Razane Tajeddine

Bochra Al Agha, Razane Tajeddine

Benchmarking presence-only passive reconnaissance in smart-grid communications is challenging because the adversary is receive-only, yet nearby observers can still alter propagation through additional shadowing and multipath that reshapes channel coherence. Public smart-grid cybersecurity datasets largely target active protocol- or measurement-layer attacks and rarely provide propagation-driven observables with tiered topology context, which limits reproducible evaluation under strictly passive threat models. This paper introduces an IEEE-inspired, literature-anchored benchmark dataset generator for passive reconnaissance over a tiered Home Area Network (HAN), Neighborhood Area Network (NAN), and Wide Area Network (WAN) communication graph with heterogeneous wireless and wireline links. Node-level time series are produced through a physically consistent channel-to-metrics mapping where channel state information (CSI) is represented via measurement-realistic amplitude and phase proxies that drive inferred signal-to-noise ratio (SNR), packet error behavior, and delay dynamics. Passive attacks are modeled only as windowed excess attenuation and coherence degradation with increased channel innovation, so reliability and latency deviations emerge through the same causal mapping without labels or feature shortcuts. The release provides split-independent realizations with burn-in removal, strictly causal temporal descriptors, adjacency-weighted neighbor aggregates and deviation features, and federated-ready per-node train, validation, and test partitions with train-only normalization metadata. Baseline federated experiments highlight technology-dependent detectability and enable standardized benchmarking of graph-temporal and federated detectors for passive reconnaissance.

Ali Al Sahili, Ali Chehab, Razane Tajeddine

Large Language Models (LLMs) are prone to memorizing training data, which poses serious privacy risks. Two of the most prominent concerns are training data extraction and Membership Inference Attacks (MIAs). Prior research has shown that these threats are interconnected: adversaries can extract training data from an LLM by querying the model to generate a large volume of text and subsequently applying MIAs to verify whether a particular data point was included in the training set. In this study, we integrate multiple MIA techniques into the data extraction pipeline to systematically benchmark their effectiveness. We then compare their performance in this integrated setting against results from conventional MIA benchmarks, allowing us to evaluate their practical utility in real-world extraction scenarios.

Bochra Al Agha, Razane Tajeddine

Smart grids are exposed to passive eavesdropping, where attackers listen silently to communication links. Although no data is actively altered, such reconnaissance can reveal grid topology, consumption patterns, and operational behavior, creating a gateway to more severe targeted attacks. Detecting this threat is difficult because the signals it produces are faint, short-lived, and often disappear when traffic is examined by a single node or along a single timeline. This paper introduces a graph-centric, multimodal detector that fuses physical-layer and behavioral indicators over ego-centric star subgraphs and short temporal windows to detect passive attacks. To capture stealthy perturbations, a two-stage encoder is introduced: graph convolution aggregates spatial context across ego-centric star subgraphs, while a bidirectional GRU models short-term temporal dependencies. The encoder transforms heterogeneous features into a unified spatio-temporal representation suitable for classification. Training occurs in a federated learning setup under FedProx, improving robustness to heterogeneous local raw data and contributing to the trustworthiness of decentralized training; raw measurements remain on client devices. A synthetic, standards-informed dataset is generated to emulate heterogeneous HAN/NAN/WAN communications with wireless-only passive perturbations, event co-occurrence, and leak-safe splits. The model achieves a testing accuracy of 98.32% per-timestep (F1_{attack}=0.972) and 93.35% per-sequence at 0.15% FPR using a simple decision rule with run-length m=2 and threshold $τ=0.55$. The results demonstrate that combining spatial and temporal context enables reliable detection of stealthy reconnaissance while maintaining low false-positive rates, making the approach suitable for non-IID federated smart-grid deployments.

Chaymaa Abbas, Mariette Awad, Razane Tajeddine

Style-conditioned data poisoning is identified as a covert vector for amplifying sociolinguistic bias in large language models. Using small poisoned budgets that pair dialectal prompts -- principally African American Vernacular English (AAVE) and a Southern dialect -- with toxic or stereotyped completions during instruction tuning, this work probes whether linguistic style can act as a latent trigger for harmful behavior. Across multiple model families and scales, poisoned exposure elevates toxicity and stereotype expression for dialectal inputs -- most consistently for AAVE -- while Standard American English remains comparatively lower yet not immune. A multi-metric audit combining classifier-based toxicity with an LLM-as-a-judge reveals stereotype-laden content even when lexical toxicity appears muted, indicating that conventional detectors under-estimate sociolinguistic harms. Additionally, poisoned models exhibit emergent jailbreaking despite the absence of explicit slurs in the poison, suggesting weakened alignment rather than memorization. These findings underscore the need for dialect-aware evaluation, content-level stereotype auditing, and training protocols that explicitly decouple style from toxicity to prevent bias amplification through seemingly minor, style-based contamination.

Linzh Zhao, Aki Rehn, Mikko A. Heikkilä et al.

Differential privacy (DP) has become an essential framework for privacy-preserving machine learning. Existing DP learning methods, however, often have disparate impacts on model predictions, e.g., for minority groups. Gradient clipping, which is often used in DP learning, can suppress larger gradients from challenging samples. We show that this problem is amplified by adaptive clipping, which will often shrink the clipping bound to tiny values to match a well-fitting majority, while significantly reducing the accuracy for others. We propose bounded adaptive clipping, which introduces a tunable lower bound to prevent excessive gradient suppression. Our method improves the accuracy of the worst-performing class on average over 10 percentage points on skewed MNIST and Fashion MNIST compared to the unbounded adaptive clipping, and over 5 percentage points over constant clipping.

Razane Tajeddine, Joonas Jälkö, Samuel Kaski et al.

In this work, we introduce a differentially private method for generating synthetic data from vertically partitioned data, \emph{i.e.}, where data of the same individuals is distributed across multiple data holders or parties. We present a differentially privacy stochastic gradient descent (DP-SGD) algorithm to train a mixture model over such partitioned data using variational inference. We modify a secure multiparty computation (MPC) framework to combine MPC with differential privacy (DP), in order to use differentially private MPC effectively to learn a probabilistic generative model under DP on such vertically partitioned data. Assuming the mixture components contain no dependencies across different parties, the objective function can be factorized into a sum of products of the contributions calculated by the parties. Finally, MPC is used to compute the aggregate between the different contributions. Moreover, we rigorously define the privacy guarantees with respect to the different players in the system. To demonstrate the accuracy of our method, we run our algorithm on the Adult dataset from the UCI machine learning repository, where we obtain comparable results to the non-partitioned case.