools like structured cospans, pushouts, pullbacks, and functor mappings, the framework supports modular composition, stratification, and seamless mapping between syntax and semantics. The approach underwrites traditional practice with firm mathematical structure, facilitates the identification of certain forms of pathways and feedback loops, the detection of simple patterns within complex diagrams, common structure between diagrams, and structure-preserving mappings between diverse diagram types. Additionally, this framework supports alternative semantics, such as stochastic transition dynamics, extending beyond traditional ordinary differential equation (ODE) representations. Applications in compositional modeling, modularity, and team-based collaboration demonstrate the practical advantages of this advanced framework. Future directions include integrating dimensional annotations, supporting hybrid and agent-based modeling paradigms, and expanding the framework's applicability to global and local temporal reasoning through temporal sheaves. By revealing and formalizing the hidden mathematical structure of System Dynamics diagrams, this work empowers practitioners to tackle complex systems with clarity, scalability, and rigor.

eess.SY↗

arXiv2025

Luminescence-Induced Tunable Superconductivity in BSCCO via GaP Quantum Dots

The enhancement of superconducting properties in high-temperature copper-oxide superconductor B(P)SCCO remains a hot research topic in the field of superconducting materials. Building on previous research, here we introduce GaP quantum dots as an heterophase into the B(P)SCCO superconductor, aiming to enhance its superconductivity through the luminescent properties of GaP quantum dots. The experimental results demonstrate that the introduction of GaP quantum dots into B(P)SCCO generates significant tunable superconducting effects, leading to enhanced critical transition temperature (Tc), critical current density (Jc), and Meissner field (Hc) of B(P)SCCO with increasing luminescent intensity of the GaP quantum dots. The enhancement effect induced by GaP quantum dots exhibits a positive correlation with luminescent intensity, meaning samples with the addition of GaP quantum dots exhibiting higher luminescent intensity show elevated Tc, Jc, and Hc values. Unlike impurity effects, a distinct critical concentration dependency is observed. Notably, this GaP quantum dot modification strategy is not only effective in conventional superconductors but also applicable to high-temperature oxide superconductors.

cond-mat.supr-con↗

arXiv2025

Beyond Single-Channel: Multichannel Signal Imaging for PPG-to-ECG Reconstruction with Vision Transformers

Reconstructing ECG from PPG is a promising yet challenging task. While recent advancements in generative models have significantly improved ECG reconstruction, accurately capturing fine-grained waveform features remains a key challenge. To address this, we propose a novel PPG-to-ECG reconstruction method that leverages a Vision Transformer (ViT) as the core network. Unlike conventional approaches that rely on single-channel PPG, our method employs a four-channel signal image representation, incorporating the original PPG, its first-order difference, second-order difference, and area under the curve. This multi-channel design enriches feature extraction by preserving both temporal and physiological variations within the PPG. By leveraging the self-attention mechanism in ViT, our approach effectively captures both inter-beat and intra-beat dependencies, leading to more robust and accurate ECG reconstruction. Experimental results demonstrate that our method consistently outperforms existing 1D convolution-based approaches, achieving up to 29% reduction in PRD and 15% reduction in RMSE. The proposed approach also produces improvements in other evaluation metrics, highlighting its robustness and effectiveness in reconstructing ECG signals. Furthermore, to ensure a clinically relevant evaluation, we introduce new performance metrics, including QRS area error, PR interval error, RT interval error, and RT amplitude difference error. Our findings suggest that integrating a four-channel signal image representation with the self-attention mechanism of ViT enables more effective extraction of informative PPG features and improved modeling of beat-to-beat variations for PPG-to-ECG mapping. Beyond demonstrating the potential of PPG as a viable alternative for heart activity monitoring, our approach opens new avenues for cyclic signal analysis and prediction.

eess.IV↗

arXiv2025

Heterogeneous Secure Transmissions in IRS-Assisted NOMA Communications: CO-GNN Approach

Intelligent Reflecting Surfaces (IRS) enhance spectral efficiency by adjusting reflection phase shifts, while Non-Orthogonal Multiple Access (NOMA) increases system capacity. Consequently, IRS-assisted NOMA communications have garnered significant research interest. However, the passive nature of the IRS, lacking authentication and security protocols, makes these systems vulnerable to external eavesdropping due to the openness of electromagnetic signal propagation and reflection. NOMA's inherent multi-user signal superposition also introduces internal eavesdropping risks during user pairing. This paper investigates secure transmissions in IRS-assisted NOMA systems with heterogeneous resource configuration in wireless networks to mitigate both external and internal eavesdropping. To maximize the sum secrecy rate of legitimate users, we propose a combinatorial optimization graph neural network (CO-GNN) approach to jointly optimize beamforming at the base station, power allocation of NOMA users, and phase shifts of IRS for dynamic heterogeneous resource allocation, thereby enabling the design of dual-link or multi-link secure transmissions in the presence of eavesdroppers on the same or heterogeneous links. The CO-GNN algorithm simplifies the complex mathematical problem-solving process, eliminates the need for channel estimation, and enhances scalability. Simulation results demonstrate that the proposed algorithm significantly enhances the secure transmission performance of the system.

cs.CR↗

arXiv2025

CLEP-GAN: An Innovative Approach to Subject-Independent ECG Reconstruction from PPG Signals

This study addresses the challenge of reconstructing unseen ECG signals from PPG signals, a critical task for non-invasive cardiac monitoring. While numerous public ECG-PPG datasets are available, they lack the diversity seen in image datasets, and data collection processes often introduce noise, complicating ECG reconstruction from PPG even with advanced machine learning models. To tackle these challenges, we first introduce a novel synthetic ECG-PPG data generation technique using an ODE model to enhance training diversity. Next, we develop a novel subject-independent PPG-to-ECG reconstruction model that integrates contrastive learning, adversarial learning, and attention gating, achieving results comparable to or even surpassing existing approaches for unseen ECG reconstruction. Finally, we examine factors such as sex and age that impact reconstruction accuracy, emphasizing the importance of considering demographic diversity during model training and dataset augmentation.

eess.SP↗

arXiv2024

Nanosecond nanothermometry in an electron microscope

Thermal transport in nanostructures plays a critical role in modern technologies. As devices shrink, techniques that can measure thermal properties at nanometer and nanosecond scales are increasingly needed to capture transient, out-of-equilibrium phenomena. We present a novel pump-probe photon-electron method within a scanning transmission electron microscope (STEM) to map temperature dynamics with unprecedented spatial and temporal resolutions. By combining focused laser-induced heating and synchronized time-resolved monochromated electron energy loss spectroscopy (EELS), we track phonon, exciton and plasmon signals in various materials, including silicon nitride, aluminum thin film, and transition metal dichalcogenides. Our results demonstrate the technique's ability to follow temperature changes at the nanometer and nanosecond scales. The experimental data closely matched theoretical heat diffusion models, confirming the method's validity. This approach opens new opportunities to investigate transient thermal phenomena in nanoscale materials, offering valuable insights for applications in thermoelectric devices and nanoelectronics.

cond-mat.mtrl-sci↗

arXiv2024

Polarization boost and ferroelectricity down to one unit cell in layered Carpy-Galy La$_{2}$Ti$_{2}$O$_{7}$ thin films

Layered perovskite-based compounds offer a range of unconventional properties enabled by their naturally anisotropic structure. While most renowned for the superconductivity observed in the Ruddlesden-Popper phases, many of these layered compounds are also ferroelectric and exhibit a sizeable in-plane polarization. Among these, the Carpy-Galy phases (A${_n}$B${_n}$O$_{3n+2}$), characterized by 110-oriented perovskite planes interleaved with additional oxygen layers, have been debated as platforms for hosting not only a robust polarization but also multiferroicity and polar metallicity. However, the challenges associated with the synthesis of ultrathin Carpy-Galy films and understanding the impact of strain on their properties limit their integration into devices. Addressing this issue, our study focuses on La$_2$Ti$_2$O$_7$, an $n$=4 (A$_2$B$_2$O$_7$) representative of the Carpy-Galy family, exploring its growth and concurrent phase stability on various substrates under different strain conditions. Remarkably, we demonstrate that a 3% tensile strain from DyScO$_3$ (100) substrates promotes a controlled layer-by-layer growth mode, while SrTiO$_3$ (110) and LaAlO$_3$-Sr$_2$TaAlO$_6$ (110), that exert negligible and compressive strains respectively, require post-deposition annealing to achieve similar results. Using scanning probe microscopy, X-ray diffraction, scanning transmission electron microscopy, and polarization switching experiments, we confirm that these films possess exceptional ferroelectric properties, including a polarization of 18 $μ$C/cm$^2$ - more than three times higher than previously reported - as well as persistence of ferroelectricity down to a single-unit-cell thickness. This study not only advances our understanding of Carpy-Galy phases in thin films but also lays a foundation for their application in advanced ferroelectric device architectures.

cond-mat.mtrl-sci↗

arXiv2024

Interlayer Engineering of Lattice Dynamics and Elastic Constants of 2D Layered Nanomaterials under Pressure

Interlayer coupling in two-dimensional (2D) layered nanomaterials can provide us novel strategies to evoke their superior properties, such as the exotic flat bands and unconventional superconductivity of twisted layers, the formation of moiré excitons and related nontrivial topology. However, to accurately quantify interlayer potential and further measure elastic properties of 2D materials remains vague, despite significant efforts. Herein, the layer-dependent lattice dynamics and elastic constants of 2D nanomaterials have been systematically investigated via pressure-engineering strategy based on ultralow frequency Raman spectroscopy. The shearing mode and layer-breathing Raman shifts of MoS2 with various thicknesses were analyzed by the linear chain model. Intriguingly, it was found that the layer-dependent dω/dP of shearing and breathing Raman modes display the opposite trends, quantitatively consistent with our molecular dynamics simulations and density functional theory calculations. These results can be generalized to other van der Waals systems, and may shed light on the potential applications of 2D materials in nanomechanics and nanoelectronics.

cond-mat.mtrl-sci↗

arXiv2024

OAG-Bench: A Human-Curated Benchmark for Academic Graph Mining

With the rapid proliferation of scientific literature, versatile academic knowledge services increasingly rely on comprehensive academic graph mining. Despite the availability of public academic graphs, benchmarks, and datasets, these resources often fall short in multi-aspect and fine-grained annotations, are constrained to specific task types and domains, or lack underlying real academic graphs. In this paper, we present OAG-Bench, a comprehensive, multi-aspect, and fine-grained human-curated benchmark based on the Open Academic Graph (OAG). OAG-Bench covers 10 tasks, 20 datasets, 70+ baselines, and 120+ experimental results to date. We propose new data annotation strategies for certain tasks and offer a suite of data pre-processing codes, algorithm implementations, and standardized evaluation protocols to facilitate academic graph mining. Extensive experiments reveal that even advanced algorithms like large language models (LLMs) encounter difficulties in addressing key challenges in certain tasks, such as paper source tracing and scholar profiling. We also introduce the Open Academic Graph Challenge (OAG-Challenge) to encourage community input and sharing. We envisage that OAG-Bench can serve as a common ground for the community to evaluate and compare algorithms in academic graph mining, thereby accelerating algorithm development and advancement in this field. OAG-Bench is accessible at https://www.aminer.cn/data/.

cs.DL↗