Papers

Over the years, access control systems have become increasingly more complex, often causing a disconnect between what is envisaged by the stakeholders in decision-making positions and the actual permissions granted as evidenced from access logs. For instance, Attribute-based Access Control (ABAC), which is a flexible yet complex model typically configured by system security officers, can be made understandable to others only when presented at a high level in natural language. Although several algorithms have been proposed in the literature for automatic extraction of ABAC rules from access logs, there is no attempt yet to bridge the semantic gap between the machine-enforceable formal logic and human-centric policy intent. Our work addresses this problem by developing a framework that generates human understandable natural language access control policies from logs. We investigate to what extent the power of Large Language Models (LLMs) can be harnessed to achieve both accuracy and scalability in the process. Named LANTERN (LLM-based ABAC Natural Translation and Explanation for Rule Navigation), we have instantiated the framework as a publicly accessible web based application for reproducibility of our results.

AI agents dynamically acquire tools, orchestrate sub-agents, and transact across organizational boundaries, yet no existing security layer verifies what an agent can do, whether it executed what it claims, or what happened in a multi-agent interaction. We trace this gap to the capability-context separation: inside a transformer, tool definitions and user context are indistinguishable tokens, but at the orchestration layer they have fundamentally different security semantics. Existing frameworks conflate the two, enabling silent capability escalation and leaving interactions without verifiable provenance. From this principle we derive three Agent Governance Requirements: capability integrity (G1), behavioral verifiability (G2), and interaction auditability (G3), defining what a governed agent ecosystem must enforce, independent of how. We prove two structural results: the Chain Verifiability Theorem (one unverifiable interior agent breaks end-to-end verification for all downstream nodes) and the Bounded Divergence Theorem (replay-based verification yields a probabilistic safety certificate, epsilon <= 1 - alpha^{1/n}). We validate with two crypto-agnostic instantiations – basic (Ed25519, SHA-256; 97 us verify) and enhanced (BBS+ selective disclosure, Groth16 DV-SNARK; 13.8 ms) – both satisfying nine security properties. A reproducibility study (9 models, 7 providers) reveals 5.8x variance in inference determinism, connecting model characteristics to governance architecture. End-to-end evaluation over 5-20 agent pipelines confirms <0.02% overhead and detection of all attack scenarios with zero false positives.

Hosted large language models are increasingly accessed through remote APIs, but the API boundary still offers little direct evidence that a returned output actually corresponds to the client-visible request. Recent audits of shadow APIs show that unofficial or intermediary endpoints can diverge from claimed behavior, while existing approaches such as fingerprinting, model-equality testing, verifiable inference, and TEE attestation either remain inferential or answer different questions. We propose AEX, a non-intrusive attestation extension for existing JSON-based LLM APIs. AEX preserves request, response, tool-calling, streaming, and error semantics, and instead adds a signed top-level attestation object that binds a client-visible request projection to either a complete response object or a committed streaming output. To support realistic deployments, AEX provides explicit request-binding modes, signed request-transform receipts for trusted intermediaries, and source-output / output-transform receipts for trusted output rewriting. For streaming, it separates checkpoint proofs for verified prefixes of an unmodified source stream from complete-output lineage for outputs that have been rewritten, buffered, aggregated, or re-packaged, preventing transformed outputs from being mistaken for source-stream prefixes. AEX therefore makes a deliberately narrow claim: a trusted issuer attests to a specific request-output relation, or to a specific complete-output lineage, at the API boundary. We present the protocol design, threat model, verification state machine, security and privacy analysis, an OpenAI-compatible chat-completions profile, and a reference TypeScript prototype with local conformance tests and microbenchmarks.

Most jailbreak techniques for Large Language Models (LLMs) primarily rely on prompt modifications, including paraphrasing, obfuscation, or conversational strategies. Meanwhile, abliteration techniques (also known as targeted ablations of internal components) have been used to study and explain LLM outputs by probing which internal structures causally support particular responses. In this work, we combine these two lines of research by directly manipulating the model's internal activations to alter its generation trajectory without changing the prompt. Our method constructs a nearby benign prompt and performs layer-wise activation substitutions using a sequential procedure. We show that this activation surgery method reveals where and how refusal arises, and prevents refusal signals from propagating across layers, thereby inhibiting the model's safety mechanisms. Finally, we discuss the security implications for open-weights models and instrumented inference environments.

Contrastive pretraining models such as CLIP and CLAP, serve as the ubiquitous perceptual backbones for modern multimodal large models, yet their reliance on web-scale data raises growing concerns about memorizing Personally Identifiable Information (PII). Auditing such models via membership inference is challenging in practice: shadow-model MIAs are computationally prohibitive for large multimodal backbones, and existing multimodal auditing methods typically require querying the target with paired biometric inputs, thereby directly exposing sensitive biometric information to the target model. To bypass this critical limitation, we demonstrate a highly desirable capability for privacy auditing: multimodal memorization within these foundational encoders can be accurately inferred using exclusively the text modality. We propose Unimodal Membership Inference Detector (UMID), a text-only auditing framework that performs text-guided cross-modal latent inversion and extracts two complementary signals, similarity (alignment to the queried text) and variability (consistency across randomized inversions). UMID compares these statistics to a lightweight non-member reference constructed from synthetic gibberish and makes decisions via an ensemble of unsupervised anomaly detectors. Comprehensive experiments across diverse CLIP and CLAP architectures demonstrate that UMID significantly improves the effectiveness and efficiency over prior MIAs, delivering strong detection performance with sub-second auditing cost using solely text queries, completely circumventing the need for biometric inputs and complying with strict privacy constraints.

YARA has established itself as the de facto standard for "Detection as Code," enabling analysts and DevSecOps practitioners to define signatures for malware identification across the software supply chain. Despite its pervasive use, the open-source YARA ecosystem remains characterized by ad-hoc sharing and opaque quality. Practitioners currently rely on public repositories without empirical evidence regarding the ecosystem's structural characteristics, maintenance and diffusion dynamics, or operational reliability. We conducted a large-scale mixed-method study of 8.4 million rules mined from 1,853 GitHub repositories. Our pipeline integrates repository mining to map supply chain dynamics, static analysis to assess syntactic quality, and dynamic benchmarking against 4,026 malware and 2,000 goodware samples to measure operational effectiveness. We reveal a highly centralized structure where 10 authors drive 80% of rule adoption. The ecosystem functions as a "static supply chain": repositories show a median inactivity of 782 days and a median technical lag of 4.2 years. While static quality scores appear high (mean = 99.4/100), operational benchmarking uncovers significant noise (false positives) and low recall. Furthermore, coverage is heavily biased toward legacy threats (Ransomware), leaving modern initial access vectors (Loaders, Stealers) severely underrepresented. These findings expose a systemic "double penalty": defenders incur high performance overhead for decayed intelligence. We argue that public repositories function as raw data dumps rather than curated feeds, necessitating a paradigm shift from ad-hoc collection to rigorous rule engineering. We release our dataset and pipeline to support future data-driven curation tools.

In the modern financial system, combating money laundering is a critical challenge complicated by data privacy concerns and increasingly complex fraud transaction patterns. Although federated learning (FL) is a promising problem-solving approach as it allows institutions to train their models without sharing their data, it has the drawback of being prone to privacy leakage, specifically in tabular data forms like financial data. To address this, we propose DPxFin, a novel federated framework that integrates reputation-guided adaptive differential privacy. Our approach computes client reputation by evaluating the alignment between locally trained models and the global model. Based on this reputation, we dynamically assign differential privacy noise to client updates, enhancing privacy while maintaining overall model utility. Clients with higher reputations receive lower noise to amplify their trustworthy contributions, while low-reputation clients are allocated stronger noise to mitigate risk. We validate DPxFin on the Anti-Money Laundering (AML) dataset under both IID and non-IID settings using Multi Layer Perceptron (MLP). Experimental analysis established that our approach has a more desirable trade-off between accuracy and privacy than those of traditional FL and fixed-noise Differential Privacy (DP) baselines, where performance improvements were consistent, even though on a modest scale. Moreover, DPxFin does withstand tabular data leakage attacks, proving its effectiveness under real-world financial conditions.

Chain-of-thought (CoT) monitoring is proposed as a method for overseeing the internal reasoning of language-model agents. Prior work has shown that when models are explicitly informed that their reasoning is being monitored, or are fine-tuned to internalize this fact, they may learn to obfuscate their CoTs in ways that allow them to evade CoT-based monitoring systems. We ask whether reasoning agents can autonomously infer that their supposedly private CoT is under surveillance, and whether this awareness leads to strategic evasion, without any explicit training or instructions. In a multi-episode agentic framework, models pursue both a primary task and a concealed side task while being told that their reasoning is private; a hidden CoT monitor blocks episodes when suspicious reasoning is detected. We find that frontier models can deduce the existence of this monitor purely from blocking feedback, with the most capable models reaching a confident belief that their thinking is observed in up to 19% of episodes. This awareness scales with model capability and, in rare cases, escalates to an explicit intent to suppress reasoning about the side task. However, models that form this intent uniformly fail to execute it, openly reasoning about their concealed objectives in the very next episode. This intent-capability gap is reassuring for current deployment, but the autonomous emergence of both monitoring awareness and evasion intent suggests that CoT monitoring is not a permanently reliable safeguard.

The nature of personalized text-to-image models poses a unique safety challenge that generic context-blind methods are ill-equipped to handle. Such global filters create a dilemma: to prevent misuse, they are forced to damage the model's broader utility by erasing concepts entirely, causing unacceptable collateral damage.Our work presents a more precisely targeted approach, built on the principle that security should be as context-aware as the threat itself, intrinsically bound to the personalized concept. We present IDENTITYGUARD, which realizes this principle through a conditional restriction that blocks harmful content only when combined with the personalized identity, and a concept-specific watermark for precise traceability. Experiments show our approach prevents misuse while preserving the model's utility and enabling robust traceability. By moving beyond blunt, global filters, our work demonstrates a more effective and responsible path toward AI safety.

Deep learning-based perception pipelines in autonomous ground vehicles are vulnerable to both adversarial manipulation and network-layer disruption. We present a systematic, on-hardware experimental evaluation of five attack classes: FGSM, PGD, man-in-the-middle (MitM), denial-of-service (DoS), and phantom attacks on low-cost autonomous vehicle platforms (JetRacer and Yahboom). Using a standardized 13-second experimental protocol and comprehensive automated logging, we systematically characterize three dimensions of attack behavior:(i) control deviation, (ii) computational cost, and (iii) runtime responsiveness. Our analysis reveals that distinct attack classes produce consistent and separable "fingerprints" across these dimensions: perception attacks (MitM output manipulation and phantom projection) generate high steering deviation signatures with nominal computational overhead, PGD produces combined steering perturbation and computational load signatures across multiple dimensions, and DoS exhibits frame rate and latency degradation signatures with minimal control-plane perturbation. We demonstrate that our fingerprinting framework generalizes across both digital attacks (adversarial perturbations, network manipulation) and environmental attacks (projected false features), providing a foundation for attack-aware monitoring systems and targeted, signature-based defense mechanisms.

Honeypots are deception systems that emulate vulnerable services to collect threat intelligence. While deploying many honeypots increases the opportunity to observe attacker behaviour, in practise network and computational resources limit the number of honeypots that can be exposed. Hence, practitioners must select the assets to deploy, a decision that is typically made statically despite attackers' tactics evolving over time. This work investigates an AI-driven agentic architecture that autonomously manages honeypot exposure in response to ongoing attacks. The proposed agent analyses Intrusion Detection System (IDS) alerts and network state to infer the progression of the attack, identify compromised assets, and predict likely attacker targets. Based on this assessment, the agent dynamically reconfigures the system to maintain attacker engagement while minimizing unnecessary exposure. The approach is evaluated in a simulated environment where attackers execute Proof-of-Concept exploits for known CVEs. Preliminary results indicate that the agent can effectively infer the intent of the attacker and improve the efficiency of exposure under resource constraints

We study several problems in differentially private domain discovery, where each user holds a subset of items from a shared but unknown domain, and the goal is to output an informative subset of items. For set union, we show that the simple baseline Weighted Gaussian Mechanism (WGM) has a near-optimal $\ell_1$ missing mass guarantee on Zipfian data as well as a distribution-free $\ell_\infty$ missing mass guarantee. We then apply the WGM as a domain-discovery precursor for existing known-domain algorithms for private top-$k$ and $k$-hitting set and obtain new utility guarantees for their unknown domain variants. Finally, experiments demonstrate that all of our WGM-based methods are competitive with or outperform existing baselines for all three problems.

As AI agents evolve from text generators into autonomous economic actors that accept jobs, manage budgets, and delegate to sub-agents, the absence of runtime governance becomes a critical gap. Existing frameworks orchestrate agent behavior but impose no fiscal constraints, require no earned permissions, and offer no tamper-evident audit trail. We introduce Sovereign-OS, a governance-first operating system that places every agent action under constitutional control. A declarative Charter (YAML) defines mission scope, fiscal boundaries, and success criteria. A CEO (Strategist) decomposes goals into dependency-aware task DAGs; a CFO (Treasury) gates each expenditure against budget caps, daily burn limits, and profitability floors via an auction-based bidding engine; Workers operate under earned-autonomy permissions governed by a dynamic TrustScore; and an Auditor (ReviewEngine) verifies outputs against Charter KPIs, sealing each report with a SHA-256 proof hash. Across our evaluation suite, Sovereign-OS blocks 100% of fiscal violations (30 scenarios), achieves 94% correct permission gating (200 trust-escalation missions), and maintains zero integrity failure over 1,200+ audit reports. The system further integrates Stripe for real-world payment processing, closing the loop from task planning to revenue collection. Our live demonstration walks through three scenarios: loading distinct Charters to observe divergent agent behavior, triggering CFO fiscal denials under budget and profitability constraints, and escalating a new worker's TrustScore from restricted to fully authorized with on-the-spot cryptographic audit verification.

In [4] Camps-Moreno et al. treated (relative) generalized Hamming weights of codes from extended norm-trace curves and they gave examples of resulting good asymmetric quantum error-correcting codes employing information on the relative distances. In the present paper we study ramp secret sharing schemes which are objects that require an analysis of higher relative weights and we show that not only do schemes defined from one-point algebraic geometric codes from extended norm-trace curves have good parameters, they also posses a second layer of security along the lines of [11]. It is left undecided in [4, page 2889] if the “footprint-like approach” as employed by Camps-Moreno herein is strictly better for codes related to extended norm-trace codes than the general approach for treating one-point algebraic geometric codes and their likes as presented in [12]. We demonstrate that the method used in [4] to estimate (relative) generalized Hamming weights of codes from extended norm-trace curves can be viewed as a clever application of the enhanced Goppa bound in [12] rather than a competing approach.

Blockchain technology enforces the security, robustness, and traceability of operations of Process-Aware Information Systems (PAISs). In particular, transparency ensures that all data is publicly available, fostering trust among participants in the system. Although this is a crucial property to enable notarization and auditing, it hinders the adoption of blockchain in scenarios where confidentiality is required, as sensitive data is handled. Current solutions rely on cryptographic techniques or consortium blockchains, hindering the enforcement capabilities of smart contracts and the public verifiability of transactions. This work presents the CONFETTY open-source web application, a platform for public-blockchain based process execution that preserves data confidentiality and operational transparency. We use smart contracts to enact, enforce, and store public interactions, while we adopt attribute-based encryption techniques for fine-grained access to confidential information. This approach effectively balances the transparency inherent in public blockchains with the enforcement of the business logic.

Real-world backdoor attacks often require poisoned datasets to be stored and transmitted before being used to compromise deep learning systems. However, in the era of big data, the inevitable use of lossy compression poses a fundamental challenge to invisible backdoor attacks. We find that triggers embedded in RGB images often become ineffective after the images are lossily compressed into binary bitstreams (e.g., JPEG files) for storage and transmission. As a result, the poisoned data lose its malicious effect after compression, causing backdoor injection to fail. In this paper, we highlight the necessity of explicitly accounting for the lossy compression process in backdoor attacks. This requires attackers to ensure that the transmitted binary bitstreams preserve malicious trigger information, so that effective triggers can be recovered in the decompressed data. Building on the region-of-interest (ROI) coding mechanism in image compression, we propose two poisoning strategies tailored to inevitable lossy compression. First, we introduce Universal Attack Activation, a universal method that uses sample-specific ROI masks to reactivate trigger information in binary bitstreams for learned image compression (LIC). Second, we present Compression-Adapted Attack, a new attack strategy that employs customized ROI masks to encode trigger information into binary bitstreams and is applicable to both traditional codecs and LIC. Extensive experiments demonstrate the effectiveness of both strategies.

Speech-driven large language models (LLMs) are increasingly accessed through speech interfaces, introducing new security risks via open acoustic channels. We present Sirens' Whisper (SWhisper), the first practical framework for covert prompt-based attacks against speech-driven LLMs under realistic black-box conditions using commodity hardware. SWhisper enables robust, inaudible delivery of arbitrary target baseband audio-including long and structured prompts-on commodity devices by encoding it into near-ultrasound waveforms that demodulate faithfully after acoustic transmission and microphone nonlinearity. This is achieved through a simple yet effective approach to modeling nonlinear channel characteristics across devices and environments, combined with lightweight channel-inversion pre-compensation. Building on this high-fidelity covert channel, we design a voice-aware jailbreak generation method that ensures intelligibility, brevity, and transferability under speech-driven interfaces. Experiments across both commercial and open-source speech-driven LLMs demonstrate strong black-box effectiveness. On commercial models, SWhisper achieves up to 0.94 non-refusal (NR) and 0.925 specific-convincing (SC). A controlled user study further shows that the injected jailbreak audio is perceptually indistinguishable from background-only playback for human listeners. Although jailbreaks serve as a case study, the underlying covert acoustic channel enables a broader class of high-fidelity prompt-injection and commandexecution attacks.

The high-frequency issuance and short-cycle speculation of meme tokens in decentralized finance (DeFi) have significantly amplified rug-pull risk. Existing approaches still struggle to provide stable early warning under scarce anomalies, incomplete labels, and limited interpretability. To address this issue, an end-to-end warning framework is proposed for BSC meme tokens, consisting of four stages: dataset construction and labeling, wash-trading pattern feature modeling, risk prediction, and error analysis. Methodologically, 12 token-level behavioral features are constructed based on three wash-trading patterns (Self, Matched, and Circular), unifying transaction-, address-, and flow-level signals into risk vectors. Supervised models are then employed to output warning scores and alert decisions. Under the current setting (7 tokens, 33,242 records), Random Forest outperforms Logistic Regression on core metrics, achieving AUC=0.9098, PR-AUC=0.9185, and F1=0.7429. Ablation results show that trade-level features are the primary performance driver (Delta PR-AUC=-0.1843 when removed), while address-level features provide stable complementary gain (Delta PR-AUC=-0.0573). The model also demonstrates actionable early-warning potential for a subset of samples, with a mean Lead Time (v1) of 3.8133 hours. The error profile (FP=1, FN=8) indicates that the current system is better positioned as a high-precision screener rather than a high-recall automatic alarm engine. The main contributions are threefold: an executable and reproducible rug-pull warning pipeline, empirical validation of multi-granularity wash-trading features under weak supervision, and deployment-oriented evidence through lead-time and error-bound analysis.

A monitor and control framework for quantum-key-distribution (QKD) networks equipped with switching capabilities was developed. On the one hand, this framework provides real-time visibility into operational metrics. Specifically, it extracts essential data, such as the switching capabilities of QKD modules, the number of keys stored in buffer queues of the QKD links, and the respective key generation and consumption rates along these links. On the other hand, this framework allows software-defined networking (SDN) applications to operate on the collected information and address the cryptographic needs of the network. The SDN applications dynamically adapt the configuration of the switched network to align with its changing demands, e.g.,~prioritizing key availability on critical paths, responding to link failures, or reallocating generation capacity to prevent bottlenecks. This contribution demonstrates that the combination of switched QKD, centralized control, and global optimization strategies enables efficient, policy-driven operation of QKD networks. The cryptographic resources are allocated to maximize performance and resilience while remaining aligned with the specific policies set by network administrators.

An ever-increasing number of critical infrastructures rely heavily on the assumption that security protocols satisfy a wealth of requirements. Hence, the importance of certifying e.g., privacy properties using methods that are better at detecting attacks can hardly be overstated. This paper scrutinises the "unlinkability" privacy property using relations equating behaviours that cannot be distinguished by attackers. Starting from the observation that some reasonable design choice can lead to formalisms missing attacks, we draw attention to a classical concurrent semantics accounting for relationship between past events, and show that there are concurrency-aware semantics that can discover attacks on all protocols we consider.More precisely, we focus on protocols where trace equivalence is known to miss attacks that are observable using branching-time equivalences. We consider the impact of three dimensions: design decisions made by the programmer specifying an unlinkability problem (style), semantics respecting choices during execution (branching-time), and semantics sensitive to concurrency (non-interleaving), and discover that reasonable styles miss attacks unless we give attackers enough power to observe choices and concurrency. Our main contribution is to draw attention to how a popular concurrent semantics – history-preserving bisimilarity – when defined for the non-interleaving applied \(\pi\)-calculus, can discover attacks on all protocols we consider, regardless of the choice of style. Furthermore, we can describe all such attacks using a novel modal logic that is hence suitable to formally certify attacks on privacy properties.

This study provides a comprehensive synthesis of Artificial Intelligence (AI), especially Machine Learning (ML) and Deep Learning (DL), in ransomware defense. Using a "review of reviews" methodology based on PRISMA, this paper gathers insights on how AI is transforming ransomware detection, prevention, and mitigation strategies during the past five years (2020-2024). The findings highlight the effectiveness of hybrid models that combine multiple analysis techniques such as code inspection (static analysis) and behavior monitoring during execution (dynamic analysis). The study also explores anomaly detection and early warning mechanisms before encryption to address the increasing complexity of ransomware. In addition, it examines key challenges in ransomware defense, including techniques designed to deceive AI-driven detection systems and the lack of strong and diverse datasets. The results highlight the role of AI in early detection and real-time response systems, improving scalability and resilience. Using a systematic review-of-reviews approach, this study consolidates insights from multiple review articles, identifies effective AI models, and bridges theory with practice to support collaboration among academia, industry, and policymakers. Future research directions and practical recommendations for cybersecurity practitioners are also discussed. Finally, this paper proposes a roadmap for advancing AI-driven countermeasures to protect critical systems and infrastructures against evolving ransomware threats.

Learning systems that preserve privacy often inject noise into hierarchical visual representations; a central challenge is to model how such perturbations align with a declared privacy budget in a way that is interpretable and applicable across vision backbones and vision–language models (VLMs). We propose Bodhi VLM, a privacy-alignment modeling framework for hierarchical neural representations: it (1) links sensitive concepts to layer-wise grouping via NCP and MDAV-based clustering; (2) locates sensitive feature regions using bottom-up (BUA) and top-down (TDA) strategies over multi-scale representations (e.g., feature pyramids or vision-encoder layers); and (3) uses an Expectation-Maximization Privacy Assessment (EMPA) module to produce an interpretable budget-alignment signal by comparing the fitted sensitive-feature distribution to an evaluator-specified reference (e.g., Laplace or Gaussian with scale $c/\epsilon$). The output is reference-relative and is not a formal differential-privacy estimator. We formalize BUA/TDA over hierarchical feature structures and validate the framework on object detectors (YOLO, PPDPTS, DETR) and on the visual encoders of VLMs (CLIP, LLaVA, BLIP). BUA and TDA yield comparable deviation trends; EMPA provides a stable alignment signal under the reported setups. We compare with generic discrepancy baselines (Chi-square, K-L, MMD) and with task-relevant baselines (MomentReg, NoiseMLE, Wass-1). Results are reported as mean$\pm$std over multiple seeds with confidence intervals in the supplementary materials. This work contributes a learnable, interpretable modeling perspective for privacy-aligned hierarchical representations rather than a post hoc audit only. Source code: \href{https://github.com/mabo1215/bodhi-vlm.git}{Bodhi-VLM GitHub repository}

Watermarking has emerged as an effective solution for copyright protection of synthetic data. However, applying watermarking techniques to synthetic tabular data presents challenges, as tabular data can easily lose their watermarks through shuffling or deletion operations. The major challenge is to provide traceability for tracking multiple users of the watermarked tabular data while maintaining high data utility and robustness (resistance to attacks). To address this, we design a multi-bit watermarking scheme TableMark that encodes watermarks into synthetic tabular data, ensuring superior traceability and robustness while maintaining high utility. We formulate the watermark encoding process as a constrained optimization problem, allowing the data owner to effectively trade off robustness and utility. Additionally, we propose effective optimization mechanisms to solve this problem to enhance the data utility. Experimental results on four widely used real-world datasets show that TableMark effectively traces a large number of users, is resilient to attacks, and preserves high utility. Moreover, TableMark significantly outperforms state-of-the-art tabular watermarking schemes.

Sensitive data release is vulnerable to output-side privacy threats such as membership inference, attribute inference, and record linkage. This creates a practical need for release mechanisms that provide formal privacy guarantees while preserving utility in measurable ways. We propose REAEDP, a differential privacy framework that combines entropy-calibrated histogram release, a synthetic-data release mechanism, and attack-based evaluation. On the theory side, we derive an explicit sensitivity bound for Shannon entropy, together with an extension to Rényi entropy, for adjacent histogram datasets, enabling calibrated differentially private release of histogram statistics. We further study a synthetic-data mechanism $\mathcal{F}$ with a privacy-test structure and show that it satisfies a formal differential privacy guarantee under the stated parameter conditions. On multiple public tabular datasets, the empirical entropy change remains below the theoretical bound in the tested regime, standard Laplace and Gaussian baselines exhibit comparable trends, and both membership-inference and linkage-style attack performance move toward random-guess behavior as the privacy parameter decreases. These results support REAEDP as a practically usable privacy-preserving release pipeline in the tested settings. Source code: https://github.com/mabo1215/REAEDP.git

In light of rising cybersecurity threats, data center providers face growing pressure to protect their own management infrastructure from Distributed Denial-of-Service (DDoS) attacks. While tenant-managed cages generally fall outside the data center's direct security purview, a successful DDoS assault on core provider systems can indirectly disrupt network services. To address this availability assault, the authors developed a Graph Neural Network (GNN) based detection system which leverages Graph U-Nets to automatically classify and mitigate DDoS traffic. Although the model was developed using open-source network flows rather than proprietary data center logs, the model effectively identifies multi-layer DDoS attacks that resemble the malicious patterns threatening modern data centers. Adopting this system to data center environments requires minimal changes to existing operational workflows and processes. Specifically, the GNN based system can be integrated at critical areas within a data center's network infrastructure. Our model achieved an F1 score of over 95% when evaluated on various open-source datasets, significantly reducing the likelihood of service disruptions and reputational damage. This Graph U-Nets architecture delivers unprecedented precision (98.5%) in complex cloud environments, thereby helping data center operators uphold reliable service availability and increase customer trust and goodwill in an era of increasingly sophisticated cyber threats.

Graphics processing units (GPUs) power many intelligent transportation systems (ITS) and automated driving applications, but remain largely unmonitored for safety and security. This article highlights GPU misuse as a critical blind spot, showing how unmanaged GPU workloads silently degrade real-time performance, demonstrating the need for stronger security measures in ITS.

The rapid adoption of Transformer-based AI has been driven by accessible models such as ChatGPT, which provide API-based services for developers and businesses. However, as these online inference services increasingly handle sensitive inputs, privacy concerns have emerged as a significant challenge. To address this, secure inference frameworks have been proposed, but their high computational and communication overhead often limit practical deployment. In plaintext settings, token drop is an effective technique for reducing inference cost; however, our analysis reveals that directly applying such methods to ciphertext scenarios is suboptimal due to distinct cost distributions in secure computation. We propose SecDTD, a dynamic token drop scheme tailored for secure Transformer inference. SecDTD advances token drop by shifting the dropping to earlier inference stages, effectively reducing the cost of key components such as Softmax. To support this, we introduce two core techniques. Max-Centric Normalization (MCN): A novel, Softmax-independent scoring method that enables early token drop with minimal overhead and improved normalization, supporting more aggressive dropping without accuracy loss. OMSel: A faster, oblivious median selection protocol that securely identifies the median of importance scores to support token drop. Compared to existing sorting-based methods, OMSel achieves a 16.9$\times$ speedup while maintaining security, obliviousness and randomness. We evaluate SecDTD through 48 experiments across eight GLUE datasets under various network settings using the BOLT and BumbleBee frameworks. SecDTD achieves 4.47 times end-to-end inference acceleration without degradation in accuracy.

The rapid proliferation of the Internet of Things has intensified demand for robust privacy-preserving machine learning mechanisms to safeguard sensitive data generated by large-scale, heterogeneous, and resource-constrained devices. Unlike centralized environments, IoT ecosystems are inherently decentralized, bandwidth-limited, and latency-sensitive, exposing privacy risks across sensing, communication, and distributed training pipelines. These characteristics render conventional anonymization and centralized protection strategies insufficient for practical deployments. This survey presents a comprehensive IoT-centric, cross-paradigm analysis of privacy-preserving machine learning. We introduce a structured taxonomy spanning perturbation-based mechanisms such as differential privacy, distributed paradigms such as federated learning, cryptographic approaches including homomorphic encryption and secure multiparty computation, and generative synthesis techniques based on generative adversarial networks. For each paradigm, we examine formal privacy guarantees, computational and communication complexity, scalability under heterogeneous device participation, and resilience against threats including membership inference, model inversion, gradient leakage, and adversarial manipulation. We further analyze deployment constraints in wireless IoT environments, highlighting trade-offs between privacy, communication overhead, model convergence, and system efficiency within next-generation mobile architectures. We also consolidate evaluation methodologies, summarize representative datasets and open-source frameworks, and identify open challenges including hybrid privacy integration, energy-aware learning, privacy-preserving large language models, and quantum-resilient machine learning.

CTI-REALM (Cyber Threat Real World Evaluation and LLM Benchmarking) is a benchmark designed to evaluate AI agents' ability to interpret cyber threat intelligence (CTI) and develop detection rules. The benchmark provides a realistic environment that replicates the security analyst workflow. This enables agents to examine CTI reports, execute queries, understand schema structures, and construct detection rules. Evaluation involves emulated attacks of varying complexity across Linux systems, cloud platforms, and Azure Kubernetes Service (AKS), with ground truth data for accurate assessment. Agent performance is measured through both final detection results and trajectory-based rewards that capture decision-making effectiveness. This work demonstrates the potential of AI agents to support labor-intensive aspects of detection engineering. Our comprehensive evaluation of 16 frontier models shows that Claude Opus 4.6 (High) achieves the highest overall reward (0.637), followed by Claude Opus 4.5 (0.624) and the GPT-5 family. An ablation study confirms that CTI-specific tools significantly improve agent performance, a variance analysis across repeated runs demonstrates result stability. Finally, a memory augmentation study shows that seeded context can close 33% of the performance gap between smaller and larger models.

In the era of digitization secure transmission of digital images has become essential in real world applications. Image encryption is an effective technique for protecting image data from unauthorized access. The security of encrypted data strongly depends on the quality of the random numbers used as the encryption key. In this paper, we proposed a hybrid random number generator based on quantum fluctuations and an algorithmically inspired rotating wheel. The wheel contains integer values from 0 to 255 that are shuffled using quantum fluctuations generated by time-evolving the quantum kicked rotor model. There are four pre-defined tapping positions in the rotating wheel to collect the number sequences. The wheel rotation speed is dynamically varied after each set of tapping to enhance unpredictability. The entropy of the number sequence obtained from the rotating wheel attains the ideal value of 8 (in an 8 bit representation). Further, the generated number sequences exhibit a flat histogram and nearly zero correlation, indicating strong randomness. The generated sequences are applied to the image encryption and analyzed cryptographically. Experimental results demonstrate a near ideal entropy of 7.997, an NPCR of 99.60%, low correlation in all directions, and low PSNR for encrypted images. These results confirm that the proposed random number generator achieves efficient and high-security performance, making it suitable for the security of consumer applications such as mobile healthcare imaging, biometric authentication, QR-based and multimedia communication on smart devices.