arXiv subjects

Xin Wang

Publications and source records attributed to Xin Wang.

At least 433 records · Page 24Linked to original sources

arXiv2026

AI-Based Vulnerability Analysis of NFT Smart Contracts

With the rapid growth of the NFT market, the security of smart contracts has become crucial. However, existing AI-based detection models for NFT contract vulnerabilities remain limited due to their complexity, while traditional manual methods are time-consuming and costly. This study proposes an AI-driven approach to detect vulnerabilities in NFT smart contracts. We collected 16,527 public smart contract codes, classifying them into five vulnerability categories: Risky Mutable Proxy, ERC-721 Reentrancy, Unlimited Minting, Missing Requirements, and Public Burn. Python-processed data was structured into training/test sets. Using the CART algorithm with Gini coefficient evaluation, we built initial decision trees for feature extraction. A random forest model was implemented to improve robustness through random data/feature sampling and multitree integration. GridSearch hyperparameter tuning further optimized the model, with 3D visualizations demonstrating parameter impacts on vulnerability detection. Results show the random forest model excels in detecting all five vulnerabilities. For example, it identifies Risky Mutable Proxy by analyzing authorization mechanisms and state modifications, while ERC-721 Reentrancy detection relies on external call locations and lock mechanisms. The ensemble approach effectively reduces single-tree overfitting, with stable performance improvements after parameter tuning. This method provides an efficient technical solution for automated NFT contract detection and lays groundwork for scaling AI applications.

cs.CR↗

arXiv2026

The Third VoicePrivacy Challenge: Preserving Emotional Expressiveness and Linguistic Content in Voice Anonymization

We present results and analyses from the third VoicePrivacy Challenge held in 2024, which focuses on advancing voice anonymization technologies. The task was to develop a voice anonymization system for speech data that conceals a speaker's voice identity while preserving linguistic content and emotional state. We provide a systematic overview of the challenge framework, including detailed descriptions of the anonymization task and datasets used for both system development and evaluation. We outline the attack model and objective evaluation metrics for assessing privacy protection (concealing speaker voice identity) and utility (content and emotional state preservation). We describe six baseline anonymization systems and summarize the innovative approaches developed by challenge participants. Finally, we provide key insights and observations to guide the design of future VoicePrivacy challenges and identify promising directions for voice anonymization research.

cs.CL↗

arXiv2026

A Safety Report on GPT-5.2, Gemini 3 Pro, Qwen3-VL, Grok 4.1 Fast, Nano Banana Pro, and Seedream 4.5

The rapid evolution of Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) has driven major gains in reasoning, perception, and generation across language and vision, yet whether these advances translate into comparable improvements in safety remains unclear, partly due to fragmented evaluations that focus on isolated modalities or threat models. In this report, we present an integrated safety evaluation of six frontier models--GPT-5.2, Gemini 3 Pro, Qwen3-VL, Grok 4.1 Fast, Nano Banana Pro, and Seedream 4.5--assessing each across language, vision-language, and image generation using a unified protocol that combines benchmark, adversarial, multilingual, and compliance evaluations. By aggregating results into safety leaderboards and model profiles, we reveal a highly uneven safety landscape: while GPT-5.2 demonstrates consistently strong and balanced performance, other models exhibit clear trade-offs across benchmark safety, adversarial robustness, multilingual generalization, and regulatory compliance. Despite strong results under standard benchmarks, all models remain highly vulnerable under adversarial testing, with worst-case safety rates dropping below 6%. Text-to-image models show slightly stronger alignment in regulated visual risk categories, yet remain fragile when faced with adversarial or semantically ambiguous prompts. Overall, these findings highlight that safety in frontier models is inherently multidimensional--shaped by modality, language, and evaluation design--underscoring the need for standardized, holistic safety assessments to better reflect real-world risk and guide responsible deployment.

cs.AI↗

arXiv2026

A Portrait of the Cosmic Reionisation History in the Context of the Early Dark Energy Model

Recent JWST observations of Lyman-$α$ emission at $z \sim 11-6$ indicate a rapid reionization of the intergalactic medium within the first $\sim700$ Myr. The required Lyman continuum (LyC) photon budget may naturally arise from the unexpectedly high galaxy number densities revealed by JWST, reducing the need for scenarios invoking very high LyC escape fractions ($f_{\rm esc}\gtrsim0.2$) or dominant contributions from ultra-faint galaxies ($M_{\rm UV}>-15$) in the standard $Λ$CDM framework. In this work, we model the reionization history under the Early Dark Energy (EDE) paradigm -- originally proposed to ease the Hubble tension -- which also explains the observed over-abundance of high-$z$ galaxies without extreme star formation efficiencies. The EDE model yields reionization histories consistent with current constraints while requiring only moderate LyC escape fractions and UV luminosity densities ($f_{\rm esc}\sim 0.05-0.1$, $M_{\rm UV}\lesssim -17$ to $-15$). Our results suggest that, once key astrophysical parameters are better constrained, the reionization history could serve as an independent and complementary probe of EDE cosmologies.

astro-ph.CO↗

arXiv2026

Discovery of a gas-enshrouded broad-line AGN at z $\sim$ 7

The Lyman-alpha (Ly$α$) absorption profile in star-forming galaxies serves as a powerful tracer of the extended, dense neutral hydrogen in their surroundings during the Epoch of Reionization (EoR). We report a unique galaxy, A2744-z7DLA, at $z\approx 6.87$ gravitationally lensed by the foreground galaxy cluster Abell 2744, which exhibits both moderate Ly$α$ emission and damped Ly$α$ absorption, suggesting the presence of a dense neutral hydrogen environment. Our analysis suggests that the UV continuum turnover near Ly$α$ is more likely shaped by a damped Ly$α$ system rather than nebular continuum from two photon process. We analyze the physical properties of A2744-z7DLA with spectroscopic and photometric data from the JWST and the HST. The galaxy shows a compact morphology ($r_e \sim 0.3\ {\rm kpc}$) and a broadened H$α$ emission line, suggesting possible AGN activity. The broad component of H$α$ has a FWHM of $2721 \pm 200\ {\rm km\ s^{-1}}$, corresponding to a black hole mass of $M_{\rm BH}=2.90^{+2.35}_{-1.28}\times 10^7 M_\odot$ and a black hole-to-stellar mass ratio of $\log (M_{\rm BH}/M_{\rm stellar}) = -1.58^{+0.45}_{-0.34}$. The Balmer decrement ($\rm Hα/Hβ$) yields a dust attenuation of $\rm A_V \approx 1.15 \pm 0.23$, indicating that this system is less dust-rich than some "little red dots". Furthermore, we perform SED fitting using both stellar and AGN models. The results show that the UV and optical wavelengths are dominated by star-forming regions, while the AGN component contributes primarily at longer wavelengths. This work provides new insights into the interplay between star formation, neutral gas, and potential AGN activity in galaxies during the EoR.

astro-ph.GA↗

arXiv2026

Simulated Ensemble Attack: Transferring Jailbreaks Across Fine-tuned Vision-Language Models

The widespread practice of fine-tuning open-source Vision-Language Models (VLMs) raises a critical security concern: jailbreak vulnerabilities in base models may persist in downstream variants, enabling transferable attacks across fine-tuned systems. To investigate this risk, we propose the Simulated Ensemble Attack (SEA), a grey-box jailbreak framework that assumes full access to the base VLM but no knowledge of the fine-tuned target. SEA enhances transferability via Fine-tuning Trajectory Simulation (FTS), which models bounded parameter variations in the vision encoder, and Targeted Prompt Guidance (TPG), which stabilizes adversarial optimization through auxiliary textual guidance. Experiments on the Qwen2-VL family demonstrate that SEA achieves consistently high transfer success and toxicity rates across diverse fine-tuned variants, including safety-enhanced models, while standard PGD-based image jailbreaks exhibit negligible transferability. Further analysis reveals that fine-tuning primarily induces localized parameter shifts around the base model, explaining why attacks optimized over a simulated neighborhood transfer effectively. We also show that SEA generalizes across different base generations (e.g., Qwen2.5/3-VL), indicating that its effectiveness arises from shared fine-tuning-induced behaviors rather than architecture- or initialization-specific factors.

cs.CV↗

arXiv2026

Diagnostic Performance of Universal-Learning Ultrasound AI Across Multiple Organs and Tasks: the UUSIC25 Challenge

IMPORTANCE: Modern ultrasound systems are universal diagnostic tools capable of imaging the entire body. However, current AI solutions remain fragmented into single-task tools. This critical gap between hardware versatility and software specificity limits workflow integration and clinical utility. OBJECTIVE: To evaluate the diagnostic accuracy, versatility, and efficiency of single general-purpose deep learning models for multi-organ classification and segmentation. DESIGN: The Universal UltraSound Image Challenge 2025 (UUSIC25) involved developing algorithms on 11,644 images aggregated from 12 sources (9 public, 3 private). Evaluation used an independent, multi-center private test set of 2,479 images, including data from a center completely unseen during training to assess generalization. OUTCOMES: Diagnostic performance (Dice Similarity Coefficient [DSC]; Area Under the Receiver Operating Characteristic Curve [AUC]) and computational efficiency (inference time, GPU memory). RESULTS: Of 15 valid algorithms, the top model (SMART) achieved a macro-averaged DSC of 0.854 across 5 segmentation tasks and AUC of 0.766 for binary classification. Models demonstrated high capability in anatomical segmentation (e.g., fetal head DSC: 0.942) but variability in complex diagnostic tasks subject to domain shift. Specifically, in breast cancer molecular subtyping, the top model's performance dropped from an AUC of 0.571 (internal) to 0.508 (unseen external center), highlighting the challenge of generalization. CONCLUSIONS: General-purpose AI models can achieve high accuracy and efficiency across multiple tasks using a single architecture. However, significant performance degradation on unseen data suggests domain generalization is critical for future clinical deployment.

cs.CV↗

arXiv2026

ShowUI-Aloha: Human-Taught GUI Agent

Graphical User Interfaces (GUIs) are central to human-computer interaction, yet automating complex GUI tasks remains a major challenge for autonomous agents, largely due to a lack of scalable, high-quality training data. While recordings of human demonstrations offer a rich data source, they are typically long, unstructured, and lack annotations, making them difficult for agents to learn from.To address this, we introduce ShowUI-Aloha, a comprehensive pipeline that transforms unstructured, in-the-wild human screen recordings from desktop environments into structured, actionable tasks. Our framework includes four key components: A recorder that captures screen video along with precise user interactions like mouse clicks, keystrokes, and scrolls. A learner that semantically interprets these raw interactions and the surrounding visual context, translating them into descriptive natural language captions. A planner that reads the parsed demonstrations, maintains task states, and dynamically formulates the next high-level action plan based on contextual reasoning. An executor that faithfully carries out these action plans at the OS level, performing precise clicks, drags, text inputs, and window operations with safety checks and real-time feedback. Together, these components provide a scalable solution for collecting and parsing real-world human data, demonstrating a viable path toward building general-purpose GUI agents that can learn effectively from simply observing humans.

cs.CV↗

arXiv2026

OpenRT: An Open-Source Red Teaming Framework for Multimodal LLMs

The rapid integration of Multimodal Large Language Models (MLLMs) into critical applications is increasingly hindered by persistent safety vulnerabilities. However, existing red-teaming benchmarks are often fragmented, limited to single-turn text interactions, and lack the scalability required for systematic evaluation. To address this, we introduce OpenRT, a unified, modular, and high-throughput red-teaming framework designed for comprehensive MLLM safety evaluation. At its core, OpenRT architects a paradigm shift in automated red-teaming by introducing an adversarial kernel that enables modular separation across five critical dimensions: model integration, dataset management, attack strategies, judging methods, and evaluation metrics. By standardizing attack interfaces, it decouples adversarial logic from a high-throughput asynchronous runtime, enabling systematic scaling across diverse models. Our framework integrates 37 diverse attack methodologies, spanning white-box gradients, multi-modal perturbations, and sophisticated multi-agent evolutionary strategies. Through an extensive empirical study on 20 advanced models (including GPT-5.2, Claude 4.5, and Gemini 3 Pro), we expose critical safety gaps: even frontier models fail to generalize across attack paradigms, with leading models exhibiting average Attack Success Rates as high as 49.14%. Notably, our findings reveal that reasoning models do not inherently possess superior robustness against complex, multi-turn jailbreaks. By open-sourcing OpenRT, we provide a sustainable, extensible, and continuously maintained infrastructure that accelerates the development and standardization of AI safety.

cs.CR↗

arXiv2026

LLM4Perf: Large Language Models Are Effective Samplers for Multi-Objective Performance Modeling

The performance of modern software systems is critically dependent on their complex configuration options. Building accurate performance models to navigate this vast space requires effective sampling strategies, yet existing methods often struggle with multi-objective optimization and cannot leverage semantic information from documentation. The recent success of Large Language Models (LLMs) motivates the central question of this work: Can LLMs serve as effective samplers for multi-objective performance modeling? To explore this, we present a comprehensive empirical study investigating the capabilities and characteristics of LLM-driven sampling. We design and implement LLM4Perf, a feedback-based framework, and use it to systematically evaluate the LLM-guided sampling process across four highly configurable, real-world systems. Our study reveals that the LLM-guided approach outperforms traditional baselines in most cases. Quantitatively, LLM4Perf achieves the best performance in nearly 68.8% (77 out of 112) of all evaluation scenarios, demonstrating its superior effectiveness. We find this effectiveness stems from the LLM's dual capabilities of configuration space pruning and feedback-driven strategy refinement. The effectiveness of this pruning is further validated by the fact that it also improves the performance of the baseline methods in nearly 91.5% (410 out of 448) of cases. Furthermore, we show how the LLM choices for each component and hyperparameters within LLM4Perf affect its effectiveness. Overall, this paper provides strong evidence for the effectiveness of LLMs in performance engineering and offers concrete insights into the mechanisms that drive their success.

cs.SE↗

arXiv2026

DVD: A Robust Method for Detecting Variant Contamination in Large Language Model Evaluation

Evaluating large language models (LLMs) is increasingly confounded by \emph{variant contamination}: the training corpus contains semantically equivalent yet lexically or syntactically altered versions of test items. Unlike verbatim leakage, these paraphrased or structurally transformed variants evade existing detectors based on sampling consistency or perplexity, thereby inflating benchmark scores via memorization rather than genuine reasoning. We formalize this problem and introduce \textbf{DVD} (\textbf{D}etection via \textbf{V}ariance of generation \textbf{D}istribution), a single-sample detector that models the local output distribution induced by temperature sampling. Our key insight is that contaminated items trigger alternation between a \emph{memory-adherence} state and a \emph{perturbation-drift} state, yielding abnormally high variance in the synthetic difficulty of low-probability tokens; uncontaminated items remain in drift with comparatively smooth variance. We construct the first benchmark for variant contamination across two domains Omni-MATH and SuperGPQA by generating and filtering semantically equivalent variants, and simulate contamination via fine-tuning models of different scales and architectures (Qwen2.5 and Llama3.1). Across datasets and models, \textbf{DVD} consistently outperforms perplexity-based, Min-$k$\%++, edit-distance (CDD), and embedding-similarity baselines, while exhibiting strong robustness to hyperparameters. Our results establish variance of the generation distribution as a principled and practical fingerprint for detecting variant contamination in LLM evaluation.

cs.AI↗

arXiv2026

Feasibility of a General-Purpose Deep Learning Dose Engine: A Multi-Site Validation Study

Conventional radiotherapy dose calculation algorithms are often computationally slow and non-differentiable, creating bottlenecks for online adaptive radiotherapy (ART) and limiting end-to-end automatic planning. Deep learning provides consistent inference performance and a differentiable framework essential for rapid optimization. In this study, we developed a generalized, site-independent deep learning dose engine using a beamlet-based input strategy. This establishes a computationally consistent and differentiable module that enables end-to-end training for autoplanning while maintaining accuracy across diverse geometries. A dataset of 3,600 plans from 120 patients across six anatomical sites was used to train two 3D convolutional neural networks, a standard U-Net and a Cascade U-Net, to predict 3D dose distributions from CT images and divergent MLC/jaw projections. Performance was validated via 3D gamma analysis on an independent cohort of 60 VMAT plans. The optimal model (U-Net with MAE loss) achieved a mean gamma passing rate of $98.9 \pm 1.6\%$ (3%/2mm, 10% threshold). Performance remained robust across all sites (passing rates $>98\%$), demonstrating that the beamlet-based strategy generalizes effectively to complex geometries without site-specific training. These results indicate that a single, site-independent model can calculate radiotherapy dose distributions with clinical accuracy. This differentiable engine is highly suitable for integration into end-to-end automatic planning, online ART, and secondary dose verification workflows.

physics.med-ph↗

arXiv2026

PALM-Bench: A Comprehensive Benchmark for Personalized Audio-Language Models

Large Audio-Language Models (LALMs) have demonstrated strong performance in audio understanding and generation. Yet, our extensive benchmarking reveals that their behavior is largely generic (e.g., summarizing spoken content) and fails to adequately support personalized question answering (e.g., summarizing what my best friend says). In contrast, human conditions their interpretation and decision-making on each individual's personal context. To bridge this gap, we formalize the task of Personalized LALMs (PALM) for recognizing personal concepts and reasoning within personal context. Moreover, we create the first benchmark (PALM-Bench) to foster the methodological advances in PALM and enable structured evaluation on several tasks across multi-speaker scenarios. Our extensive experiments on representative open-source LALMs, show that existing training-free prompting and supervised fine-tuning strategies, while yield improvements, remains limited in modeling personalized knowledge and transferring them across tasks robustly. Data and code will be released.

cs.CL↗

arXiv2026

BEACON: JWST NIRCam Pure-parallel Imaging Survey. II. Physical Properties of $z=7-14$ Galaxies

We present photometric properties of 161 galaxy candidates at $z=7-14$ selected from the second data release (DR2) of BEACON, a JWST Cycle 2 pure-parallel NIRCam imaging program. Carefully selected from 36 independent pointings (corresponding to $\sim350$\,arcmin$^2$ sky coverage), and hence with reduced cosmic variance, our galaxy candidates provide an unbiased sample for investigating galaxy properties over a wide range of environments. We measure the physical properties, including UV continuum slope ($β_{\rm UV}$), stellar mass ($M_*$), star formation rate (SFR), and sizes. Our highest redshift galaxy candidate at $z=13.71\pm0.15$ has a remarkably bright UV luminosity of $M_{\rm UV}=-21.19\pm0.08$, making it the brightest galaxy at $z>12$ if spectroscopically confirmed. With an extremely blue UV slope, compact morphology, and high star formation rate surface density ($Σ_{\rm SFR}$), this candidate may have extremely low metallicity, high ionizing photon escape fraction, or contributions from an AGN. Among our multiple independent sightlines, we identify three fields of galaxy number overdensity with $>3σ$ significance. The properties of galaxies in various environments do not exhibit significant differences, implying either that accelerated galaxy evolution in overdense regions is not yet widespread at $z>7$, or that the current constraints are limited by sample size. Our simulations indicate that increasing the sample by an order of magnitude would allow such environmental trends to be robustly confirmed or ruled out, underscoring the importance of future pure-parallel observations.

astro-ph.GA↗

arXiv2026

An Uncertainty-Aware Generalization Framework for Cardiovascular Image Segmentation

Deep learning models have achieved significant success in segmenting cardiovascular structures, but there is a growing need to improve their generalization and robustness. Current methods often face challenges such as overfitting and limited accuracy, largely due to their reliance on large annotated datasets and limited optimization techniques. This paper introduces the UU-Mamba model, an extension of the U-Mamba architecture, designed to address these challenges in both cardiac and vascular segmentation. By incorporating Sharpness-Aware Minimization (SAM), the model enhances generalization by seeking flatter minima in the loss landscape. Additionally, we propose an uncertainty-aware loss function that integrates region-based, distribution-based, and pixel-based components, improving segmentation accuracy by capturing both local and global features. We expand our evaluations on the ImageCAS (coronary artery) and Aorta (aortic branches and zones) datasets, which present more complex segmentation challenges than the ACDC dataset (left and right ventricles) used in prior work, showcasing the model's adaptability and resilience. Our results confirm UU-Mamba's superior performance compared to leading models such as TransUNet, Swin-Unet, nnUNet, and nnFormer. We also provide a more in-depth assessment of the model's robustness and segmentation accuracy through extensive experiments.

cs.AI↗

arXiv2026

Neural-network-based Self-triggered Observed Platoon Control for Autonomous Vehicles

This paper investigates autonomous vehicle (AV) platoon control under uncertain dynamics and intermittent communication, which remains a critical challenge in intelligent transportation systems. To address these issues, this paper proposes an adaptive consensus tracking control framework for nonlinear multi-agent systems (MASs). The proposed approach integrates backstepping design, a nonlinear sampled-data observer, radial basis function neural networks, and a self-triggered communication mechanism. The radial basis function neural networks approximate unknown nonlinearities and time-varying disturbances, thereby enhancing system robustness. A distributed observer estimates neighboring states based on limited and intermittent measurements, thereby reducing dependence on continuous communication. Moreover, self-triggered mechanism is developed to determine triggering instants, guaranteeing a strictly positive minimum inter-event time and preventing Zeno behavior. The theoretical analysis proves that all closed-loop signals are uniformly ultimately bounded (UUB), and tracking errors converge to a compact set. Simulation results demonstrate that the proposed approach achieves high robustness, adaptability, and communication efficiency, making it suitable for real-world networked vehicle systems.

eess.SY↗

arXiv2026

Aerial World Model for Long-horizon Visual Generation and Navigation in 3D Space

Unmanned aerial vehicles (UAVs) have emerged as powerful embodied agents. One of the core abilities is autonomous navigation in large-scale three-dimensional environments. Existing navigation policies, however, are typically optimized for low-level objectives such as obstacle avoidance and trajectory smoothness, lacking the ability to incorporate high-level semantics into planning. To bridge this gap, we propose ANWM, an aerial navigation world model that predicts future visual observations conditioned on past frames and actions, thereby enabling agents to rank candidate trajectories by their semantic plausibility and navigational utility. ANWM is trained on 4-DoF UAV trajectories and introduces a physics-inspired module: Future Frame Projection (FFP), which projects past frames into future viewpoints to provide coarse geometric priors. This module mitigates representational uncertainty in long-distance visual generation and captures the mapping between 3D trajectories and egocentric observations. Empirical results demonstrate that ANWM significantly outperforms existing world models in long-distance visual forecasting and improves UAV navigation success rates in large-scale environments.

cs.RO↗

arXiv2026

From "The Cliff" to "Virgil": Mapping the Spectral Diversity of Little Red Dots with JWST/NIRSpec

One of JWST's most unexpected discoveries is the emergence of "Little Red Dots'' (LRDs): compact sources at $z \gtrsim 3$ with blue rest-frame UV continua, red optical slopes, and broad Balmer emission lines that challenge standard models and suggest a population of early, unusual active galactic nuclei (AGNs). Using a comprehensive photometric selection and public NIRSpec/PRISM spectroscopy across six JWST deep fields, we identify a large sample of 118 LRDs with high-S/N spectra, enabling a population-wide analysis of their UV-optical continuum and emission lines. We find clear correlations between rest-frame color ([0.3-0.9\,$μ$m]) and slopes: bluer LRDs have blue UV slopes ($β_{ν,\mathrm{UV}} \sim 0.3$) and red optical slopes, while redder LRDs exhibit redder UV slopes ($β_{ν,\mathrm{UV}} \sim 1.1$). The continuum shape shows a similar trend: redder LRDs display prominent Balmer breaks and curvature, while bluer LRDs follow power-law-like optical SEDs. From literature compilations, $\sim$60% of known broad-line AGNs satisfy our LRD criteria, and up to 90% of LRDs show broad Balmer lines. Emission-line diagnostics reveal a shift from high H$_α$/H$_β$ and low [OIII]$\lambda5007$/H$_β$ in redder LRDs to the opposite in bluer ones, along with stronger narrow-line equivalent widths, suggesting a transition from AGN- to host-dominated emission. We fit the spectra with a two-component model combining a gas-enshrouded black hole (BH) and a galaxy host. Redder LRDs require higher-luminosity, unreddened BHs and modestly reddened hosts; bluer LRDs require lower-luminosity, reddened BHs and dust-free galaxies. This framework reproduces the diversity in colors and spectral shape by varying BH luminosity, obscuration, and host-to-BH luminosity ratio.

astro-ph.GA↗