Papers

As the primary mechanism of digital authentication, user-created passwords exhibit common patterns and regularities that can be learned from leaked datasets. Password choices are profoundly shaped by external factors, including social contexts, cultural trends, and popular vocabulary. Prevailing password guessing models primarily emphasize patterns derived from leaked passwords, while neglecting these external influences – a limitation that hampers their adaptability to emerging password trends and erodes their effectiveness over time. To address these challenges, we propose KAPG, a knowledge-augmented password guessing framework that adaptively integrates external lexical knowledge into the guessing process. KAPG couples internal statistical knowledge learned from leaked passwords with external information that reflects real-world trends. By using password prefixes as anchors for knowledge lookup, it dynamically injects relevant external cues during generation while preserving the structural regularities of authentic passwords. Experiments on twelve leaked datasets show that KnowGuess achieves average improvements of 36.5% and 74.7% over state-of-the-art models in intra-site and cross-site scenarios, respectively. Further analyses of password overlap and model efficiency highlight its robustness and computational efficiency. To counter these attacks, we further develop KAPSM, a trend-aware and site-specific password strength meter. Experiments demonstrate that KAPSM significantly outperforms existing tools in accuracy across diverse evaluation settings.

Linguistic steganography enables covert communication through embedding secret messages into innocuous texts; however, current methods face critical limitations in payload capacity and security. Traditional modification-based methods introduce detectable anomalies, while retrieval-based strategies suffer from low embedding capacity. Modern generative steganography leverages language models to generate natural stego text but struggles with limited entropy in token predictions, further constraining capacity. To address these issues, we propose an entropy-driven framework called RTMStega that integrates rank-based adaptive coding and context-aware decompression with normalized entropy. By mapping secret messages to token probability ranks and dynamically adjusting sampling via context-aware entropy-based adjustments, RTMStega achieves a balance between payload capacity and imperceptibility. Experiments across diverse datasets and models demonstrate that RTMStega triples the payload capacity of mainstream generative steganography, reduces processing time by over 50%, and maintains high text quality, offering a trustworthy solution for secure and efficient covert communication.

Binarized Neural Networks (BNNs) are a class of deep neural networks designed to utilize minimal computational resources, which drives their popularity across various applications. Recent studies highlight the potential of mapping BNN model parameters onto emerging non-volatile memory technologies, specifically using crossbar architectures, resulting in improved inference performance compared to traditional CMOS implementations. However, the common practice of protecting model parameters from theft attacks by storing them in an encrypted format and decrypting them at runtime introduces significant computational overhead, thus undermining the core principles of in-memory computing, which aim to integrate computation and storage. This paper presents a robust strategy for protecting BNN model parameters, particularly within in-memory computing frameworks. Our method utilizes a secret key derived from a physical unclonable function to transform model parameters prior to storage in the crossbar. Subsequently, the inference operations are performed on the encrypted weights, achieving a very special case of Fully Homomorphic Encryption (FHE) with minimal runtime overhead. Our analysis reveals that inference conducted without the secret key results in drastically diminished performance, with accuracy falling below 15%. These results validate the effectiveness of our protection strategy in securing BNNs within in-memory computing architectures while preserving computational efficiency.

Virtual Reality (VR) has gained increasing traction among various domains in recent years, with major companies such as Meta, Pico, and Microsoft launching their application stores to support third-party developers in releasing their applications (or simply apps). These apps offer rich functionality but inherently collect privacy-sensitive data, such as user biometrics, behaviors, and the surrounding environment. Nevertheless, there is still a lack of domain-specific regulations to govern the data handling of VR apps, resulting in significant variations in their privacy practices among app stores. In this work, we present the first comprehensive multi-store study of privacy practices in the current VR app ecosystem, covering a large-scale dataset involving 6,565 apps collected from five major app stores. We assess both declarative and behavioral privacy practices of VR apps, using a multi-faceted approach based on natural language processing, reverse engineering, and static analysis. Our assessment reveals significant privacy compliance issues across all stores, underscoring the premature status of privacy protection in this rapidly growing ecosystem. For instance, one third of apps fail to declare their use of sensitive data, and 21.5% of apps neglect to provide valid privacy policies. Our work sheds light on the status quo of privacy protection within the VR app ecosystem for the first time. Our findings should raise an alert to VR app developers and users, and encourage store operators to implement stringent regulations on privacy compliance among VR apps.

Introduced by Juels and Rivest in 2013, Honeywords, which are decoy passwords stored alongside a real password, appear to be a proactive method to help detect password credentials misuse. However, despite over a decade of research, this technique has not been adopted by major authentication platforms. This position paper argues that the core concept of Honeywords has potential but requires more research on issues such as flatness, integration, and reliability, in order to be a practical deployable solution. This paper examines the current work on Honeyword generation, attacker modeling, and honeychecker architecture, analyzing the subproblems that have been addressed and ongoing issues that prevent this system from being more widely used. The paper then suggests a deployable framework that combines the attacker-resilient, context-aware decoy creation that Honeywords provide with easy integration into existing systems. Honeywords will only move from an academic idea to a practical security tool if technical advances are paired with secure and straightforward architectures, along with adaptive response handling and detailed configuration checks.

Prompt compression is increasingly deployed in LLM agents to reduce latency and cost, but it also determines what the backend LLM ultimately sees. We show that, when trusted and untrusted inputs are compressed under a shared budget, this lossy transformation creates a new attack surface: by perturbing only untrusted inputs before compression, an adversary can cause the compressor to discard task-critical evidence or safety guardrails before inference. Unlike prompt injection, jailbreaks, or RAG poisoning, the attack target is the compressor rather than the backend LLM; the perturbation need not encode a meaningful instruction or survive compression. We formalize this vulnerability as adversarial information loss (AIL), the excess downstream distortion caused by adversarially steering a lossy compressor beyond benign compression alone. To exploit AIL, we present COMA, a transfer-based black-box attack that optimizes pre-compression perturbations using attacker-side surrogate compressors and backend LLMs. Across three tasks and six compressors, COMA achieves 0.71 average ASR, versus 0.21 for the strongest baseline, and transfers to two real-world agent case studies.

In this paper, we propose a groundbreaking quantum-secure federated learning (QFL) framework designed to safeguard distributed learning systems against the emerging threat of quantum-enabled adversaries. As classical cryptographic methods become increasingly vulnerable to quantum attacks, our framework establishes a resilient security architecture that remains robust even in the presence of quantum-capable attackers. We integrate and rigorously evaluate advanced quantum and post-quantum protocols including Quantum Key Distribution (QKD), Quantum Teleportation, Key Encapsulation Mechanisms (KEM) and Post-Quantum Cryptography (PQC) to fortify the QFL process against both classical and quantum threats. These mechanisms are systematically analyzed and implemented to demonstrate their seamless interoperability within a secure and scalable QFL ecosystem. Through comprehensive theoretical modeling and experimental validation, this work provides a detailed security and performance assessment of the proposed framework. Our findings lay a strong foundation for next-generation federated learning systems that are inherently secure in the quantum era.

Large language models (LLMs) have become indispensable for automated code generation, yet the quality and security of their outputs remain a critical concern. Existing studies predominantly concentrate on adversarial attacks or inherent flaws within the models. However, a more prevalent yet underexplored issue concerns how the quality of a benign but poorly formulated prompt affects the security of the generated code. To investigate this, we first propose an evaluation framework for prompt quality encompassing three key dimensions: goal clarity, information completeness, and logical consistency. Based on this framework, we construct and publicly release CWE-BENCH-PYTHON, a large-scale benchmark dataset containing tasks with prompts categorized into four distinct levels of normativity (L0-L3). Extensive experiments on multiple state-of-the-art LLMs reveal a clear correlation: as prompt normativity decreases, the likelihood of generating insecure code consistently and markedly increases. Furthermore, we demonstrate that advanced prompting techniques, such as Chain-of-Thought and Self-Correction, effectively mitigate the security risks introduced by low-quality prompts, substantially improving code safety. Our findings highlight that enhancing the quality of user prompts constitutes a critical and effective strategy for strengthening the security of AI-generated code.

SpoofTrackBench is a reproducible, modular benchmark for evaluating adversarial robustness in real-time localization and tracking (RTLS) systems under radar spoofing. Leveraging the Hampton University Skyler Radar Sensor dataset, we simulate drift, ghost, and mirror-type spoofing attacks and evaluate tracker performance using both Joint Probabilistic Data Association (JPDA) and Global Nearest Neighbor (GNN) architectures. Our framework separates clean and spoofed detection streams, visualizes spoof-induced trajectory divergence, and quantifies assignment errors via direct drift-from-truth metrics. Clustering overlays, injection-aware timelines, and scenario-adaptive visualizations enable interpretability across spoof types and configurations. Evaluation figures and logs are auto-exported for reproducible comparison. SpoofTrackBench sets a new standard for open, ethical benchmarking of spoof-aware tracking pipelines, enabling rigorous cross-architecture analysis and community validation.

Post-quantum multivariate public key cryptography (MPKC) schemes resist quantum threats but require heavy operations, such as rejection sampling, which challenge resource-limited devices. Prior hardware designs have addressed various aspects of MPKC signature generation. However, rejection sampling remains largely unexplored in such contexts. This paper presents RejSCore, a lightweight hardware accelerator for rejection sampling in post-quantum cryptography. It specifically targets the QR-UOV scheme, which is a prominent candidate under the second-round of the National Institute of Standards and Technology (NIST) additional digital signature standardization process. The architecture includes an AES-CTR-128-based pseudorandom number generator. Moreover, a lightweight iterative method is employed in rejection sampling, offering reduced resource consumption and area overhead while slightly increasing latency. The performance of RejSCore is comprehensively evaluated on Artix-7 FPGAs and 65 nm CMOS technology using the Area-Delay Product (ADP) and Power-Delay Product (PDP). On Artix-7 and 65 nm CMOS, RejSCore achieves an area of 2042 slices and 464,866~$\mu m^2$, with operating frequencies of 222 MHz and 565 MHz, respectively. Using the QR-UOV parameters for security level I ($q = 127$, $v = 156$, $m = 54$, $l = 3$), the core completes its operation in 8525 clock cycles. The ADP and PDP evaluations confirm RejSCore's suitability for deployment in resource-constrained and security-critical environments.

This paper presents a real-time modular defense system named Sentra-Guard. The system detects and mitigates jailbreak and prompt injection attacks targeting large language models (LLMs). The framework uses a hybrid architecture with FAISS-indexed SBERT embedding representations that capture the semantic meaning of prompts, combined with fine-tuned transformer classifiers, which are machine learning models specialized for distinguishing between benign and adversarial language inputs. It identifies adversarial prompts in both direct and obfuscated attack vectors. A core innovation is the classifier-retriever fusion module, which dynamically computes context-aware risk scores that estimate how likely a prompt is to be adversarial based on its content and context. The framework ensures multilingual resilience with a language-agnostic preprocessing layer. This component automatically translates non-English prompts into English for semantic evaluation, enabling consistent detection across over 100 languages. The system includes a HITL feedback loop, where decisions made by the automated system are reviewed by human experts for continual learning and rapid adaptation under adversarial pressure. Sentra-Guard maintains an evolving dual-labeled knowledge base of benign and malicious prompts, enhancing detection reliability and reducing false positives. Evaluation results show a 99.96% detection rate (AUC = 1.00, F1 = 1.00) and an attack success rate (ASR) of only 0.004%. This outperforms leading baselines such as LlamaGuard-2 (1.3%) and OpenAI Moderation (3.7%). Unlike black-box approaches, Sentra-Guard is transparent, fine-tunable, and compatible with diverse LLM backends. Its modular design supports scalable deployment in both commercial and open-source environments. The system establishes a new state-of-the-art in adversarial LLM defense.

The misuse of advanced generative AI models has resulted in the widespread proliferation of falsified data, particularly forged human-centric audiovisual content, which poses substantial societal risks (e.g., financial fraud and social instability). In response to this growing threat, several works have preliminarily explored countermeasures. However, the lack of sufficient and diverse training data, along with the absence of a standardized benchmark, hinder deeper exploration. To address this challenge, we first build Mega-MMDF, a large-scale, diverse, and high-quality dataset for multimodal deepfake detection. Specifically, we employ 21 forgery pipelines through the combination of 10 audio forgery methods, 12 visual forgery methods, and 6 audio-driven face reenactment methods. Mega-MMDF currently contains 0.1 million real samples and 1.1 million forged samples, making it one of the largest and most diverse multimodal deepfake datasets, with plans for continuous expansion. Building on it, we present DeepfakeBench-MM, the first unified benchmark for multimodal deepfake detection. It establishes standardized protocols across the entire detection pipeline and serves as a versatile platform for evaluating existing methods as well as exploring novel approaches. DeepfakeBench-MM currently supports 5 datasets and 11 multimodal deepfake detectors. Furthermore, our comprehensive evaluations and in-depth analyses uncover several key findings from multiple perspectives (e.g., augmentation, stacked forgery). We believe that DeepfakeBench-MM, together with our large-scale Mega-MMDF, will serve as foundational infrastructures for advancing multimodal deepfake detection.

AI agents powered by large language models (LLMs) are being deployed at scale, yet we lack a systematic understanding of how the choice of backbone LLM affects agent security. The non-deterministic sequential nature of AI agents complicates security modeling, while the integration of traditional software with AI components entangles novel LLM vulnerabilities with conventional security risks. Existing frameworks only partially address these challenges as they either capture specific vulnerabilities only or require modeling of complete agents. To address these limitations, we introduce threat snapshots: a framework that isolates specific states in an agent's execution flow where LLM vulnerabilities manifest, enabling the systematic identification and categorization of security risks that propagate from the LLM to the agent level. We apply this framework to construct the $b^3$ benchmark, a security benchmark based on 194,331 unique crowdsourced adversarial attacks. We then evaluate 34 popular LLMs with it, revealing, among other insights, that enhanced reasoning capabilities improve security, while model size does not correlate with security. We release our benchmark, dataset, and evaluation code to facilitate widespread adoption by LLM providers and practitioners, offering guidance for agent developers and incentivizing model developers to prioritize backbone security improvements.

Recent work has revealed MOLE, the first practical attack to compromise GPU Trusted Execution Environments (TEEs), by injecting malicious firmware into the embedded Microcontroller Unit (MCU) of Arm Mali GPUs. By exploiting the absence of cryptographic verification during initialization, adversaries with kernel privileges can bypass memory protections, exfiltrate sensitive data at over 40 MB/s, and tamper with inference results, all with negligible runtime overhead. This attack surface affects commodity mobile SoCs and cloud accelerators, exposing a critical firmware-level trust gap in existing GPU TEE designs. To address this gap, this paper presents FAARM, a lightweight Firmware Attestation and Authentication framework that prevents MOLE-style firmware subversion. FAARM integrates digital signature verification at the EL3 secure monitor using vendor-signed firmware bundles and an on-device public key anchor. At boot, EL3 verifies firmware integrity and authenticity, enforces version checks, and locks the firmware region, eliminating both pre-verification and time-of-check-to-time-of-use (TOCTOU) attack vectors. We implement FAARM as a software-only prototype on a Mali GPU testbed, using a Google Colab-based emulation framework that models the firmware signing process, the EL1 to EL3 load path, and secure memory configuration. FAARM reliably detects and blocks malicious firmware injections, rejecting tampered images before use and denying overwrite attempts after attestation. Firmware verification incurs only 1.34 ms latency on average, demonstrating that strong security can be achieved with negligible overhead. FAARM thus closes a fundamental gap in shim-based GPU TEEs, providing a practical, deployable defense that raises the security baseline for both mobile and cloud GPU deployments.

Blockchain systems rely on decentralized ledgers and strong security guarantees. A key requirement is non-repudiation, which prevents denial of transaction authorship and supports integrity of recorded data. This work surveys digital signature schemes used in blockchain platforms and analyzes how they deliver non-repudiation and contribute to overall system security. We examine representative scheme families and their cryptographic foundations, security assumptions, and properties relevant to deployment, including unforgeability, resistance to malleability, support for aggregation and multisignature or threshold settings, key and signature sizes, and verification cost. Using these criteria, we compare the suitability of different designs for consensus protocols, smart contract constraints, and resource limits. We highlight practical tradeoffs that affect throughput, storage, scalability, and attack surfaces, and summarize benefits and limitations of each scheme in blockchain contexts. The study underscores that carefully chosen digital signatures are central to achieving non-repudiation and preserving information integrity, and it outlines implementation considerations and open directions such as interoperability and post-quantum readiness.

Graph Neural Networks(GNNs) are vulnerable to backdoor attacks, where adversaries implant malicious triggers to manipulate model predictions. Existing trigger generators are often simplistic in structure and overly reliant on specific features, confining them to a single graph learning paradigm, such as graph supervised learning, graph contrastive learning, or graph prompt learning. Such paradigm-specific designs lead to poor transferability across different learning frameworks, limiting attack success rates in general testing scenarios. To bridge this gap, we propose Cross-Paradigm Graph Backdoor Attacks with Promptable Subgraph Triggers(CP-GBA), which employs Graph Prompt Learning(GPL) to synthesize transferable subgraph triggers. Specifically, we first distill a compact yet expressive trigger set into a queryable repository, jointly optimizing for class-awareness, feature richness, and structural fidelity. Furthermore, we pioneer the theoretical exploration of GPL transferability under prompt-based objectives, ensuring robust generalization to diverse and unseen test-time paradigms. Extensive experiments across multiple real-world datasets and defense scenarios show that CP-GBA achieves state-of-the-art attack success rates.

We formalize a cross-domain "ZK coprocessor bridge" that lets Solana programs request private execution on Aztec L2 (via Ethereum) using Wormhole Verifiable Action Approvals (VAAs) as authenticated transport. The system comprises: (i) a Solana program that posts messages to Wormhole Core with explicit finality; (ii) an EVM Portal that verifies VAAs, enforces a replay lock, parses a bound payload secretHash||m from the attested VAA, derives a domain-separated field commitment, and enqueues an L1->L2 message into the Aztec Inbox (our reference implementation v0.1.0 currently uses consumeWithSecret(vaa, secretHash); we provide migration guidance to the payload-bound interface); (iii) a minimal Aztec contract that consumes the message privately; and (iv) an off-chain relayer that ferries VAAs and can record receipts on Solana. We present state machines, message formats, and proof sketches for replay-safety, origin authenticity, finality alignment, parameter binding (no relayer front-running of Aztec parameters), privacy, idempotence, and liveness. Finally, we include a concise Reproducibility note with pinned versions and artifacts to replicate a public testnet run.

Provenance analysis (PA) has recently emerged as an important solution for cyber attack investigation. PA leverages system monitoring to monitor system activities as a series of system audit events and organizes these events as a provenance graph to show the dependencies among system activities, which can reveal steps of cyber attacks. Despite their potential, existing PA techniques face two critical challenges: (1) they are inflexible and non-extensible, making it difficult to incorporate analyst expertise, and (2) they are memory inefficient, often requiring>100GB of RAM to hold entire event streams, which fundamentally limits scalability and deployment in real-world environments. To address these limitations, we propose the ProGQL framework, which provides a domain-specific graph search language with a well-engineered query engine, allowing PA over system audit events and expert knowledge to be jointly expressed as a graph search query and thereby facilitating the investigation of complex cyberattacks. In particular, to support dependency searches from a starting edge required in PA, ProGQL introduces new language constructs for constrained graph traversal, edge weight computation, value propagation along weighted edges, and graph merging to integrate multiple searches. Moreover, the ProGQL query engine is optimized for efficient incremental graph search across heterogeneous database backends, eliminating the need for full in-memory materialization and reducing memory overhead. Our evaluations on real attacks demonstrate the effectiveness of the ProGQL language in expressing a diverse set of complex attacks compared with the state-of-the-art graph query language Cypher, and the comparison with the SOTA PA technique DEPIMPACT further demonstrates the significant improvement of the scalability brought by our ProGQL framework's design.

Patch backporting, the process of migrating mainline security patches to older branches, is an essential task in maintaining popular open-source projects (e.g., Linux kernel). However, manual backporting can be labor-intensive, while existing automated methods, which heavily rely on predefined syntax or semantic rules, often lack agility for complex patches. In this paper, we introduce PORTGPT, an LLM-agent for end-to-end automation of patch backporting in real-world scenarios. PORTGPT enhances an LLM with tools to access code on-demand, summarize Git history, and revise patches autonomously based on feedback (e.g., from compilers), hence, simulating human-like reasoning and verification. PORTGPT achieved an 89.15% success rate on existing datasets (1815 cases), and 62.33% on our own dataset of 146 complex cases, both outperforms state-of-the-art of backporting tools. We contributed 9 backported patches from PORTGPT to the Linux kernel community and all patches are now merged.

Deep neural networks can convert ECG page images into analyzable waveforms, yet centralized training often conflicts with cross-institutional privacy and deployment constraints. A cross-silo federated digitization framework is presented that trains a full-model nnU-Net segmentation backbone without sharing images and aggregates updates across sites under realistic non-IID heterogeneity (layout, grid style, scanner profile, noise). The protocol integrates three standard server-side aggregators–FedAvg, FedProx, and FedAdam–and couples secure aggregation with central, user-level differential privacy to align utility with formal guarantees. Key features include: (i) end-to-end full-model training and synchronization across clients; (ii) secure aggregation so the server only observes a clipped, weighted sum once a participation threshold is met; (iii) central Gaussian DP with Renyi accounting applied post-aggregation for auditable user-level privacy; and (iv) a calibration-aware digitization pipeline comprising page normalization, trace segmentation, grid-leakage suppression, and vectorization to twelve-lead signals. Experiments on ECG pages rendered from PTB-XL show consistently faster convergence and higher late-round plateaus with adaptive server updates (FedAdam) relative to FedAvg and FedProx, while approaching centralized performance. The privacy mechanism maintains competitive accuracy while preventing exposure of raw images or per-client updates, yielding deployable, auditable guarantees suitable for multi-institution settings.

Using risky text prompts, such as pornography and violent prompts, to test the safety of text-to-image (T2I) models is a critical task. However, existing risky prompt datasets are limited in three key areas: 1) limited risky categories, 2) coarse-grained annotation, and 3) low effectiveness. To address these limitations, we introduce T2I-RiskyPrompt, a comprehensive benchmark designed for evaluating safety-related tasks in T2I models. Specifically, we first develop a hierarchical risk taxonomy, which consists of 6 primary categories and 14 fine-grained subcategories. Building upon this taxonomy, we construct a pipeline to collect and annotate risky prompts. Finally, we obtain 6,432 effective risky prompts, where each prompt is annotated with both hierarchical category labels and detailed risk reasons. Moreover, to facilitate the evaluation, we propose a reason-driven risky image detection method that explicitly aligns the MLLM with safety annotations. Based on T2I-RiskyPrompt, we conduct a comprehensive evaluation of eight T2I models, nine defense methods, five safety filters, and five attack strategies, offering nine key insights into the strengths and limitations of T2I model safety. Finally, we discuss potential applications of T2I-RiskyPrompt across various research fields. The dataset and code are provided in https://github.com/datar001/T2I-RiskyPrompt.

Wide-bandgap (WBG) technologies offer unprecedented improvements in power system efficiency, size, and performance, but also introduce unique sensor corruption and cybersecurity risks in industrial control systems (ICS), particularly due to high-frequency noise and sophisticated cyber-physical threats. This proof-of-concept (PoC) study demonstrates the adaptation of a noise-driven physically unclonable function (PUF) and machine learning (ML)-assisted anomaly detection framework to the demanding environment of WBG-based ICS sensor pathways. By extracting entropy from unavoidable WBG switching noise (up to 100 kHz) as a PUF source, and simultaneously using this noise as a real-time threat indicator, the proposed system unites hardware-level authentication and anomaly detection. Our approach integrates hybrid machine learning (ML) models with adaptive Bayesian filtering, providing robust and low-latency detection capabilities resilient to both natural electromagnetic interference (EMI) and active adversarial manipulation. Through detailed simulations of WBG modules under benign and attack scenarios–including EMI injection, signal tampering, and node impersonation–we achieve 95% detection accuracy and sub-millisecond processing latency. These results demonstrate the feasibility of physics-driven, dual-use noise exploitation as a scalable ICS defense primitive. Our findings lay the groundwork for next-generation security strategies that leverage inherent device characteristics, bridging hardware and artificial intelligence (AI) for enhanced protection of critical ICS infrastructure.

Data poisoning attacks are a potential threat to machine learning (ML) models, aiming to manipulate training datasets to disrupt their performance. Existing defenses are mostly designed to mitigate specific poisoning attacks or are aligned with particular ML algorithms. Furthermore, most defenses are developed to secure deep neural networks or binary classifiers. However, traditional multiclass classifiers need attention to be secure from data poisoning attacks, as these models are significant in developing multi-modal applications. Therefore, this paper proposes SecureLearn, a two-layer attack-agnostic defense to defend multiclass models from poisoning attacks. It comprises two components of data sanitization and a new feature-oriented adversarial training. To ascertain the effectiveness of SecureLearn, we proposed a 3D evaluation matrix with three orthogonal dimensions: data poisoning attack, data sanitization and adversarial training. Benchmarking SecureLearn in a 3D matrix, a detailed analysis is conducted at different poisoning levels (10%-20%), particularly analysing accuracy, recall, F1-score, detection and correction rates, and false discovery rate. The experimentation is conducted for four ML algorithms, namely Random Forest (RF), Decision Tree (DT), Gaussian Naive Bayes (GNB) and Multilayer Perceptron (MLP), trained with three public datasets, against three poisoning attacks and compared with two existing mitigations. Our results highlight that SecureLearn is effective against the provided attacks. SecureLearn has strengthened resilience and adversarial robustness of traditional multiclass models and neural networks, confirming its generalization beyond algorithm-specific defenses. It consistently maintained accuracy above 90%, recall and F1-score above 75%. For neural networks, SecureLearn achieved 97% recall and F1-score against all selected poisoning attacks.

Provenance analysis based on system audit data has emerged as a fundamental approach for investigating Advanced Persistent Threat (APT) attacks. Due to the high concealment and long-term persistence of APT attacks, they are only represented as a minimal part of the critical path in the provenance graph. While existing techniques employ behavioral pattern matching and data flow feature matching to uncover latent associations in attack sequences through provenance graph path reasoning, their inability to establish effective attack context associations often leads to the conflation of benign system operations with real attack entities, that fail to accurately characterize real APT behaviors. We observe that while the causality of entities in the provenance graph exhibit substantial complexity, attackers often follow specific attack patterns-specifically, clear combinations of tactics and techniques to achieve their goals. Based on these insights, we propose TPPR, a novel framework that first extracts anomaly subgraphs through abnormal node detection, TTP-annotation and graph pruning, then performs attack path reasoning using mined TTP sequential pattern, and finally reconstructs attack scenarios through confidence-based path scoring and merging. Extensive evaluation on real enterprise logs (more than 100 million events) and DARPA TC dataset demonstrates TPPR's capability to achieve 99.9% graph simplification (700,000 to 20 edges) while preserving 91% of critical attack nodes, outperforming state-of-the-art solutions (SPARSE, DepImpact) by 63.1% and 67.9% in reconstruction precision while maintaining attack scenario integrity.

Federated learning (FL) has emerged as a promising paradigm for decentralized model training, enabling multiple clients to collaboratively learn a shared model without exchanging their local data. However, the decentralized nature of FL also introduces vulnerabilities, as malicious clients can compromise or manipulate the training process. In this work, we introduce dictator clients, a novel, well-defined, and analytically tractable class of malicious participants capable of entirely erasing the contributions of all other clients from the server model, while preserving their own. We propose concrete attack strategies that empower such clients and systematically analyze their effects on the learning process. Furthermore, we explore complex scenarios involving multiple dictator clients, including cases where they collaborate, act independently, or form an alliance in order to ultimately betray one another. For each of these settings, we provide a theoretical analysis of their impact on the global model's convergence. Our theoretical algorithms and findings about the complex scenarios including multiple dictator clients are further supported by empirical evaluations on both computer vision and natural language processing benchmarks.

Wi-Fi Channel State Information (CSI) has been extensively studied for sensing activities. However, its practical application in user authentication still needs to be explored. This study presents a novel approach to biometric authentication using Wi-Fi Channel State Information (CSI) data for palm recognition. The research delves into utilizing a Raspberry Pi encased in a custom-built box with antenna power reduced to 1dBm, which was used to capture CSI data from the right hands of 20 participants (10 men and 10 women). The dataset was normalized using MinMax scaling to ensure uniformity and accuracy. By focusing on biophysical aspects such as hand size, shape, angular spread between fingers, and finger phalanx lengths, among other characteristics, the study explores how these features affect electromagnetic signals, which are then reflected in Wi-Fi CSI, allowing for precise user identification. Five classification algorithms were evaluated, with the Random Forest classifier achieving an average F1-Score of 99.82% using 10-fold cross-validation. Amplitude and Phase data were used, with each capture session recording approximately 1000 packets per second in five 5-second intervals for each User. This high accuracy highlights the potential of Wi-Fi CSI in developing robust and reliable user authentication systems based on palm biometric data.

The Internet of Things (IoT) relies heavily on resource-limited devices to communicate critical (e.g., military data) information under low-energy adversarial environments and low-latency wireless channels. Authenticated Encryption (AE) guarantees confidentiality, authenticity, and integrity, making it a vital security service for IoT. However, current deployed (lightweight) AE standards lack essential features like key compromise resiliency and compact authentication tags, as well as performance enhancements such as offline-online cryptography. To address these gaps, we propose Graphene, the first (to our knowledge) symmetric Forward-secure and Aggregate Authenticated Encryption (FAAE) framework designed for the performance and security demands of low-end IoT infrastructures. Graphene innovates by synergizing key evolution strategies and offline-online cryptographic processing with Universal Message Authentication Codes (UMACs) to guarantee breach-resiliency, near-optimal online latency, and compactness. We demonstrate Graphene efficiency through two distinct instantiations, each balancing unique performance trade-offs with extensibility for diverse MACs. Our experimental evaluation on commodity hardware and 32-bit ARM Cortex-M4 microcontroller shows Graphene significant performance gains over existing alternatives. Graphene is also backward compatible with standard-compliant cryptographic implementations. We release our implementation as open source for public testing and adaptation.

The globalized semiconductor supply chain has made Hardware Trojans (HT) a significant security threat to embedded systems, necessitating the design of efficient and adaptable detection mechanisms. Despite promising machine learning-based HT detection techniques in the literature, they suffer from ad hoc feature selection and the lack of adaptivity, all of which hinder their effectiveness across diverse HT attacks. In this paper, we propose SAND, a selfsupervised and adaptive NAS-driven framework for efficient HT detection. Specifically, this paper makes three key contributions. (1) We leverage self-supervised learning (SSL) to enable automated feature extraction, eliminating the dependency on manually engineered features. (2) SAND integrates neural architecture search (NAS) to dynamically optimize the downstream classifier, allowing for seamless adaptation to unseen benchmarks with minimal fine-tuning. (3) Experimental results show that SAND achieves a significant improvement in detection accuracy (up to 18.3%) over state-of-the-art methods, exhibits high resilience against evasive Trojans, and demonstrates strong generalization.

Large language models (LLMs) remain vulnerable to sophisticated prompt engineering attacks that exploit contextual framing to bypass safety mechanisms, posing significant risks in cybersecurity applications. We introduce Jailbreak Mimicry, a systematic methodology for training compact attacker models to automatically generate narrative-based jailbreak prompts in a one-shot manner. Our approach transforms adversarial prompt discovery from manual craftsmanship into a reproducible scientific process, enabling proactive vulnerability assessment in AI-driven security systems. Developed for the OpenAI GPT-OSS-20B Red-Teaming Challenge, we use parameter-efficient fine-tuning (LoRA) on Mistral-7B with a curated dataset derived from AdvBench, achieving an 81.0% Attack Success Rate (ASR) against GPT-OSS-20B on a held-out test set of 200 items. Cross-model evaluation reveals significant variation in vulnerability patterns: our attacks achieve 66.5% ASR against GPT-4, 79.5% on Llama-3 and 33.0% against Gemini 2.5 Flash, demonstrating both broad applicability and model-specific defensive strengths in cybersecurity contexts. This represents a 54x improvement over direct prompting (1.5% ASR) and demonstrates systematic vulnerabilities in current safety alignment approaches. Our analysis reveals that technical domains (Cybersecurity: 93% ASR) and deception-based attacks (Fraud: 87.8% ASR) are particularly vulnerable, highlighting threats to AI-integrated threat detection, malware analysis, and secure systems, while physical harm categories show greater resistance (55.6% ASR). We employ automated harmfulness evaluation using Claude Sonnet 4, cross-validated with human expert assessment, ensuring reliable and scalable evaluation for cybersecurity red-teaming. Finally, we analyze failure mechanisms and discuss defensive strategies to mitigate these vulnerabilities in AI for cybersecurity.

Sparse variable selection improves interpretability and generalization in high-dimensional learning by selecting a small subset of informative features. Recent advances in Mixed Integer Programming (MIP) have enabled solving large-scale non-private sparse regression - known as Best Subset Selection (BSS) - with millions of variables in minutes. However, extending these algorithmic advances to the setting of Differential Privacy (DP) has remained largely unexplored. In this paper, we introduce two new pure differentially private estimators for sparse variable selection, levering modern MIP techniques. Our framework is general and applies broadly to problems like sparse regression or classification, and we provide theoretical support recovery guarantees in the case of BSS. Inspired by the exponential mechanism, we develop structured sampling procedures that efficiently explore the non-convex objective landscape, avoiding the exhaustive combinatorial search in the exponential mechanism. We complement our theoretical findings with extensive numerical experiments, using both least squares and hinge loss for our objective function, and demonstrate that our methods achieve state-of-the-art empirical support recovery, outperforming competing algorithms in settings with up to $p=10^4$.