Department of Electrical Engineering

Permanent URI for this collection

http://hdl.handle.net/20.500.12519/625

Browse

Browsing Department of Electrical Engineering by Title

  • Item
    Arithmetic properties of complex fibonacci numbers and fibonacci quaternions
    (SAS International Publications, 2021-09) Leung, Ho-Hon ; Kamalov, Firuz
    In this paper, we investigate certain arithmetic properties of complex Fibonacci numbers and Fibonacci quaternions. More specifically, we look at the divisibility properties of complex Fibonacci numbers and Fibonacci quaternions. Our results make use of some well-known Fibonacci identities. Since quaternions are non-commutative algebra, extra care has been taken to investigate the various divisibility properties of the Fibonacci quaternions. © SAS International Publications.
  • Item
    Audio steganalysis based on lossless data-compression techniques
    (Springer Nature Switzerland AG, 2012) Djebbar, Fatiha ; Ayad, Beghdad
    In this paper, we introduce a new blind steganalysis method that can reliably detect modifications in audio signals due to steganography. Lossless data-compression ratios are computed from the testing signals and their reference versions and used as features for the classifier design. Additionally, we propose to extract additional features from different energy parts of each tested audio signal to retrieve more informative data and enhance the classifier capability. Support Vector Machine (SVM) is employed to discriminate between the cover- and the stego-audio signals. Experimental results show that our method performs very well and achieves very good detection rates of stego-audio signals produced by S-tools4, Steghide and Hide4PGP. © 2012 Springer-Verlag.
  • Item
    Auditory-based subband blind source separation using sample-by-sample and Infomax algorithms
    ( 2010) Salem, Abderraouf Ben ; Selouani, Sid Ahmed ; Hamam, Habib
    We present a new subband decomposition method for the separation of convolutive mixtures of speech. This method uses a sample-by-sample algorithm to perform the subband decomposition by mimicking the processing performed by the human ear. The unknown source signals are separated by maximizing the entropy of a transformed set of signal mixtures through the use of a gradient ascent algorithm. Experimental results show the efficiency of the proposed approach in terms of signal-to-interference ratio. Compared with the fullband method that uses the Infomax algorithm, our method shows an important improvement of the output signal-to-noise ratio when the sensor inputs are severely degraded by additive noise.
  • Item
    Autocorrelation for time series with linear trend
    (Institute of Electrical and Electronics Engineers Inc., 2021-09-29) Kamalov, Firuz ; Thabtah, Fadi ; Gurrib, Ikhlaas
    The autocorrelation function (ACF) is a fundamental concept in time series analysis including financial forecasting. In this note, we investigate the properties of the sample ACF for a time series with linear trend. In particular, we show that the sample ACF of the time series approaches 1 for all lags as the number of time steps increases. The theoretical results are supported by numerical experiments. Our result helps researchers better understand the ACF patterns and make correct ARMA selection. © 2021 IEEE.
  • Item
    Autoencoder-based Intrusion Detection System
    (Institute of Electrical and Electronics Engineers Inc., 2021) Kamalov, Firuz ; Zgheib, Rita ; Leung, Ho Hon ; Al-Gindy, Ahmed ; Moussa, Sherif
    Given the dependence of the modern society on networks, the importance of effective intrusion detection systems (IDS) cannot be underestimated. In this paper, we consider an autoencoder-based IDS for detecting distributed denial of service attacks (DDoS). The advantage of autoencoders over traditional machine learning methods is the ability to train on unlabeled data. As a result, autoencoders are well-suited for detecting unknown attacks. The key idea of the proposed approach is that anomalous traffic flows will have higher reconstruction loss which can be used to flag the intrusions. The results of numerical experiments show that the proposed method outperforms benchmark unsupervised algorithms in detecting DDoS attacks. © 2021 IEEE.
  • Item
    Autoregressive and neural network models: A comparative study with linearly lagged series
    (Institute of Electrical and Electronics Engineers Inc., 2021-09-29) Kamalov, Firuz ; Gurrib, Ikhlaas ; Thabtah, Fadi
    Time series analysis such as stock price forecasting is an important part of financial research. In this regard, autoregressive (AR) and neural network (NN) models offer contrasting approaches to time series modeling. Although AR models remain widely used, NN models and their variant long short-term memory (LSTM) networks have grown in popularity. In this paper, we compare the performance of AR, NN, and LSTM models in forecasting linearly lagged time series. To test the models we carry out extensive numerical experiments based on simulated data. The results of the experiments reveal that despite the inherent advantage of AR models in modeling linearly lagged data, NN models perform just as well, if not better, than AR models. Furthermore, the NN models outperform LSTMs on the same data. We find that a simple multi-layer perceptron can achieve highly accurate out of sample forecasts. The study shows that NN models perform well even in the case of linearly lagged time series. © 2021 IEEE.
  • Item
    Batch-based power-controlled channel assignment for improved throughput in software-defined networks
    (Institute of Electrical and Electronics Engineers Inc., 2019) Salameh, Haythem Bany ; Musa, Ahmed ; Outoom, Ruba ; Halloush, Rami ; Aloqaily, Moayad ; Jararweh, Yaser
    Software-defined networking (SDN) along with transmission power control (TPC) have a great potential in enabling efficient wireless networking. Power control aims at increasing network throughput, while SDN provides cognition and intelligent capabilities to network devices. The key challenge in enabling efficient operation of such networks is how to perform efficient power-controlled MAC protocols that includes channel assignment and power allocation such that network throughput is enhanced while using the least number of channels. Traditional MAC protocols for SDNs employ an exclusive channel-occupancy between neighboring secondary users (SUs), which significantly limits network performance. In this paper, we develop a novel power-controlled spectrum access protocol for SDNs based on the interference-channel occupancy model with the objective of increasing network throughput. It allows several concurrent interference-limited transmissions to simultaneously proceed over the same channel in the same neighborhood. Unlike most of previous power-control MAC protocols that perform the channel assignment and power allocation sequentially, our protocol simultaneously makes distributed channel and power assignment decisions for multiple SU transmissions (batch-based method). Batching can be achieved by using an admission control window for SUs to exchange their collision-avoidance control information. Simulation results reveal that compared with CSMA/CA variants, our protocol greatly improve spectrum efficiency, which improves network throughput while reducing energy consumption. © 2019 IEEE.
  • Item
    Cluster aware mobility encounter dataset enlargement
    (Institute of Electrical and Electronics Engineers Inc., 2019) Haldar, Rajarshi ; Bacanli, Salih Safa ; Aloqaily, Moayad ; Mnaouer, Adel Ben ; Turgut, Damla
    The recent emerging fields in data processing and manipulation has facilitated the need for synthetic data generation. This is also valid for mobility encounter dataset generation. Synthetic data generation might be useful to run research-based simulations and also create mobility encounter models. Our approach in this paper is to generate a larger dataset by using a given dataset which includes the clusters of people. Based on the cluster information, we created a framework. Using this framework, we can generate a similar dataset that is statistically similar to the input dataset. We have compared the statistical results of our approach with the real dataset and an encounter mobility model generation technique in the literature. The results showed that the created datasets have similar statistical structure with the given dataset. © 2019 IEEE.
  • Item
    Comparative analysis of activation functions in neural networks
    (Institute of Electrical and Electronics Engineers Inc., 2021) Kamalov, Firuz ; Nazir, Amril ; Safaraliev, Murodbek ; Cherukuri, Aswani Kumar ; Zgheib, Rita
    Although the impact of activations on the accuracy of neural networks has been covered in the literature, there is little discussion about the relationship between the activations and the geometry of neural network model. In this paper, we examine the effects of various activation functions on the geometry of the model within the feature space. In particular, we investigate the relationship between the activations in the hidden and output layers, the geometry of the trained neural network model, and the model performance. We present visualizations of the trained neural network models to help researchers better understand and intuit the effects of activation functions on the models. © 2021 IEEE.
  • Item
    Comparative study of digital audio steganography techniques
    ( 2012) Djebbar, Fatiha ; Ayad, Beghdad ; Meraim, Karim Abed ; Hamam, Habib
    The rapid spread in digital data usage in many real life applications have urged new and effective ways to ensure their security. Efficient secrecy can be achieved, at least in part, by implementing steganograhy techniques. Novel and versatile audio steganographic methods have been proposed. The goal of steganographic systems is to obtain secure and robust way to conceal high rate of secret data. We focus in this paper on digital audio steganography, which has emerged as a prominent source of data hiding across novel telecommunication technologies such as covered voice-over-IP, audio conferencing, etc. The multitude of steganographic criteria has led to a great diversity in these system design techniques. In this paper, we review current digital audio steganographic techniques and we evaluate their performance based on robustness, security and hiding capacity indicators. Another contribution of this paper is the provision of a robustness-based classification of steganographic models depending on their occurrence in the embedding process. A survey of major trends of audio steganography applications is also discussed in this paper. © 2012 Djebbar et al.; licensee Springer.
  • Item
    Controlled distortion for high capacity data-in-speech spectrum steganography
    (Institute of Electrical and Electronics Engineers Inc., 2010) Djebbar, Fatiha ; Hamam, Habib ; Abed-Meraim, Karim ; Guerchi, Driss
    Methods applied to ensure privacy of digital data became essential in many real life applications. Efficient secrecy can be achieved, at least in part, by implementing steganography techniques. In this paper, we present a technique that limits the impact of high data capacity embedding on the quality of stego wideband speech. Our method uses the energy of each frequency bin component to determine the maximum number of bits that can be confined without inducing any noticeable distortion on the cover speech. To guarantee good quality of stego speech, the embedding in the selected frequency components occurs below a well defined distortion level to limit the impact of the hiding on the stego-speech. The algorithm uses multiple parameters that can be adjusted by the sender to render the steganalysis work more challenging. The objective and subjective results show that this approach is robust to noise addition and maintains a very good quality of the cover signal while achieving high hiding capacity. © 2010 IEEE.
  • Item
    A cooperative and conversational virtual agent for M-commerce applications
    (Institute of Electrical and Electronics Engineers Inc., 2009) Rekik, Siwar ; Selouani, Sid-Ahmed ; Hamam, Habib
    The aim of this paper is to present a cooperative approach to improve the Human-System spoken dialogues. The main advantage of the proposed approach is its ability to reach both of the user and the system goals more efficiently. The strategy that underlines our system is well adapted to the mobile applications since it involves effective spoken exchanges to reach the users' and system mutual goals. The proposed fram ework is built in order to use the Hidden Markov Models (HMMs) based CMU-Sphinx speech recognition engine f or mobile communications. To evaluate our approach a case-study in the M-trading field is considered. The analysis of the case study shows the efficiency ofour strategy in comparison with the usual ones. ©2009 IEEE.
  • Item
    Covariant representations of C*-dynamical systems with compact groups
    (The Theta Foundation, 2013) Kamalov, Firuz
    Let (A, G, σ) be a C*-dynamical system, where G is compact. We show that every irreducible covariant representation (π,U) of (A, G, σ) is induced from an irreducible covariant representation (π0,U0) of a subsystem (A, G0, σ) such that π0 is a factor representation. We show that if (π,U) is an irreducible covariant representation of (A, GP, σ) with ker π = P, then π is a homogenous representation. Hence, (A, G, σ) satisfies the strong-EHI property. © THETA, 2013.
  • Item
    Data and service management in densely crowded environments : challenges, opportunities, and recent developments
    (Institute of Electrical and Electronics Engineers Inc., 2019) Aloqaily, Moayad ; Ridhawi, Ismaeel Al ; Salameh, Haythem Bany ; Jararweh, Yaser
    Densely crowded environments such as stadiums and metro stations have shown shortcomings when users request data and services simultaneously. This is due to the excessive amount of requested and generated traffic from the user side. Based on the wide availability of user smart-mobile devices, and noting their technological advancements, devices are not being categorized only as data/service requesters anymore, but are readily being transformed to data/service providing network-side tools. In essence, to offload some of the workload burden from the cloud, data can be either fully or partially replicated to edge and mobile devices for faster and more efficient data access in such dense environments. Moreover, densely crowded environments provide an opportunity to deliver, in a timely manner, through node collaboration, enriched user-specific services using the replicated data and device-specific capabilities. In this article, we first highlight the challenges that arise in densely crowded environments in terms of data/service management and delivery. Then we show how data replication and service composition are considered promising solutions for data and service management in densely crowded environments. Specifically, we describe how to replicate data from the cloud to the edge, and then to mobile devices to provide faster data access for users. We also discuss how services can be composed in crowded environments using service-specific overlays. We conclude the article with most of the open research areas that remain to be investigated. © 2019 IEEE.
  • Item
    Data imbalance in classification : experimental evaluation
    (Elsevier Inc., 2020-03) Thabtah, Fadi ; Hammoud, Suhel ; Kamalov, Firuz ; Gonsalves, Amanda
    The advent of Big Data has ushered a new era of scientific breakthroughs. One of the common issues that affects raw data is class imbalance problem which refers to imbalanced distribution of values of the response variable. This issue is present in fraud detection, network intrusion detection, medical diagnostics, and a number of other fields where negatively labeled instances significantly outnumber positively labeled instances. Modern machine learning techniques struggle to deal with imbalanced data by focusing on minimizing the error rate for the majority class while ignoring the minority class. The goal of our paper is demonstrate the effects of class imbalance on classification models. Concretely, we study the impact of varying class imbalance ratios on classifier accuracy. By highlighting the precise nature of the relationship between the degree of class imbalance and the corresponding effects on classifier performance we hope to help researchers to better tackle the problem. To this end, we carry out extensive experiments using 10-fold cross validation on a large number of datasets. In particular, we determine that the relationship between the class imbalance ratio and the accuracy is convex. © 2019 Elsevier Inc.
  • Item
    Deep learning regularization in imbalanced data
    (Institute of Electrical and Electronics Engineers Inc., 2020-11-03) Kamalov, Firuz ; Leung, Ho Hon
    Deep neural networks are known to have a large number of parameters which can lead to overfitting. As a result various regularization methods designed to mitigate the model overfitting have become an indispensable part of many neural network architectures. However, it remains unclear which regularization methods are the most effective. In this paper, we examine the impact of regularization on neural network performance in the context of imbalanced data. We consider three main regularization approaches: L{1}, L{2}, and dropout regularization. Numerical experiments reveal that the L{1} regularization method can be an effective tool to prevent overfitting in neural network models for imbalanced data. Index Terms-regularization, neural networks, imbalanced data. © 2020 IEEE.
  • Item
    Deepkover - an adaptive artful intelligent assistance system for cognitively impaired people
    (Taylor and Francis Ltd., 2010) Najjar, Mehdi ; Courtemanche, Francois ; Hamam, Habib ; Mayers, Andre
    This article presents a novel modular adaptive artful intelligent assistance system for cognitively and/or memory impaired people engaged in the realisation of their activities of daily living (ADLs). The goal of this assistance system is to help disabled persons moving/evolving within a controlled environment in order to provide logistic support in achieving their ADLs. Empirical results of practical tests are presented and interpreted. Some deductions about the key features that represent originalities of the assistance system are drawn and future works are announced. © 2010 Taylor & Francis Group, LLC.
  • Item
    The dual structure of crossed product C*-algebras with finite groups
    ( 2013) Kamalov, Firuz
    We study the space of irreducible representations of a crossed product C*-algebra A⋊σ G, where G is a finite group. We construct a space Γ which consists of pairs of irreducible representations of A and irreducible projective representations of subgroups of G. We show that there is a natural action of G on Γ and that the orbit space G\Γ corresponds bijectively to the dual of A⋊σG. © 2013 Australian Mathematical Publishing Association Inc.
  • Item
    Ensemble Learning with Resampling for Imbalanced Data
    (Springer Science and Business Media Deutschland GmbH, 2021) Kamalov, Firuz ; Elnagar, Ashraf ; Leung, Ho Hon
    Imbalanced class distribution is an issue that appears in various applications. In this paper, we undertake a comprehensive study of the effects of sampling on the performance of bootstrap aggregating in the context of imbalanced data. Concretely, we carry out a comparison of sampling methods applied to single and ensemble classifiers. The experiments are conducted on simulated and real-life data using a range of sampling methods. The contributions of the paper are twofold: i) demonstrate the effectiveness of ensemble techniques based on resampled data over a single base classifier and ii) compare the effectiveness of different resampling techniques when used during the bagging stage for ensemble classifiers. The results reveal that ensemble methods overwhelmingly outperform single classifiers based on resampled data. In addition, we discover that NearMiss and random oversampling (ROS) are the optimal sampling algorithms for ensemble learning. © 2021, Springer Nature Switzerland AG.
  • Item
    Evaluation of Arabic-Based Contextualized Word Embedding Models
    (Institute of Electrical and Electronics Engineers Inc., 2021) Yagi, Sane Mo ; Mansour, Youssef ; Kamalov, Firuz ; Elnagar, Ashraf
    The distributed representation of words, as in Word2Vec, FastText, and GloVe, results in the production of a single vector for each word type regardless of the polysemy or homonymy that many words may have. Context-sensitive representation as implemented in deep learning neural networks, on the other hand, produces different vectors for the multiple senses of a word. Several contextualized word embeddings have been produced for the Arabic language (e.g., AraBERT, QARiB, AraGPT, etc.). The majority of these were tested on a few NLP tasks but there was no direct comparison between them. As a result, we do not know which of these is most efficient and for which tasks. This paper is a first step in an endeavor to establish evaluation criteria for them. It describes 24 such embeddings, then conducts exploratory intrinsic and extrinsic evaluation of them. Afterwards, it tests relational knowledge in them, covering four semantic relations: colors of fruits, capitals of countries, causation, and general information. It also evaluates the utility of these models in Named Entity Recognition and Sentiment Analysis tasks. It has been demonstrated here that AraBERTv02 and MARBERT are the best on both types of evaluation; therefore, both are recommended for fine-tuning Arabic NLP tasks. The ultimate conclusion is that it is feasible to test higher order reasoning relations in these embeddings. © 2021 IEEE