Digital Commons Network^™

Applications Of Quantitative Information Flow To Property-Revealing Encryption And Differential Privacy, Mireya Anita Jurado

FIU Electronic Theses and Dissertations

Across our digital lives, two powerful forces of data utility and data privacy push and pull against each other. In response, technologies have been designed to help navigate this tension, by revealing, or leaking, information that could be useful to an adversary in exchange for some added utility. For many of these technologies, there is uncertainty about how harmful these leaks are. How much security and privacy are we sacrificing? Under what circumstances are these technologies safe to use?

To bring clarity to these compromises between security and utility, this dissertation applies the information-theoretic framework of quantitative information flow (QIF), …

An Evolutionary Optimization Algorithm For Automated Classical Machine Learning, Leila Zahedi

FIU Electronic Theses and Dissertations

Machine learning is an evolving branch of computational algorithms that allow computers to learn from experiences, make predictions, and solve different problems without being explicitly programmed. However, building a useful machine learning model is a challenging process, requiring human expertise to perform various proper tasks and ensure that the machine learning's primary objective --determining the best and most predictive model-- is achieved. These tasks include pre-processing, feature selection, and model selection. Many machine learning models developed by experts are designed manually and by trial and error. In other words, even experts need the time and resources to create good predictive …

Anomaly Detection In Sequential Data: A Deep Learning-Based Approach, Jayesh Soni

FIU Electronic Theses and Dissertations

Anomaly Detection has been researched in various domains with several applications in intrusion detection, fraud detection, system health management, and bio-informatics. Conventional anomaly detection methods analyze each data instance independently (univariate or multivariate) and ignore the sequential characteristics of the data. Anomalies in the data can be detected by grouping the individual data instances into sequential data and hence conventional way of analyzing independent data instances cannot detect anomalies. Currently: (1) Deep learning-based algorithms are widely used for anomaly detection purposes. However, significant computational overhead time is incurred during the training process due to static constant batch size and learning …

Integrating Cultural Knowledge Into Artificially Intelligent Systems: Human Experiments And Computational Implementations, Anurag Acharya

FIU Electronic Theses and Dissertations

With the advancement of Artificial Intelligence, it seems as if every aspect of our lives is impacted by AI in one way or the other. As AI is used for everything from driving vehicles to criminal justice, it becomes crucial that it overcome any biases that might hinder its fair application. We are constantly trying to make AI be more like humans. But most AI systems so far fail to address one of the main aspects of humanity: our culture and the differences between cultures. We cannot truly consider AI to have understood human reasoning without understanding culture. So it …

Distributed Machine Learning Algorithms For Resource-Constrained Heterogeneous Internet-Of-Things Environments, Ahmed Imteaj

FIU Electronic Theses and Dissertations

With the improvement of network infrastructures and advancement of IoT technologies, now it is desirable to perform computation at the edges, rather than sharing data with a central fusion center, which is privacy-intrusive. Both conventional (centralized) and distributed machine learning (ML) algorithms fail to address underlying challenges related to users’ privacy or capturing global knowledge of the whole network. To properly handle such challenges, a recently invented distributed ML technique, called Federated Learning was invented that shows us a pathway to construct a global model without exposing any user’s private data through on-device model training utilizing edge resources. However, FL …

Intelligent Data Analytics Using Deep Learning For Data Science, Maria E. Presa Reyes

FIU Electronic Theses and Dissertations

Nowadays, data science stimulates the interest of academics and practitioners because it can assist in the extraction of significant insights from massive amounts of data. From the years 2018 through 2025, the Global Datasphere is expected to rise from 33 Zettabytes to 175 Zettabytes, according to the International Data Corporation. This dissertation proposes an intelligent data analytics framework that uses deep learning to tackle several difficulties when implementing a data science application. These difficulties include dealing with high inter-class similarity, the availability and quality of hand-labeled data, and designing a feasible approach for modeling significant correlations in features gathered from …

Communicating With Culture: How Humans And Machines Detect Narrative Elements, Wolfgang Victor H. Yarlott

FIU Electronic Theses and Dissertations

To understand how people communicate, we must understand how they leverage shared stories and all the knowledge, information, and associations contained within those stories. I examine three classes of narrative elements that convey a wealth of cultural knowledge: Propp's morphology, motifs, and discourse structure. Propp's morphology communicates how roles and actions drive a narrative forward; motifs fill those roles and actions with specific, remarkable events; discourse groups these into a coherent structure to convey a point.

My thesis has three aims: first, to demonstrate that people can reliably detect and identify all three of these narrative elements; second, to develop …

Detecting The Emotions Of Animate Beings In Narrative, Samira Zad

FIU Electronic Theses and Dissertations

Identifying emotions as expressed in text (a.k.a. text emotion recognition) has received a lot of attention over the past decade. Narratives often involve a great deal of emotional expression, and so emotion recognition on narrative text is of great interest to computational approaches to narrative understanding. The meaning and impact of narratives is strongly bound up with the emotions expressed therein. Emotions may be experienced by characters in a story (which may include the narrator), by a story-external narrator, or by the reader. There has been so far two separate streams of work relevant to this observation: (1) emotion detection, …

A Machine Learning Framework For Identifying Molecular Biomarkers From Transcriptomic Cancer Data, Md Abdullah Al Mamun

FIU Electronic Theses and Dissertations

Cancer is a complex molecular process due to abnormal changes in the genome, such as mutation and copy number variation, and epigenetic aberrations such as dysregulations of long non-coding RNA (lncRNA). These abnormal changes are reflected in transcriptome by turning oncogenes on and tumor suppressor genes off, which are considered cancer biomarkers.

However, transcriptomic data is high dimensional, and finding the best subset of genes (features) related to causing cancer is computationally challenging and expensive. Thus, developing a feature selection framework to discover molecular biomarkers for cancer is critical.

Traditional approaches for biomarker discovery calculate the fold change for each …

Deep Learning For Multiclass Classification, Predictive Modeling And Segmentation Of Disease Prone Regions In Alzheimer’S Disease, Maryamossadat Aghili

FIU Electronic Theses and Dissertations

One of the challenges facing accurate diagnosis and prognosis of Alzheimer’s Disease (AD) is identifying the subtle changes that define the early onset of the disease. This dissertation investigates three of the main challenges confronted when such subtle changes are to be identified in the most meaningful way. These are (1) the missing data challenge, (2) longitudinal modeling of disease progression, and (3) the segmentation and volumetric calculation of disease-prone brain areas in medical images. The scarcity of sufficient data compounded by the missing data challenge in many longitudinal samples exacerbates the problem as we seek statistical meaningfulness in multiclass …

An Exploration Of Controlling The Content Learned By Deep Neural Networks, Liqun Yang

FIU Electronic Theses and Dissertations

With the great success of the Deep Neural Network (DNN), how to get a trustworthy model attracts more and more attention. Generally, people intend to provide the raw data to the DNN directly in training. However, the entire training process is in a black box, in which the knowledge learned by the DNN is out of control. There are many risks inside. The most common one is overfitting. With the deepening of research on neural networks, additional and probably greater risks were discovered recently. The related research shows that unknown clues can hide in the training data because of the …

A Study Of Sparse Representation Of Boolean Functions, Yekun Xu

FIU Electronic Theses and Dissertations

Boolean function is one of the most fundamental computation models in theoretical computer science. The two most common representations of Boolean functions are Fourier transform and real polynomial form. Applying analytic tools under these representations to the study Boolean functions has led to fruitful research in many areas such as complexity theory, learning theory, inapproximability, pseudorandomness, metric embedding, property testing, threshold phenomena, social choice, etc. In this thesis, we focus on \emph{sparse representations} of Boolean function in both Fourier transform and polynomial form, and obtain the following new results. A classical result of Rothschild and van Lint asserts that if …

Interpretability Of Ai In Computer Systems And Public Policy, Farzana Beente Yusuf

FIU Electronic Theses and Dissertations

Advances in Artificial Intelligence (AI) have led to spectacular innovations and sophisticated systems for tasks that were thought to be capable only by humans. Examples include playing chess and Go, face and voice recognition, driving vehicles, and more. In recent years, the impact of AI has moved beyond offering mere predictive models into building interpretable models that appeal to human logic and intuition because they ensure transparency and simplicity and can be used to make meaningful decisions in real-world applications. A second trend in AI is characterized by important advancements in the realm of causal reasoning. Identifying causal relationships is …

Inducing Stereotypical Character Roles From Plot Structure, Labiba Jahan

FIU Electronic Theses and Dissertations

If we are to understand stories, we must understand characters: characters are central to every narrative and drive the action forward. Critically, many stories (especially cultural ones) employ stereotypical character roles in their stories for different purposes, including efficient communication among bundles of default characteristics and associations, ease understanding of those characters' role in the overall narrative, and many more. These roles include ideas such as hero, villain, or victim, as well as culturally-specific roles such as, for example, the donor (in Russian tales) or the trickster (in Native American tales). My thesis aims to learn these roles automatically, inducing …

Technical Interviews: Another Barrier To Broadening Participation In Computing?, Stephanie Jill Lunn

FIU Electronic Theses and Dissertations

What does it take to obtain a computing position in the industry? Although anecdotal reports state that ``hiring is broken,'' empirical evidence is necessary to identify the flaws in the existing system. The goal of this dissertation was to understand what expectations companies have for job seekers in computing, and to explore students' experiences with technical interviews and their pathways to job attainment. In particular, this work considered how hiring practices may impact populations already underrepresented in computing such as women, Black/African American students, and Hispanic/Latinx students. It also sought to understand how minoritized populations leverage their own inherent capital …

Development Of A Real-Time Single-Lead Single-Beat Frequency-Independent Myocardial Infarction Detector, Harold Martin

FIU Electronic Theses and Dissertations

The central aim of this research is the development and deployment of a novel multilayer machine learning design with unique application for the diagnosis of myocardial infarctions (MIs) from individual heartbeats of single-lead electrocardiograms (EKGs) irrespective of their sampling frequencies over a given range. To the best of our knowledge, this design is the first to attempt inter-patient myocardial infarction detection from individual heartbeats of single-lead (lead II) electrocardiograms that achieves high accuracy and near real-time diagnosis. The processing time of 300 milliseconds to a diagnosis is just at the time range in between extremely fast heartbeats of around 300 …

Evaluation Of Parametric And Nonparametric Statistical Models In Wrong-Way Driving Crash Severity Prediction, Sajidur Rahman Nafis

FIU Electronic Theses and Dissertations

Wrong-way driving (WWD) crashes result in more fatalities per crash, involve more vehicles, and cause extended road closures compared to other types of crashes. Although crashes involving wrong-way drivers are relatively few, they often lead to fatalities and serious injuries. Researchers have been using parametric statistical models to identify factors that affect WWD crash severity. However, these parametric models are generally based on several assumptions, and the results could generate numerous errors and become questionable when these assumptions are violated. On the other hand, nonparametric methods such as data mining or machine learning techniques do not use a predetermined functional …

Correlating Water Quality And Profile Data In The Florida Keys Using Machine Learning Methods, Alejandro M. Torres Castellanos

FIU Electronic Theses and Dissertations

Water quality is a very active subject of research in the water science field, where its importance includes maintaining the environment, managing wastewater, and securing fresh water. However, the increase of human development has led to problems that are affecting the ecosystem. Motivated by these problems, this research aims to find a solution for understanding the coastal water of the Florida Keys. The research used machine learning methods to find a correlation between water quality dataset and profile measurements dataset. To achieve this objective, the research first went through cleaning, rescuing, and structuring a readable dataset of the profile measurements …

An Angle-Based Stochastic Gradient Descent Method For Machine Learning: Principle And Application, Chongya Song

FIU Electronic Theses and Dissertations

In deep learning, optimization algorithms are employed to expedite the resolution to accurate models through the calibrations of the current gradient and the associated learning rate. A major shortcoming of these existing methods is the manner in which the calibration terms are computed, only utilizing the previous gradients during their computations. Because the gradient is a time-sensitive variable computed at a specific moment in time, it is possible that older gradients can introduce significant deviation into the calibration terms. Although most algorithms alleviate this situation by combining the exponential moving average of the previous gradients, we found that this method …

Digital And Mixed Domain Hardware Reduction Algorithms And Implementations For Massive Mimo, Najath A. Mohomed

FIU Electronic Theses and Dissertations

Emerging 5G and 6G based wireless communications systems largely rely on multiple-input-multiple-output (MIMO) systems to reduce inherently extensive path losses, facilitate high data rates, and high spatial diversity. Massive MIMO systems used in mmWave and sub-THz applications consists of hundreds perhaps thousands of antenna elements at base stations. Digital beamforming techniques provide the highest flexibility and better degrees of freedom for phased antenna arrays as compared to its analog and hybrid alternatives but has the highest hardware complexity.

Conventional digital beamformers at the receiver require a dedicated analog to digital converter (ADC) for every antenna element, leading to ADCs for …

Defense By Deception Against Stealthy Attacks In Power Grids, Md Hasan Shahriar

FIU Electronic Theses and Dissertations

Cyber-physical Systems (CPSs) and the Internet of Things (IoT) are converging towards a hybrid platform that is becoming ubiquitous in all modern infrastructures. The integration of the complex and heterogeneous systems creates enormous space for the adversaries to get into the network and inject cleverly crafted false data into measurements, misleading the control center to make erroneous decisions. Besides, the attacker can make a critical part of the system unavailable by compromising the sensor data availability. To obfuscate and mislead the attackers, we propose DDAF, a deceptive data acquisition framework for CPSs' hierarchical communication network. Each switch in the hierarchical …

Temporal And Causal Inference With Longitudinal Multi-Omics Microbiome Data, Daniel Ruiz-Perez

FIU Electronic Theses and Dissertations

Microbiomes are communities of microbes inhabiting an environmental niche. Thanks to next generation sequencing technologies, it is now possible to study microbial communities, their impact on the host environment, and their role in specific diseases and health. Technology has also triggered the increased generation of multi-omics microbiome data, including metatranscriptomics (quantitative survey of the complete metatranscriptome of the microbial community), metabolomics (quantitative profile of the entire set of metabolites present in the microbiome's environmental niche), and host transcriptomics (gene expression profile of the host). Consequently, another major challenge in microbiome data analysis is the integration of multi-omics data sets and …

Causality In Microbiomes, Md Musfiqur Rahman Sazal

FIU Electronic Theses and Dissertations

No abstract provided.

Automatic Learning Of Document Section Structure For Ontology-Based Semantic Search, Deya Banisakher

FIU Electronic Theses and Dissertations

Modeling natural human behavior in understanding written language is crucial for developing true artificial intelligence. For people, words convey certain semantic concepts. While documents represent an abstract concept---they are collections of text organized in some logical structure, that is, sentences, paragraphs, sections, and so on. Similar to words, these document structures, are used to convey a logical flow of semantic concepts. Machines however, only view words as spans of characters and documents as mere collections of free-text, missing any underlying meanings behind words and the logical structure of those documents.

Automatic semantic concept detection is the process by which the …

A Comprehensive Security Framework For Securing Sensors In Smart Devices And Applications, Amit Kumar Sikder

FIU Electronic Theses and Dissertations

This doctoral dissertation introduces novel security frameworks to detect sensor-based threats on smart devices and applications in smart settings such as smart home, smart office, etc. First, we present a formal taxonomy and in-depth impact analysis of existing sensor-based threats to smart devices and applications based on attack characteristics, targeted components, and capabilities. Then, we design a novel context-aware intrusion detection system, 6thSense, to detect sensor-based threats in standalone smart devices (e.g., smartphone, smart watch, etc.). 6thSense considers user activity-sensor co-dependence in standalone smart devices to learn the ongoing user activity contexts and builds a context-aware model to distinguish malicious …

Demystifying Search Rank Fraud, Nestor G. Hernandez

FIU Electronic Theses and Dissertations

Search rank fraud, i.e., the posting of large numbers of fake activities for products hosted in commercial peer-opinion services such as those provided by Google, Apple, Amazon, seeks to give the illusion of grassroots engagement, and boost financial gains, promote malware and even assist censorship efforts. Search rank fraud continues to be a significant problem, after years of investment from service providers and the academic community. In this thesis we envision that knowledge of the authentic capabilities, behaviors and strategies employed by empirically validated workers, will enable us to develop solutions that efficiently manage and contain search rank fraud, by …

Mitigating Stealthy Link Flooding Ddos Attacks Using Sdn-Based Moving Target Defense, Abdullah Aydeger

FIU Electronic Theses and Dissertations

With the increasing diversity and complication of Distributed Denial-of-Service (DDoS) attacks, it has become extremely challenging to design a fully protected network. For instance, recently, a new type of attack called Stealthy Link Flooding Attack (SLFA) has been shown to cause critical network disconnection problems, where the attacker targets the communication links in the surrounding area of a server. The existing defense mechanisms for this type of attack are based on the detection of some unusual traffic patterns; however, this might be too late as some severe damage might already be done. These mechanisms also do not consider countermeasures during …

Development Of Gaussian Learning Algorithms For Early Detection Of Alzheimer's Disease, Chen Fang

FIU Electronic Theses and Dissertations

Alzheimer’s disease (AD) is the most common form of dementia affecting 10% of the population over the age of 65 and the growing costs in managing AD are estimated to be $259 billion, according to data reported in the 2017 by the Alzheimer's Association. Moreover, with cognitive decline, daily life of the affected persons and their families are severely impacted. Taking advantage of the diagnosis of AD and its prodromal stage of mild cognitive impairment (MCI), an early treatment may help patients preserve the quality of life and slow the progression of the disease, even though the underlying disease cannot …

Reputation-Aware Trajectory-Based Data Mining In The Internet Of Things (Iot), Samia Tasnim

FIU Electronic Theses and Dissertations

Internet of Things (IoT) is a critically important technology for the acquisition of spatiotemporally dense data in diverse applications, ranging from environmental monitoring to surveillance systems. Such data helps us improve our transportation systems, monitor our air quality and the spread of diseases, respond to natural disasters, and a bevy of other applications. However, IoT sensor data is error-prone due to a number of reasons: sensors may be deployed in hazardous environments, may deplete their energy resources, have mechanical faults, or maybe become the targets of malicious attacks by adversaries. While previous research has attempted to improve the quality of …

Trajectory Privacy Preservation And Lightweight Blockchain Techniques For Mobility-Centric Iot, Abdur Bin Shahid

FIU Electronic Theses and Dissertations

Various research efforts have been undertaken to solve the problem of trajectory privacy preservation in the Internet of Things (IoT) of resource-constrained mobile devices. Most attempts at resolving the problem have focused on the centralized model of IoT, which either impose high delay or fail against a privacy-invading attack with long-term trajectory observation. These proposed solutions also fail to guarantee location privacy for trajectories with both geo-tagged and non-geo-tagged data, since they are designed for geo-tagged trajectories only. While a few blockchain-based techniques have been suggested for preserving trajectory privacy in decentralized model of IoT, they require large storage capacity …