arXiv subjects

Mohamed Nafea

Publications and source records attributed to Mohamed Nafea.

10 recordsLinked to original sources

arXiv2026

Causal Discovery in Linear Models with Unobserved Variables and Measurement Error

The presence of unobserved common causes and measurement error poses two major obstacles to causal structure learning, since ignoring either source of complexity can induce spurious causal relations among variables of interest. We study causal structure learning in linear systems where both challenges may occur simultaneously. We introduce a causal model called LV-SEM-ME, which contains four types of variables: directly observed variables, variables that are not directly observed but are measured with error, the corresponding measurements, and variables that are neither observed nor measured. Under a separability condition-namely, identifiability of the mixing matrix associated with the exogenous noise terms of the observed variables-together with certain faithfulness assumptions, we characterize the extent of identifiability and the corresponding observational equivalence classes. We provide graphical characterizations of these equivalence classes and develop recovery algorithms that enumerate all models in the equivalence class of the ground truth. We also establish, via a four-node union model that subsumes instrumental variable, front-door, and negative-control-outcome settings, a form of identification robustness: the target effect remains identifiable in the broader LV-SEM-ME model even when the assumptions underlying the specialized identification formulas for the corresponding submodels need not all hold simultaneously.

cs.LG↗

arXiv2025

Hierarchical Attention Network for Interpretable ECG-based Heart Disease Classification

Cardiovascular disease remains one of the leading causes of mortality worldwide, underscoring the need for accurate as well as interpretable diagnostic machine learning tools. In this work, we investigate heart disease classification using electrocardiogram (ECG) data from two widely-utilized datasets: The MIT-BIH Arrhythmia and the PTB-XL datasets. We adapt a hierarchical attention network (HAN), originally developed for text classification, into an ECG-based heart-disease classification task. Our adapted HAN incorporates two attention layers that focus on ECG data segments of varying sizes. We conduct a comparative analysis between our adapted HAN and a more sophisticated state-of-the-art architecture, featuring a network with convolution, attention, and transformer layers (CAT-Net). Our empirical evaluation encompasses multiple aspects including test accuracy (quantified by 0-1 loss); model complexity (measured by the number of model parameters); and interpretability (through attention map visualization). Our adapted HAN demonstrates comparable test accuracy with significant reductions in model complexity and enhanced interpretability analysis: For the MIT-BIH dataset, our adapted HAN achieves 98.55\% test accuracy compared to 99.14\% for CAT-Net, while reducing the number of model parameters by a factor of 15.6. For the PTB-XL dataset, our adapted HAN achieves a 19.3-fold reduction in model complexity compared to CAT-Net, with only a 5\% lower test accuracy. From an interpretability perspective, the significantly simpler architecture and the hierarchical nature of our adapted HAN model facilitate a more straightforward interpretability analysis based on visualizing attention weights. Building on this advantage, we conduct an interpretability analysis of our HAN that highlights the regions of the ECG signal most relevant to the model's decisions.

eess.SP↗

arXiv2024

Centralized and Federated Heart Disease Classification Models Using UCI Dataset and their Shapley-value Based Interpretability

Cardiovascular diseases are a leading cause of mortality worldwide, highlighting the need for accurate diagnostic methods. This study benchmarks centralized and federated machine learning algorithms for heart disease classification using the UCI dataset which includes 920 patient records from four hospitals in the USA, Hungary and Switzerland. Our benchmark is supported by Shapley-value interpretability analysis to quantify features' importance for classification. In the centralized setup, various binary classification algorithms are trained on pooled data, with a support vector machine (SVM) achieving the highest testing accuracy of 83.3\%, surpassing the established benchmark of 78.7\% with logistic regression. Additionally, federated learning algorithms with four clients (hospitals) are explored, leveraging the dataset's natural partition to enhance privacy without sacrificing accuracy. Federated SVM, an uncommon approach in the literature, achieves a top testing accuracy of 73.8\%. Our interpretability analysis aligns with existing medical knowledge of heart disease indicators. Overall, this study establishes a benchmark for efficient and interpretable pre-screening tools for heart disease while maintaining patients' privacy. This work is available at https://github.com/padillma1/Heart-Disease-Classification-on-UCI-dataset-and-Shapley-Interpretability-Analysis.

cs.LG↗

arXiv2022

Causal Discovery in Linear Structural Causal Models with Deterministic Relations

Linear structural causal models (SCMs) -- in which each observed variable is generated by a subset of the other observed variables as well as a subset of the exogenous sources -- are pervasive in causal inference and casual discovery. However, for the task of causal discovery, existing work almost exclusively focus on the submodel where each observed variable is associated with a distinct source with non-zero variance. This results in the restriction that no observed variable can deterministically depend on other observed variables or latent confounders. In this paper, we extend the results on structure learning by focusing on a subclass of linear SCMs which do not have this property, i.e., models in which observed variables can be causally affected by any subset of the sources, and are allowed to be a deterministic function of other observed variables or latent confounders. This allows for a more realistic modeling of influence or information propagation in systems. We focus on the task of causal discovery form observational data generated from a member of this subclass. We derive a set of necessary and sufficient conditions for unique identifiability of the causal structure. To the best of our knowledge, this is the first work that gives identifiability results for causal discovery under both latent confounding and deterministic relationships. Further, we propose an algorithm for recovering the underlying causal structure when the aforementioned conditions are satisfied. We validate our theoretical results both on synthetic and real datasets.

cs.LG↗

arXiv2022

Causal Discovery in Linear Latent Variable Models Subject to Measurement Error

We focus on causal discovery in the presence of measurement error in linear systems where the mixing matrix, i.e., the matrix indicating the independent exogenous noise terms pertaining to the observed variables, is identified up to permutation and scaling of the columns. We demonstrate a somewhat surprising connection between this problem and causal discovery in the presence of unobserved parentless causes, in the sense that there is a mapping, given by the mixing matrix, between the underlying models to be inferred in these problems. Consequently, any identifiability result based on the mixing matrix for one model translates to an identifiability result for the other model. We characterize to what extent the causal models can be identified under a two-part faithfulness assumption. Under only the first part of the assumption (corresponding to the conventional definition of faithfulness), the structure can be learned up to the causal ordering among an ordered grouping of the variables but not all the edges across the groups can be identified. We further show that if both parts of the faithfulness assumption are imposed, the structure can be learned up to a more refined ordered grouping. As a result of this refinement, for the latent variable model with unobserved parentless causes, the structure can be identified. Based on our theoretical results, we propose causal structure learning methods for both models, and evaluate their performance on synthetic data.

cs.LG↗

arXiv2021

Information Theoretic Measures for Fairness-aware Feature Selection

Machine learning algorithms are increasingly used for consequential decision making regarding individuals based on their relevant features. Features that are relevant for accurate decisions may however lead to either explicit or implicit forms of discrimination against unprivileged groups, such as those of certain race or gender. This happens due to existing biases in the training data, which are often replicated or even exacerbated by the learning algorithm. Identifying and measuring these biases at the data level is a challenging problem due to the interdependence among the features, and the decision outcome. In this work, we develop a framework for fairness-aware feature selection which takes into account the correlation among the features and the decision outcome, and is based on information theoretic measures for the accuracy and discriminatory impacts of features. In particular, we first propose information theoretic measures which quantify the impact of different subsets of features on the accuracy and discrimination of the decision outcomes. We then deduce the marginal impact of each feature using Shapley value function; a solution concept in cooperative game theory used to estimate marginal contributions of players in a coalitional game. Finally, we design a fairness utility score for each feature (for feature selection) which quantifies how this feature influences accurate as well as nondiscriminatory decisions. Our framework depends on the joint statistics of the data rather than a particular classifier design. We examine our proposed framework on real and synthetic data to evaluate its performance.

cs.LG↗

arXiv2021

Coded Caching in the Presence of a Wire and a Cache Tapping Adversary of Type II

This paper introduces the notion of cache-tapping into the information theoretic models of coded caching. The wiretap channel II in the presence of multiple receivers equipped with fixed-size cache memories, and an adversary which selects symbols to tap into from cache placement and/or delivery is introduced. The legitimate terminals know neither whether placement, delivery, or both are tapped, nor the positions in which they are tapped. Only the size of the overall tapped set is known. For two receivers and two files, the strong secrecy capacity -- the maximum achievable file rate while keeping the overall library strongly secure -- is identified. Lower and upper bounds on the strong secrecy file rate are derived when the library has more than two files. Achievability relies on a code design which combines wiretap coding, security embedding codes, one-time pad keys, and coded caching. A genie-aided upper bound, in which the transmitter is provided with user demands before placement, establishes the converse for the two-files case. For more than two files, the upper bound is constructed by three successive channel transformations. Our results establish provable security guarantees against a powerful adversary which optimizes its tapping over both phases of communication in a cache-aided system.

cs.IT↗

arXiv2018

Generalizing Multiple Access Wiretap and Wiretap II Channel Models: Achievable Rates and Cost of Strong Secrecy

In this paper, new two-user multiple access wiretap channel models are studied. First, the multiple access wiretap channel II with a discrete memoryless main channel, under three different wiretapping scenarios, is introduced. The wiretapper, as in the classical wiretap channel II model, chooses a fixed-length subset of the channel uses on which she obtains noise-free observations of one of the codewords, a superposition of the two codewords, or each of the two codewords. These thus extend the recently examined wiretap channel II with a noisy main channel to a multiple access setting with a variety of attack models for the wiretapper. Next, a new multiple access wiretap channel model, which further generalizes the multiple access wiretap channel II under the third wiretapping scenario, i.e., that which features the strongest adversarial model, is proposed. In this model, the wiretapper, besides choosing a subset of the channel uses to noiselessly observe the transmitted codeword symbols of both users, observes the remainder of the two codewords through a discrete memoryless multiple access channel. Achievable strong secrecy rate regions for all the proposed models are derived. Achievability is established by solving dual multi-terminal secret key agreement problems in the source model, and converting the solution to the original channel models using probability distribution approximation arguments. The derived achievable rate regions quantify the secrecy cost due to the additional capabilities of the wiretapper with respect to the previous multiple access wiretap models.

cs.IT↗

arXiv2017

A New Wiretap Channel Model and its Strong Secrecy Capacity

In this paper, a new wiretap channel model is proposed, where the legitimate transmitter and receiver communicate over a discrete memoryless channel. The wiretapper has perfect access to a fixed-length subset of the transmitted codeword symbols of her choosing. Additionally, she observes the remainder of the transmitted symbols through a discrete memoryless channel. This new model subsumes the classical wiretap channel and wiretap channel II with noisy main channel as its special cases. The strong secrecy capacity of the proposed channel model is identified. Achievability is established by solving a dual secret key agreement problem in the source model, and converting the solution to the original channel model using probability distribution approximation arguments. In the dual problem, a source encoder and decoder, who observe random sequences independent and identically distributed according to the input and output distributions of the legitimate channel in the original problem, communicate a confidential key over a public error-free channel using a single forward transmission, in the presence of a compound wiretapping source who has perfect access to the public discussion. The security of the key is guaranteed for the exponentially many possibilities of the subset chosen at wiretapper by deriving a lemma which provides a doubly-exponential convergence rate for the probability that, for a fixed choice of the subset, the key is uniform and independent from the public discussion and the wiretapping source's observation. The converse is derived by using Sanov's theorem to upper bound the secrecy capacity of the new wiretap channel model by the secrecy capacity when the tapped subset is randomly chosen by nature.

cs.IT↗

arXiv2015

Secure Degrees of Freedom for the MIMO Wire-tap Channel with a Multi-antenna Cooperative Jammer

In this paper, a multiple antenna wire-tap channel in the presence of a multi-antenna cooperative jammer is studied. In particular, the secure degrees of freedom (s.d.o.f.) of this channel is established, with $N_t$ antennas at the transmitter, $N_r$ antennas at the legitimate receiver, and $N_e$ antennas at the eavesdropper, for all possible values of the number of antennas, $N_c$, at the cooperative jammer. In establishing the result, several different ranges of $N_c$ need to be considered separately. The lower and upper bounds for these ranges of $N_c$ are derived, and are shown to be tight. The achievability techniques developed rely on a variety of signaling, beamforming, and alignment techniques which vary according to the (relative) number of antennas at each terminal and whether the s.d.o.f. is integer valued. Specifically, it is shown that, whenever the s.d.o.f. is integer valued, Gaussian signaling for both transmission and cooperative jamming, linear precoding at the transmitter and the cooperative jammer, and linear processing at the legitimate receiver, are sufficient for achieving the s.d.o.f. of the channel. By contrast, when the s.d.o.f. is not an integer, the achievable schemes need to rely on structured signaling at the transmitter and the cooperative jammer, and joint signal space and signal scale alignment. The converse is established by combining an upper bound which allows for full cooperation between the transmitter and the cooperative jammer, with another upper bound which exploits the secrecy and reliability constraints.

cs.IT↗