Papers

The integration of Generative Artificial Intelligence (GenAI) in healthcare is impeded by significant security challenges unaddressed by traditional frameworks, precisely the data-in-use gap where sensitive patient data and proprietary AI models are exposed during active processing. To address this, the paper proposes the Confidential Zero-Trust Framework (CZF), a novel security paradigm that synergistically combines Zero-Trust Architecture for granular access control with the hardware-enforced data isolation of Confidential Computing. We detailed a multi-tiered architectural blueprint for implementing the CZF on Google Cloud and analyzed its efficacy against real-world threats. The CZF provides a defense-in-depth architecture where data remains encrypted while in-use within a hardware-based Trusted Execution Environment (TEE). The framework's use of remote attestation offers cryptographic proof of workload integrity, transforming compliance from a procedural exercise into a verifiable technical fact and enabling secure, multi-party collaborations previously blocked by security and intellectual property concerns. By closing the data-in-use gap and enforcing Zero-Trust principles, the CZF provides a robust and verifiable framework that establishes the necessary foundation of trust to enable the responsible adoption of transformative AI technologies in healthcare.

Jailbreak attacks designed to bypass safety mechanisms pose a serious threat by prompting LLMs to generate harmful or inappropriate content, despite alignment with ethical guidelines. Crafting universal filtering rules remains difficult due to their inherent dependence on specific contexts. To address these challenges without relying on threshold calibration or model fine-tuning, this work introduces a semantic consistency analysis between successful and unsuccessful responses, demonstrating that a negation-aware scoring approach captures meaningful patterns. Building on this insight, a novel detection framework called NegBLEURT Forest is proposed to evaluate the degree of alignment between outputs elicited by adversarial prompts and expected safe behaviors. It identifies anomalous responses using the Isolation Forest algorithm, enabling reliable jailbreak detection. Experimental results show that the proposed method consistently achieves top-tier performance, ranking first or second in accuracy across diverse models using the crafted dataset, while competing approaches exhibit notable sensitivity to model and data variations.

We present ModularSubsetSelection (MSS), a new algorithm for locally differentially private (LDP) frequency estimation. Given a universe of size $k$ and $n$ users, our $\varepsilon$-LDP mechanism encodes each input via a Residue Number System (RNS) over $\ell$ pairwise-coprime moduli $m_0, \ldots, m_{\ell-1}$, and reports a randomly chosen index $j \in [\ell]$ along with the perturbed residue using the statistically optimal SubsetSelection (SS) (Wang et al. 2016). This design reduces the user communication cost from $\Theta\bigl(\omega \log_2(k/\omega)\bigr)$ bits required by standard SS (with $\omega \approx k/(e^\varepsilon+1)$) down to $\lceil \log_2 \ell \rceil + \lceil \log_2 m_j \rceil$ bits, where $m_j < k$. Server-side decoding runs in $\Theta(n + r k \ell)$ time, where $r$ is the number of LSMR (Fong and Saunders 2011) iterations. In practice, with well-conditioned moduli (i.e., constant $r$ and $\ell = \Theta(\log k)$), this becomes $\Theta(n + k \log k)$. We prove that MSS achieves worst-case MSE within a constant factor of state-of-the-art protocols such as SS and ProjectiveGeometryResponse (PGR) (Feldman et al. 2022) while avoiding the algebraic prerequisites and dynamic-programming decoder required by PGR. Empirically, MSS matches the estimation accuracy of SS, PGR, and RAPPOR (Erlingsson, Pihur, and Korolova 2014) across realistic $(k, \varepsilon)$ settings, while offering faster decoding than PGR and shorter user messages than SS. Lastly, by sampling from multiple moduli and reporting only a single perturbed residue, MSS achieves the lowest reconstruction-attack success rate among all evaluated LDP protocols.

Verifying user attributes to provide fine-grained access control to databases is fundamental to attribute-based authentication. Either a single (central) authority verifies all the attributes, or multiple independent authorities verify the attributes distributedly. In the central setup, the authority verifies all user attributes, and the user downloads only the authorized record. While this is communication efficient, it reveals all user attributes to the authority. A distributed setup prevents this privacy breach by letting each authority verify and learn only one attribute. Motivated by this, Jafarpisheh~et~al. introduced an information-theoretic formulation, called distributed attribute-based private access control (DAPAC). With $N$ non-colluding authorities (servers), $N$ attributes and $K$ possible values for each attribute, the DAPAC system lets each server learn only the single attribute value that it verifies, and is oblivious to the remaining $N-1$. The user retrieves its designated record, without learning anything about the remaining database records. The goal is to maximize the rate, i.e., the ratio of desired message size to total download size. However, not all attributes are sensitive, and DAPAC's privacy constraints can be too restrictive, negatively affecting the rate. To leverage the heterogeneous privacy requirements of user attributes, we propose heterogeneous (Het)DAPAC, a framework which off-loads verification of $N-D$ of the $N$ attributes to a central server, and retains DAPAC's architecture for the $D$ sensitive attributes. We first present a HetDAPAC scheme, which improves the rate from $\frac{1}{2K}$ to $\frac{1}{K+1}$, while sacrificing the privacy of a few non-sensitive attributes. Unlike DAPAC, our scheme entails a download imbalance across servers; we propose a second scheme achieving a balanced per-server download and a rate of $\frac{D+1}{2KD}$.

Bitcoin's (BTC) Difficulty Adjustment Algorithm (DAA) has been a source of vulnerability for incentive attacks such as selfish mining, block withholding and coin hopping strategies. In this paper, first, we rigorously study the short-term revenue change per hashpower of the adversarial and honest miners for these incentive attacks. To study the long-term effects, we introduce a new efficiency metric defined as the revenue/cost per hashpower per time for the attacker and the honest miners. Our results indicate that the short-term benefits of intermittent mining strategies are negligible compared to the original selfish mining attack, and in the long-term, selfish mining provides better efficiency. We further demonstrate that a coin hopping strategy between BTC and Bitcoin Cash (BCH) relying on BTC DAA benefits the loyal honest miners of BTC in the same way and to the same extent per unit of computational power as it does the hopper in the short-term. For the long-term, we establish a new boundary between the selfish mining and coin hopping attack, identifying the optimal efficient strategy for each parameter. For block withholding strategies, it turns out, the honest miners outside the pool profit from the attack, usually even more than the attacker both in the short-term and the long-term. Moreover, a Power Adjusting Withholding (PAW) attacker does not necessarily observe a profit lag in the short-term. In other words, even without a difficulty adjustment, a PAW attacker makes profits. It has been long thought that the profit lag of selfish mining is among the main reasons why such an attack has not been observed in practice. We show that such a barrier does not apply to PAW and relatively small pools are at an immediate threat.

The routing protocol for low-power and lossy networks (RPL) has become the de facto routing standard for resource-constrained IoT systems, but its lightweight design exposes critical vulnerabilities to a wide range of routing-layer attacks such as hello flood, decreased rank, and version number manipulation. Traditional countermeasures, including protocol-level modifications and machine learning classifiers, can achieve high accuracy against known threats, yet they fail when confronted with novel or zero-day attacks unless fully retrained, an approach that is impractical for dynamic IoT environments. In this paper, we investigate incremental learning as a practical and adaptive strategy for intrusion detection in RPL-based networks. We systematically evaluate five model families, including ensemble models and deep learning models. Our analysis highlights that incremental learning not only restores detection performance on new attack classes but also mitigates catastrophic forgetting of previously learned threats, all while reducing training time compared to full retraining. By combining five diverse models with attack-specific analysis, forgetting behavior, and time efficiency, this study provides systematic evidence that incremental learning offers a scalable pathway to maintain resilient intrusion detection in evolving RPL-based IoT networks.

Smart Contract Reusable Components(SCRs) play a vital role in accelerating the development of business-specific contracts by promoting modularity and code reuse. However, the risks associated with SCR usage violations have become a growing concern. One particular type of SCR usage violation, known as a logic-level usage violation, is becoming especially harmful. This violation occurs when the SCR adheres to its specified usage rules but fails to align with the specific business logic of the current context, leading to significant vulnerabilities. Detecting such violations necessitates a deep semantic understanding of the contract's business logic, including the ability to extract implicit usage patterns and analyze fine-grained logical behaviors. To address these challenges, we propose SCRUTINEER, the first automated and practical system for detecting logic-level usage violations of SCRs. First, we design a composite feature extraction approach that produces three complementary feature representations, supporting subsequent analysis. We then introduce a Large Language Model-powered knowledge construction framework, which leverages comprehension-oriented prompts and domain-specific tools to extract logic-level usage and build the SCR knowledge base. Next, we develop a Retrieval-Augmented Generation-driven inspector, which combines a rapid retrieval strategy with both comprehensive and targeted analysis to identify potentially insecure logic-level usages. Finally, we implement a logic-level usage violation analysis engine that integrates a similarity-based checker and a snapshot-based inference conflict checker to enable accurate and robust detection. We evaluate SCRUTINEER from multiple perspectives on 3 ground-truth datasets. The results show that SCRUTINEER achieves a precision of 80.77%, a recall of 82.35%, and an F1-score of 81.55% in detecting logic-level usage violations of SCRs.

Formal verification of cryptographic protocols typically relies on symbolic models that abstract away compiled code and microarchitectural side channels, leaving a gap between verified specifications and deployed executables. We present a toolchain that extracts protocol-relevant models from real binaries and analyzes them under explicit leakage contracts for constant-time and Spectre-PHT-style speculative observations. Starting from a selected binary region, we lift machine code to an intermediate representation, instrument it with leakage contracts, symbolically execute it to obtain event/observation traces, and translate these traces into Sapic+ for analysis with Tamarin, ProVerif, and DeepSec. As case studies, we extract models of WhatsApp Desktop's session-management and double-ratchet components from its binary and analyze forward secrecy and post-compromise security under a state-cloning compromise. For side-channel analysis, we study the Basic Access Control (BAC) protocol used in e-passports and WhatsApp's session establishment. Under our observation models, we identify an instruction-cache side channel in WhatsApp Desktop enabling social-graph inference, and we reproduce known unlinkability issues in BAC under microarchitectural observations.

Wi-Fi Channel State Information (CSI) has been repeatedly proposed as a biometric modality, often with reports of high accuracy and operational feasibility. However, the field lacks a consolidated understanding of its security properties, adversarial resilience, and methodological consistency. This Systematization of Knowledge (SoK) examines CSI-based biometric authentication through a security lens, analyzing how existing works diverge in sensing infrastructure, signal representations, feature pipelines, learning models, and evaluation methodologies. Our synthesis reveals systemic inconsistencies: reliance on aggregate accuracy metrics, limited reporting of FAR/FRR/EER, absence of per-user risk analysis, and scarce consideration of threat models or adversarial feasibility. To this end, we construct a unified evaluation framework to expose these issues empirically and demonstrate how security-relevant metrics such as per-class EER, Frequency Count of Scores (FCS), and the Gini Coefficient uncover risk concentration that remains hidden under traditional reporting practices. The resulting analysis highlights concrete attack surfaces–including replay, geometric mimicry, and environmental perturbation–and shows how methodological choices materially influence vulnerability profiles. Based on these findings, we articulate the security boundaries of current CSI biometrics and provide guidelines for rigorous evaluation, reproducible experimentation, and future research directions. This SoK offers the security community a structured, evidence-driven reassessment of Wi-Fi CSI biometrics and their suitability as an authentication primitive.

Modern electrical power grids represent complex cyber-physical systems requiring specialized cybersecurity frameworks beyond traditional IT security models. Existing threat intelligence standards such as STIX 2.1 and MITRE ATT&CK lack coverage for grid-specific assets, operational technology relationships, and cyber-physical interdependencies essential for power system security. We present Grid-STIX, a domain-specific extension of STIX 2.1 for electrical grid cybersecurity applications. Grid-STIX employs a modular architecture encompassing physical assets, operational technology components, cyber-physical relationships, and security policies that capture modern power systems including distributed energy resources, advanced metering infrastructure, and nuclear energy facilities. The framework provides threat modeling capabilities through systematic representation of attack patterns, supply chain risks, and cross-domain impact analysis while maintaining STIX 2.1 compliance. Grid-STIX includes modules for nuclear safeguards and non-proliferation verification, enabling cybersecurity modeling across conventional and nuclear energy sectors. The ontology supports Zero Trust enforcement through policy decision points and operational context integration. Our implementation includes validation pipelines, Python code generation, and visualizations. Use cases demonstrate applications including cross-utility threat intelligence sharing, supply chain risk assessment, and nuclear facility cybersecurity. Grid-STIX is available as an open-source framework to advance collaborative cybersecurity research across the electrical power sector.

Large language models (LLMs) have achieved remarkable success across a wide range of natural language processing tasks, demonstrating human-level performance in text generation, reasoning, and question answering. However, training such models requires substantial computational resources, large curated datasets, and sophisticated alignment procedures. As a result, they constitute highly valuable intellectual property (IP) assets that warrant robust protection mechanisms. Existing IP protection approaches suffer from critical limitations. Model fingerprinting techniques can identify model architectures but fail to establish ownership of specific model instances. In contrast, traditional backdoor-based watermarking methods embed behavioral anomalies that can be easily removed through common post-processing operations such as fine-tuning or knowledge distillation. We propose SEAL, a subspace-anchored watermarking framework that embeds multi-bit signatures directly into the model's latent representational space, supporting both white-box and black-box verification scenarios. Our approach leverages model editing techniques to align the hidden representations of selected anchor samples with predefined orthogonal bit vectors. This alignment embeds the watermark while preserving the model's original factual predictions, rendering the watermark functionally harmless and stealthy. We conduct comprehensive experiments on multiple benchmark datasets and six prominent LLMs, comparing SEAL with 11 existing fingerprinting and watermarking methods to demonstrate its superior effectiveness, fidelity, efficiency, and robustness. Furthermore, we evaluate SEAL under potential knowledgeable attacks and show that it maintains strong verification performance even when adversaries possess knowledge of the watermarking mechanism and the embedded signatures.

Retrieval-augmented generation (RAG) has rapidly emerged as a transformative approach for integrating large language models into clinical and biomedical workflows. However, privacy risks, such as protected health information (PHI) exposure, remain inconsistently mitigated. This review provides a thorough analysis of the current landscape of RAG applications in healthcare, including (i) sensitive data type across clinical scenarios, (ii) the associated privacy risks, (iii) current and emerging data-privacy protection mechanisms and (iv) future direction for patient data privacy protection. We synthesize 23 articles on RAG applications in healthcare and systematically analyze privacy challenges through a pipeline-structured framework encompassing data storage, transmission, retrieval and generation stages, delineating potential failure modes, their underlying causes in threat models and system mechanisms, and their practical implications. Building on this analysis, we critically review 17 articles on privacy-preserving strategies for RAG systems. Our evaluation reveals critical gaps, including insufficient clinical validation, absence of standardized evaluation frameworks, and lack of automated assessment tools. We propose actionable directions based on these limitations and conclude with a call to action. This review provides researchers and practitioners with a structured framework for understanding privacy vulnerabilities in healthcare RAG and offers a roadmap toward developing systems that achieve both clinical effectiveness and robust privacy preservation.

Rank aggregation is a task of combining the rankings of items from multiple users into a single ranking that best represents the users' rankings. Alabi et al. (AAAI'22) presents differentially-private (DP) polynomial-time approximation schemes (PTASes) and $5$-approximation algorithms with certain additive errors for the Kemeny rank aggregation problem in both central and local models. In this paper, we present improved DP PTASes with smaller additive error in the central model. Furthermore, we are first to study the footrule rank aggregation problem under DP. We give a near-optimal algorithm for this problem; as a corollary, this leads to 2-approximation algorithms with the same additive error as the $5$-approximation algorithms of Alabi et al. for the Kemeny rank aggregation problem in both central and local models.

High-assurance cryptography provides strong guarantees that source implementations are functionally correct and provably secure. In this paper, we demonstrate that the Jasmin compiler preserves functional correctness and KEM-IND-CCA security (which were established in prior work) of a highly optimized Jasmin implementation of ML-KEM used in the popular messenger Signal. Our proof of preservation is fully mechanized in the Rocq prover and is based on three general contributions: (1) A general framework for modeling game-based security and for reasoning about preservation of game-based security under compilation. (2) A new, interaction-trees-based semantics of Jasmin and assembly programs. Our new semantics supports features required by ML-KEM, such as probabilistic computations and rejection sampling routines. (3) A new relational Hoare logic for interaction trees, which we use to prove correctness of the JASMIN compiler under our new semantics.

Software quality research increasingly relies on large-scale datasets that measure both the product and process aspects of software systems. However, existing resources often focus on limited dimensions, such as code smells, technical debt, or refactoring activity, thereby restricting comprehensive analyses across time and quality dimensions. To address this gap, we present the Software Quality Dataset (SQuaD), a multi-dimensional, time-aware collection of software quality metrics extracted from 450 mature open-source projects across diverse ecosystems, including Apache, Mozilla, FFmpeg, and the Linux kernel. By integrating nine state-of-the-art static analysis tools, i.e., SonarQube, CodeScene, PMD, Understand, CK, JaSoMe, RefactoringMiner, RefactoringMiner++, and PyRef, our dataset unifies over 700 unique metrics at method, class, file, and project levels. Covering a total of 63,586 analyzed project releases, SQuaD also provides version control and issue-tracking histories, software vulnerability data (CVE/CWE), and process metrics proven to enhance Just-In-Time (JIT) defect prediction. The SQuaD enables empirical research on maintainability, technical debt, software evolution, and quality assessment at unprecedented scale. We also outline emerging research directions, including automated dataset updates and cross-project quality modeling to support the continuous evolution of software analytics. The dataset is publicly available on ZENODO (DOI: 10.5281/zenodo.17566690).

Smart contracts have emerged as key components within decentralized environments, enabling the automation of transactions through self-executing programs. While these innovations offer significant advantages, they also present potential drawbacks if the smart contract code is not carefully designed and implemented. This paper investigates the capability of large language models (LLMs) to detect OWASP-inspired vulnerabilities in smart contracts beyond the Ethereum Virtual Machine (EVM) ecosystem, focusing specifically on Solana and Algorand. Given the lack of labeled datasets for non-EVM platforms, we design a synthetic dataset of annotated smart contract snippets in Rust (for Solana) and PyTeal (for Algorand), structured around a vulnerability taxonomy derived from OWASP. We evaluate LLMs under three configurations: prompt engineering, fine-tuning, and a hybrid of both, comparing their performance on different vulnerability categories. Experimental results show that prompt engineering achieves general robustness, while fine-tuning improves precision and recall on less semantically rich languages such as TEAL. Additionally, we analyze how the architectural differences of Solana and Algorand influence the manifestation and detectability of vulnerabilities, offering platform-specific mappings that highlight limitations in existing security tooling. Our findings suggest that LLM-based approaches are viable for static vulnerability detection in smart contracts, provided domain-specific data and categorization are integrated into training pipelines.

Data normalization is a crucial preprocessing step for enhancing model performance and training stability. In federated learning (FL), where data remains distributed across multiple parties during collaborative model training, normalization presents unique challenges due to the decentralized and often heterogeneous nature of the data. Traditional methods rely on either independent client-side processing, i.e., local normalization, or normalizing the entire dataset before distributing it to parties, i.e., pooled normalization. Local normalization can be problematic when data distributions across parties are non-IID, while the pooled normalization approach conflicts with the decentralized nature of FL. In this paper, we explore the adaptation of widely used normalization techniques to FL and define the term federated normalization. Federated normalization simulates pooled normalization by enabling the collaborative exchange of normalization parameters among parties. Thus, it achieves performance on par with pooled normalization without compromising data locality. However, sharing normalization parameters such as the mean introduces potential privacy risks, which we further mitigate through a robust privacy-preserving solution. Our contributions include: (i) We systematically evaluate the impact of various federated and local normalization techniques in heterogeneous FL scenarios, (ii) We propose a novel homomorphically encrypted $k$-th ranked element (and median) calculation tailored for the federated setting, enabling secure and efficient federated normalization, (iii) We propose privacy-preserving implementations of widely used normalization techniques for FL, leveraging multiparty fully homomorphic encryption (MHE).

This paper investigates neural network-based biometric authentication using finger-drawn digits on touchscreen devices. We evaluated CNN and autoencoder architectures for user authentication through simple digit patterns (0-9) traced with finger input. Twenty participants contributed 2,000 finger-drawn digits each on personal touchscreen devices. We compared two CNN architectures: a modified Inception-V1 network and a lightweight shallow CNN for mobile environments. Additionally, we examined Convolutional and Fully Connected autoencoders for anomaly detection. Both CNN architectures achieved ~89% authentication accuracy, with the shallow CNN requiring fewer parameters. Autoencoder approaches achieved ~75% accuracy. The results demonstrate that finger-drawn symbol authentication provides a viable, secure, and user-friendly biometric solution for touchscreen devices. This approach can be integrated with existing pattern-based authentication methods to create multi-layered security systems for mobile applications.

Cyberattacks are becoming increasingly frequent and sophisticated, often exploiting the software supply chain (SSC) as an attack vector. Attack graphs provide a detailed representation of the sequence of events and vulnerabilities that could lead to a successful security breach in a system. MulVal is a widely used open-source tool for logical attack graph generation in networked systems. However, its current lack of support for capturing and reasoning about SSC threat propagation makes it unsuitable for addressing modern SSC attacks, such as the XZ compromise or the 3CX double SSC attack. To address this limitation, we propose an extension to MulVal that integrates SSC threat propagation analysis with existing network-based threat analysis. This extension introduces a new set of predicates within the familiar MulVal syntax, enabling seamless integration. The new facts and interaction rules model SSC assets, their dependencies, interactions, compromises, additional security mechanisms, initial system states, and known threats. We explain how this integration operates in both directions and demonstrate the practical application of the extension.

Vehicle-to-Everything (V2X) communication faces a critical authentication dilemma: traditional public-key schemes like ECDSA provide strong security but impose 2 ms verification delays unsuitable for collision avoidance, while symmetric approaches like TESLA achieve microsecond-level efficiency at the cost of 20-100 ms key disclosure latency. Neither meets 5G New Radio (NR)-V2X's stringent requirements for both immediate authentication and computational efficiency. This paper presents SALT-V, a novel hybrid authentication framework that reconciles this fundamental trade-off through intelligent protocol stratification. SALT-V employs ECDSA signatures for 10% of traffic (BOOT frames) to establish sender trust, then leverages this trust anchor to authenticate 90% of messages (DATA frames) using lightweight GMAC operations. The core innovation - an Ephemeral Session Tag (EST) whitelist mechanism - enables 95% of messages to achieve immediate verification without waiting for key disclosure, while Bloom filter integration provides O(1) revocation checking in 1 us. Comprehensive evaluation demonstrates that SALT-V achieves 0.035 ms average computation time (57x faster than pure ECDSA), 1 ms end-to-end latency, 41-byte overhead, and linear scalability to 2000 vehicles, making it the first practical solution to satisfy all safety-critical requirements for real-time V2X deployment.

Healthcare AI systems face major vulnerabilities to data poisoning that current defenses and regulations cannot adequately address. We analyzed eight attack scenarios in four categories: architectural attacks on convolutional neural networks, large language models, and reinforcement learning agents; infrastructure attacks exploiting federated learning and medical documentation systems; critical resource allocation attacks affecting organ transplantation and crisis triage; and supply chain attacks targeting commercial foundation models. Our findings indicate that attackers with access to only 100-500 samples can compromise healthcare AI regardless of dataset size, often achieving over 60 percent success, with detection taking an estimated 6 to 12 months or sometimes not occurring at all. The distributed nature of healthcare infrastructure creates many entry points where insiders with routine access can launch attacks with limited technical skill. Privacy laws such as HIPAA and GDPR can unintentionally shield attackers by restricting the analyses needed for detection. Supply chain weaknesses allow a single compromised vendor to poison models across 50 to 200 institutions. The Medical Scribe Sybil scenario shows how coordinated fake patient visits can poison data through legitimate clinical workflows without requiring a system breach. Current regulations lack mandatory adversarial robustness testing, and federated learning can worsen risks by obscuring attribution. We recommend multilayer defenses including required adversarial testing, ensemble-based detection, privacy-preserving security mechanisms, and international coordination on AI security standards. We also question whether opaque black-box models are suitable for high-stakes clinical decisions, suggesting a shift toward interpretable systems with verifiable safety guarantees.

Software vulnerabilities are increasing at an alarming rate. However, manual patching is both time-consuming and resource-intensive, while existing automated vulnerability repair (AVR) techniques remain limited in effectiveness. Recent advances in large language models (LLMs) have opened a new paradigm for AVR, demonstrating remarkable progress. To examine the capability of LLMs in AVR, several vulnerability benchmarks have been proposed recently. However, they still suffer from key limitations of outdated vulnerabilities, limited language coverage, unreliable patch validation, and insufficient reproducibility. To overcome these challenges, we introduce PATCHEVAL, a multilingual benchmark for Go, JavaScript, and Python, languages for which existing benchmarks remain unexplored. PATCHEVAL curates a dataset of 1,000 vulnerabilities drawn from CVEs reported between 2015 and 2025, covering 65 distinct CWEs. A subset of 230 CVEs is further equipped with runtime sandbox environments, enabling patch verification through both security tests and functionality tests. To provide a systematic comparison of LLM-based vulnerability repair, we evaluate a series of state-of-the-art LLMs and agents, presenting an in-depth analysis that empirically yields key insights to guide future research in AVR.

Large language models (LLMs) are increasingly tasked with generating structured outputs. While structured generation methods ensure validity, they often lack output diversity, a critical limitation that we confirm in our preliminary study. We propose a novel method to enhance diversity in automaton-based structured generation. Our approach utilizes automata traversal history to steer LLMs towards novel structural patterns. Evaluations show our method significantly improves structural and content diversity while maintaining comparable generation efficiency. Furthermore, we conduct a case study showcasing the effectiveness of our method in generating diverse test cases for testing open-source libraries.

Cheating in online games poses significant threats to the gaming industry, yet most prior research has concentrated on Massively Multiplayer Online Role-Playing Games (MMORPGs). Competitive genres-such as Multiplayer Online Battle Arena (MOBA), First Person Shooter (FPS), Real Time Strategy (RTS), and Action games-remain underexplored due to the difficulty of detecting cheating users and the demand for complex data and techniques. To address this gap, many game companies rely on kernel-level anti-cheat solutions, which, while effective, raise serious concerns regarding user privacy and system security. In this paper, we propose SYNOPTICON, a novel cheating detection framework that leverages user consensus to identify abnormal behavior. SYNOPTICON integrates a lightweight client-side detection mechanism with a server-side voting system: when suspicious activity is identified, clients cast votes to the server, which aggregates them to establish consensus and distinguish cheaters from legitimate players. This architecture enables transparency, reduces reliance on intrusive monitoring, and mitigates privacy risks. We evaluate SYNOPTICON in both a controlled simulation and a real-world FPS environment. Simulation results verify its feasibility and requirements, while real-world experiments confirm its effectiveness in reliably detecting cheating users. Furthermore, we demonstrate the system's applicability and sustainability for long-term game management using public datasets. SYNOPTICON represents a user-driven, consensus-based alternative to conventional anti-cheat systems, offering a practical and privacy-preserving solution for competitive online games.

Graph unlearning (GU) has emerged as a promising solution to comply with "the right to be forgotten" regulations by enabling the removal of sensitive information upon request. However, this solution is not foolproof. The involvement of multiple parties creates new attack surfaces, and residual traces of deleted data can persist within the unlearned graph neural networks (GNNs). These vulnerabilities can be exploited by attackers to recover the supposedly erased samples, undermining the intended functionality of GU. In this work, we propose GraphToxin, the first full graph reconstruction attack against GU. We show that GraphToxin can recover not only a deleted individual's information and personal links but also sensitive content of their neighbors, thereby posing substantially more detrimental threats than prior membership inference attacks (MIAs). Specifically, we introduce a novel curvature matching module to provide fine-grained guidance for full unlearned graph recovery. We further extend GraphToxin to multiple-node removal under both white-box and black-box settings, showcasing its practical feasibility and potential to cause considerable harm. We highlight the necessity of worst-case analysis and propose a systematic evaluation framework to assess attack performance under both random and worst-case node removal scenarios. Our extensive experiments demonstrate the effectiveness and flexibility of GraphToxin. Notably, current defense mechanisms are largely ineffective against this attack. Additionally, our findings reveal that existing GU verification standards based on MIAs can be misleading: they achieve membership-level protection, while the full unlearned graph remains recoverable through GraphToxin.

Robust invisible watermarking embeds hidden information in images such that the watermark can survive various manipulations. However, the emergence of powerful diffusion-based image generation and editing techniques poses a new threat to these watermarking schemes. In this paper, we investigate the intersection of diffusion-based image editing and robust image watermarking. We analyze how diffusion-driven image edits can significantly degrade or even fully remove embedded watermarks from state-of-the-art robust watermarking systems. Both theoretical formulations and empirical experiments are provided. We prove that as a image undergoes iterative diffusion transformations, the mutual information between the watermarked image and the embedded payload approaches zero, causing watermark decoding to fail. We further propose a guided diffusion attack algorithm that explicitly targets and erases watermark signals during generation. We evaluate our approach on recent deep learning-based watermarking schemes and demonstrate near-zero watermark recovery rates after attack, while maintaining high visual fidelity of the regenerated images. Finally, we discuss ethical implications of such watermark removal capablities and provide design guidelines for future watermarking strategies to be more resilient in the era of generative AI.

Modern text-to-speech (TTS) systems, particularly those built on Large Audio-Language Models (LALMs), generate high-fidelity speech that faithfully reproduces input text and mimics specified speaker identities. While prior misuse studies have focused on speaker impersonation, this work explores a distinct content-centric threat: exploiting TTS systems to produce speech containing harmful content. Realizing such threats poses two core challenges: (1) LALM safety alignment frequently rejects harmful prompts, yet existing jailbreak attacks are ill-suited for TTS because these systems are designed to faithfully vocalize any input text, and (2) real-world deployment pipelines often employ input/output filters that block harmful text and audio. We present HARMGEN, a suite of five attacks organized into two families that address these challenges. The first family employs semantic obfuscation techniques (Concat, Shuffle) that conceal harmful content within text. The second leverages audio-modality exploits (Read, Spell, Phoneme) that inject harmful content through auxiliary audio channels while maintaining benign textual prompts. Through evaluation across five commercial LALMs-based TTS systems and three datasets spanning two languages, we demonstrate that our attacks substantially reduce refusal rates and increase the toxicity of generated speech. We further assess both reactive countermeasures deployed by audio-streaming platforms and proactive defenses implemented by TTS providers. Our analysis reveals critical vulnerabilities: deepfake detectors underperform on high-fidelity audio; reactive moderation can be circumvented by adversarial perturbations; while proactive moderation detects 57-93% of attacks. Our work highlights a previously underexplored content-centric misuse vector for TTS and underscore the need for robust cross-modal safeguards throughout training and deployment.

This paper presents a secure aggregation system Armadillo that has disruptive resistance against adversarial clients, such that any coalition of malicious clients (within the tolerated threshold) can affect the aggregation result only by misreporting their private inputs in a pre-defined legitimate range. Armadillo is designed for federated learning setting, where a single powerful server interacts with many weak clients iteratively to train models on client's private data. While a few prior works consider disruption resistance under such setting, they either incur high per-client cost (Chowdhury et al. CCS '22) or require many rounds (Bell et al. USENIX Security '23). Although disruption resistance can be achieved generically with zero-knowledge proof techniques (which we also use in this paper), we realize an efficient system with two new designs: 1) a simple two-layer secure aggregation protocol that requires only simple arithmetic computation; 2) an agreement protocol that removes the effect of malicious clients from the aggregation with low round complexity. With these techniques, Armadillo completes each secure aggregation in 3 rounds while keeping the server and clients computationally lightweight.

We present an end-to-end demonstration of how attackers can exploit AI safety failures to harm vulnerable populations: from jailbreaking LLMs to generate phishing content, to deploying those messages against real targets, to successfully compromising elderly victims. We systematically evaluated safety guardrails across six frontier LLMs spanning four attack categories, revealing critical failures where several models exhibited near-complete susceptibility to certain attack vectors. In a human validation study with 108 senior volunteers, AI-generated phishing emails successfully compromised 11% of participants. Our work uniquely demonstrates the complete attack pipeline targeting elderly populations, highlighting that current AI safety measures fail to protect those most vulnerable to fraud. Beyond generating phishing content, LLMs enable attackers to overcome language barriers and conduct multi-turn trust-building conversations at scale, fundamentally transforming fraud economics. While some providers report voluntary counter-abuse efforts, we argue these remain insufficient.

High-precision clock synchronization is essential for a wide range of network-distributed applications. In the quantum space, these applications include communication, sensing, and positioning. However, current synchronization techniques are vulnerable to attacks, such as intercept-resend attacks, spoofing, and delay attacks. Here, we propose and experimentally demonstrate a new quantum secure time transfer (QSTT) system, subsequently used for clock synchronization, that largely negates such attacks. Novel to our system is the optimal use of self-generated quantum keys within the QSTT to information-theoretically secure the maximum amount of timing data; as well as the introduction, within a hybrid quantum/post-quantum architecture, of an information-theoretic secure obfuscated encryption sequence of the remaining timing data. With these enhancements, we argue that our new system represents the most robust implementation of QSTT to date.