4 sonuçlar
Arama Sonuçları
Listeleniyor 1 - 4 / 4
Yayın ANN activation function estimators for homomorphic encrypted inference(Institute of Electrical and Electronics Engineers Inc., 2025-06-13) Harb, Mhd Raja Abou; Çeliktaş, BarışHomomorphic Encryption (HE) enables secure computations on encrypted data, facilitating machine learning inference in sensitive environments such as healthcare and finance. However, efficiently handling non-linear activation functions, specifically Sigmoid and Tanh, remains a significant computational challenge for encrypted inference using Artificial Neural Networks (ANNs). This study introduces a lightweight, ANN-based estimator designed to accurately approximate activation functions under homomorphic encryption. Unlike traditional polynomial and piecewise linear approximations, the proposed ANN estimators achieve superior accuracy with lower computational overhead associated with bootstrapping or high-degree polynomial techniques. These estimators are trained on plaintext data and seamlessly integrated into encrypted inference pipelines, significantly outperforming conventional methods. Experimental evaluations demonstrate notable improvements, with ANN estimators enhancing accuracy by approximately 2% for Sigmoid and up to 73% for Tanh functions, improving F1-scores by approximately 2% for Sigmoid and up to 88% for Tanh, and markedly reducing Mean Square Error (MSE) by up to 96% compared to polynomial approximations. The ANN estimator achieves an accuracy of 97.70% and an AUC of 0.9997 when integrated into a CNN architecture on the MNIST dataset, and an accuracy of 85.25% with an AUC of 0.9459 on the UCI Heart Disease dataset during ciphertext inference. These results underscore the estimator’s practical effectiveness and computational feasibility, making it suitable for secure and efficient ANN inference in encrypted environments.Yayın Intelligent health monitoring in 6G networks: machine learning-enhanced VLC-based medical body sensor networks(Multidisciplinary Digital Publishing Institute (MDPI), 2025-05-23) Antaki, Bilal; Dalloul, Ahmed Hany; Miramirkhani, FarshadRecent advances in Artificial Intelligence (AI)-driven wireless communication are driving the adoption of Sixth Generation (6G) technologies in crucial environments such as hospitals. Visible Light Communication (VLC) leverages existing lighting infrastructure to deliver high data rates while mitigating electromagnetic interference (EMI); however, patient movement induces fluctuating signal strength and dynamic channel conditions. In this paper, we present a novel integration of site-specific ray tracing and machine learning (ML) for VLC-enabled Medical Body Sensor Networks (MBSNs) channel modeling in distinct hospital settings. First, we introduce a Q-learning-based adaptive modulation scheme that meets target symbol error rates (SERs) in real time without prior environmental information. Second, we develop a Long Short-Term Memory (LSTM)-based estimator for path loss and Root Mean Square (RMS) delay spread under dynamic hospital conditions. To our knowledge, this is the first study combining ray-traced channel impulse response modeling (CIR) with ML techniques in hospital scenarios. The simulation results demonstrate that the Q-learning method consistently achieves SERs with a spectral efficiency (SE) lower than optimal near the threshold. Furthermore, LSTM estimation shows that D1 has the highest Root Mean Square Error (RMSE) for path loss (1.6797 dB) and RMS delay spread (1.0567 ns) in the Intensive Care Unit (ICU) ward, whereas D3 exhibits the highest RMSE for path loss (1.0652 dB) and RMS delay spread (0.7657 ns) in the Family-Type Patient Rooms (FTPRs) scenario, demonstrating high estimation accuracy under realistic conditions.Yayın Adaptive incident escalation in SOCs via AI-driven skill-aware assignment and tier optimization(Institute of Electrical and Electronics Engineers Inc., 2026-04-15) Abuaziz, Ahmed; Çeliktaş, BarışModern Security Operations Centers (SOCs) face significant operational bottlenecks driven by escalating alert volumes, increasingly sophisticated cyberattack vectors, and chronic imbalances in analyst workloads. Conventional rule-based escalation models frequently fail to account for the multi-dimensional nature of incident characteristics, the nuances of analyst expertise, and fluctuating operational demands. This study proposes a comprehensive AI-driven framework for intelligent incident assignment and workload optimization. The framework introduces five primary contributions: 1) a multi-factor scoring model that integrates severity and complexity metrics with dynamic workload balancing; 2) two novel optimization algorithms, Quantile-Targeted Normality-Regularized Optimization (QT-NRO) and Joint Optimization of Weights and Thresholds (JOWT), to calibrate scoring coefficients against target analyst utilization; 3) a Large Language Model (LLM) engine leveraging Retrieval-Augmented Generation (RAG) for semantic alignment between incident requirements and analyst expertise; 4) an Adaptive Capacity Zoning mechanism for dynamic workload management; and 5) a novel RAG Relevance Score metric—a pre-resolution, semantic alignment indicator that quantifies analyst-incident assignment quality independently of resolution time, addressing a fundamental limitation of traditional temporal metrics such as Mean Time to Resolution (MTTR) and providing a reusable benchmark applicable to any skill-aware assignment system. In addition, the framework incorporates a feedback-based continuous learning mechanism that utilizes historical resolution data to inform future assignments. An experimental evaluation using 10,021 real-world incidents from Microsoft Defender demonstrates that the JOWT algorithm achieves a tier distribution alignment within 0.8% of targets. LLM-enhanced semantic matching yields improvements between 26.7% and 126.8% in skill alignment across both normal-load and high-load evaluations, while simulations indicate a 31.8% reduction in MTTR. These results substantiate the efficacy of AI-driven methodologies in enhancing SOC operational efficiency and response precision.Yayın A deployment-oriented privacy-preserving CTI framework: integrating PIR, federated learning, differential privacy, and practical hardenings(Institute of Electrical and Electronics Engineers Inc., 2026) Çamalan, Emre; Çeliktaş, BarışThreat Intelligence Platforms (TIPs) enable organizations to share indicators of compromise (IoCs), yet the operational CTI lifecycle exposes multiple, largely independent privacy surfaces: query content and access-pattern leakage during IoC lookup, gradient and membership inference risks during collaborative model training, and residual metadata side-channels in network traffic. Existing work addresses these surfaces in isolation; no prior framework orchestrates their joint mitigation within a single, deploymentoriented CTI pipeline under explicit guarantee boundaries. We present a prototype workflow-level privacy orchestration for cyber threat intelligence that coordinates four mechanisms across the query-learn-update lifecycle: (i) Private Information Retrieval (PIR) to hide queried IoC indices, (ii) cross-silo federated learning (FL) to keep raw CTI data local, (iii) a formal client-level Differential Privacy (DP) mechanism for federated model training to protect against inversion and membership inference attacks, and (iv) practical privacy hardenings, namely fixed-shape PIR batching (a traffic-shaping mechanism, not a cryptographic PIR guarantee) and secure aggregation simulated under an honest-but-curious coordinator assumption, to mitigate residual side-channel leakage. The contribution is therefore one of CTI-specific workflow orchestration and systematic evaluation, not of new cryptographic primitives: formal (ε, δ) guarantees apply exclusively to the differentially private federated learning component, while the remaining mechanisms serve as deployment-oriented hardenings under stated assumptions. We implement a working prototype over a two-million-row AbuseIPDB-style IoC dataset. Under a two-server non-colluding assumption, PIR queries complete in approximately 40 seconds with 16MB transfer per fixed batch. Local Random Forest and Logistic Regression baselines reach 89.0% and 77.00% accuracy, respectively, while federated variants with DP-FedAvg (gradient clipping and RDP-based privacy accounting) demonstrate a quantified privacy–utility trade-off across multiple noise levels. A corrected canonical single-round (T=1) baseline establishes the reconciled reference operating point; reviewer-driven multi-round experiments (T ∈ {1, 10, 20}) and an auxiliary clip-norm sensitivity analysis (C ∈ {0.5, 1.0, 2.0}) further characterize how privacy budgets, model utility, and training stability evolve beyond the single-round setting, with all (ε, δ) values computed via RDP composition for the corresponding configuration. The framework aligns with recent advances in secure aggregation and privacy-preserving CTI analytics, and is designed to be compatible with GDPR, CCPA, ISO/IEC 27701, and NIST 800-53 privacy principles, demonstrating prototype-level feasibility for regulation-aware CTI collaboration across organizations.
