Physical Sciences and Mathematics Commons^™

Global Cyber Attack Forecast Using Ai Techniques, Nusrat Kabir Samia

Electronic Thesis and Dissertation Repository

The advancement of internet technology and growing involvement in the cyber world have made us prone to cyber-attacks inducing severe damage to individuals and organizations, including financial loss, identity theft, and reputational damage. The rapid emergence and evolution of new networks and new opportunities for businesses and technologies are increasing threats to security vulnerabilities. Hence cyber-crime analysis is one of the wide range applications of Data Mining that can be eventually used to predict and detect crime. However, there are several constraints while analyzing cyber-attacks, which are yet to be resolved for more accurate cyber security inspection.

Although there are …

Predicting Network Failures With Ai Techniques, Chandrika Saha

Electronic Thesis and Dissertation Repository

Network failure is the unintentional interruption of internet services, resulting in widespread client frustration. It is especially true for time-sensitive services in the healthcare industry, smart grid control, and mobility control, among others. In addition, the COVID-19 pandemic has compelled many businesses to operate remotely, making uninterrupted internet access essential. Moreover, Internet Service Providers (ISPs) lose millions of dollars annually due to network failure, which has a negative impact on their businesses. Currently, redundant network equipment is used as a restoration technique to resolve this issue of network failure. This technique requires a strategy for failure identification and prediction to …

Reducing Negative Transfer Of Random Data In Source-Free Unsupervised Domain Adaptation, Anthony Wong

Electronic Thesis and Dissertation Repository

In domain adaptation, a model trained on one dataset (source domain) is applied to a different but related dataset (target domain). The most cutting-edge method is unsupervised source-free domain adaptation (SFDA), in which source data, source labels, and target labels are not available during adaptation. This thesis explores a realistic scenario where the target dataset includes some images that are unrelated to the adaptation process. This scenario can occur from errors in data collection or processing. We provide experiments and analysis to show that current state-of-the-art (SOTA) SFDA methods suffer significant performance drops under a specific domain adaptation setup when …

Attention-Based Multi-Source-Free Domain Adaptation For Eeg Emotion Recognition, Amir Hesam Salimnia

Electronic Thesis and Dissertation Repository

Electroencephalography (EEG) based emotion recognition in affective brain-computer interfaces has advanced significantly in recent years. Unsupervised domain adaptation (UDA) methods have been successfully used to mitigate the need for large amounts of training data, which is required due to the inter-subject variability of EEG signals. Typical UDA solutions require access to raw source data to leverage the knowledge learned from the labelled source domains (previous subjects) across the target domain (a new subject), raising privacy concerns. To tackle this issue, we propose Attention-based Multi-Source-Free Domain Adaptation (AMFDA) for EEG emotion recognition. AMFDA attempts to transfer knowledge of source models to …

Algorithmic Improvements In Deep Reinforcement Learning, Norman L. Tasfi

Electronic Thesis and Dissertation Repository

Reinforcement Learning (RL) has seen exponential performance improvements over the past decade, achieving super-human performance across many domains. Deep Reinforcement Learning (DRL), the combination of RL methods with deep neural networks (DNN) as function approximators, has unlocked much of this progress. The path to generalized artificial intelligence (GAI) will depend on deep learning (DL) and RL. However, much work is required before the technology reaches anything resembling GAI. Therefore, this thesis focuses on a subset of areas within RL that require additional research to advance the field, specifically: sample efficiency, planning, and task transfer. The first area, sample efficiency, refers …

Improving Deep Entity Resolution By Constraints, Soudeh Nilforoushan

Electronic Thesis and Dissertation Repository

Entity resolutions the problem of finding duplicate data in a dataset and resolving possible differences and inconsistencies. ER is a long-standing data management and information retrieval problem and a core data integration and cleaning task. There are diverse solutions for ER that apply rule-based techniques, pairwise binary classification, clustering, and probabilistic inference, among other techniques. Deep learning (DL) has been extensively used for ER and has shown competitive performance compared to conventional ER solutions. The state-of-the-art (SOTA) ER solutions using DL are based on pairwise comparison and binary classification. They transform pairs of records into a latent space that can …

Machine Learning With Big Data For Electrical Load Forecasting, Alexandra L'Heureux

Electronic Thesis and Dissertation Repository

Today, the amount of data collected is exploding at an unprecedented rate due to developments in Web technologies, social media, mobile and sensing devices and the internet of things (IoT). Data is gathered in every aspect of our lives: from financial information to smart home devices and everything in between. The driving force behind these extensive data collections is the promise of increased knowledge. Therefore, the potential of Big Data relies on our ability to extract value from these massive data sets. Machine learning is central to this quest because of its ability to learn from data and provide data-driven …

Deep Learning For Load Forecasting With Smart Meter Data: Online And Federated Learning, Mohammad Navid Fekri

Electronic Thesis and Dissertation Repository

Electricity load forecasting has been attracting increasing attention because of its importance for energy management, infrastructure planning, and budgeting. In recent years, the proliferation of smart meters has created new opportunities for forecasting on the building and even individual household levels. Machine learning (ML) has achieved great successes in this domain; however, conventional ML techniques require data transfer to a centralized location for model training, therefore, increasing network traffic and exposing data to privacy and security risks. Also, traditional approaches employ offline learning, which means that they are only trained once and miss out on the possibility to learn from …

The Role Of Transient Vibration Of The Skull On Concussion, Rodrigo Dalvit Carvalho Da Silva

Electronic Thesis and Dissertation Repository

Concussion is a traumatic brain injury usually caused by a direct or indirect blow to the head that affects brain function. The maximum mechanical impedance of the brain tissue occurs at 450±50 Hz and may be affected by the skull resonant frequencies. After an impact to the head, vibration resonance of the skull damages the underlying cortex. The skull deforms and vibrates, like a bell for 3 to 5 milliseconds, bruising the cortex. Furthermore, the deceleration forces the frontal and temporal cortex against the skull, eliminating a layer of cerebrospinal fluid. When the skull vibrates, the force spreads directly to …

Evaluating Cranial Nonmetric Traits In Mummies From Pachacamac, Peru: The Utility Of Semi-Automated Image Segmentation In Paleoradiology, Cameron J. Beason

Electronic Thesis and Dissertation Repository

Anthropologists employ biodistance analysis to understand past population interactions and relatedness. The objectives of this thesis are twofold: to determine whether a sample of five mummies from the pilgrimage centre, Pachacamac, on the Central Coast of Peru comprised local or non-local individuals through an analysis of cranial nonmetric traits using comparative samples from the North and Central Coasts of Peru and Chile; and to test the utility of machine-learning-based image segmentation in the image analysis software, Dragonfly, to automatically segment CT scans of the mummies such that the cranial nonmetric traits are visible. Results show that while fully automated segmentation …

A Generative-Discriminative Approach To Human Brain Mapping, Deepanshu Wadhwa

Electronic Thesis and Dissertation Repository

During everyday behaviours, the brain shows complex spatial patterns of activity. These activity maps are very replicable within an individual, but vary significantly across individuals, even though they are evoked by the same behaviour. It is unknown how differences in these spatial patterns relate to differences in behavior or function. More fundamentally, the structural, developmental, and genetic factors that determine the spatial organisation of these brain maps in each individual are unclear. Here we propose a new quantitative approach for uncovering the basic principles by which functional brain maps are organized. We propose to take an generative-discriminative approach to human …

Generating Effective Sentence Representations: Deep Learning And Reinforcement Learning Approaches, Mahtab Ahmed

Electronic Thesis and Dissertation Repository

Natural language processing (NLP) is one of the most important technologies of the information age. Understanding complex language utterances is also a crucial part of artificial intelligence. Many Natural Language applications are powered by machine learning models performing a large variety of underlying tasks. Recently, deep learning approaches have obtained very high performance across many NLP tasks. In order to achieve this high level of performance, it is crucial for computers to have an appropriate representation of sentences. The tasks addressed in the thesis are best approached having shallow semantic representations. These representations are vectors that are then embedded in …

Protein Interaction Sites Prediction Using Deep Learning, Sourajit Basak

Electronic Thesis and Dissertation Repository

The accurate prediction of protein-protein interaction (PPI) binding sites is a fundamental problem in bioinformatics, since most of the time proteins perform their functions by interacting with some other proteins. Experimental methods are slow, expensive and not very accurate, hence the need for efficient computational methods.

In this thesis, we perform a study aiming to improve the performance of the currently best program for binding site prediction, DELPHI. We have employed some of the currently best techniques from machine learning, including attention and various embedding techniques, such as BERT and ELMo. This is the first time such tools are being …

A Deep Topical N-Gram Model And Topic Discovery On Covid-19 News And Research Manuscripts, Yuan Du

Electronic Thesis and Dissertation Repository

Topic modeling with the latent semantic analysis (LSA), the latent Dirichlet allocation (LDA) and the biterm topic model (BTM) has been successfully implemented and used in many areas, including movie reviews, recommender systems, and text summarization, etc. However, these models may become computationally intensive if tested on a humongous corpus. Considering the wide acceptance of machine learning based on deep neural networks, this research proposes two deep neural network (NN) variants, 2-layer NN and 3-layer NN of the LDA modeling techniques. The primary goal is to deal with problems with a large corpus using manageable computational resources.

This thesis analyze …

Edge-Cloud Iot Data Analytics: Intelligence At The Edge With Deep Learning, Ananda Mohon M. Ghosh

Electronic Thesis and Dissertation Repository

Rapid growth in numbers of connected devices, including sensors, mobile, wearable, and other Internet of Things (IoT) devices, is creating an explosion of data that are moving across the network. To carry out machine learning (ML), IoT data are typically transferred to the cloud or another centralized system for storage and processing; however, this causes latencies and increases network traffic. Edge computing has the potential to remedy those issues by moving computation closer to the network edge and data sources. On the other hand, edge computing is limited in terms of computational power and thus is not well suited for …

Deep Learning On Smart Meter Data: Non-Intrusive Load Monitoring And Stealthy Black-Box Attacks, Junfei Wang

Electronic Thesis and Dissertation Repository

Climate change and environmental concerns are instigating widespread changes in modern electricity sectors due to energy policy initiatives and advances in sustainable technologies. To raise awareness of sustainable energy usage and capitalize on advanced metering infrastructure (AMI), a novel deep learning non-intrusive load monitoring (NILM) model is proposed to disaggregate smart meter readings and identify the operation of individual appliances. This model can be used by Electric power utility (EPU) companies and third party entities, and then utilized to perform active or passive consumer power demand management. Although machine learning (ML) algorithms are powerful, these remain vulnerable to adversarial attacks. …

A New Method To Solve Same-Different Problems With Few-Shot Learning, Yuanyuan Han

Electronic Thesis and Dissertation Repository

Visual learning of highly abstract concepts is often simple for humans but very challenging for machines. Same-different (SD) problems are a visual reasoning task with highly abstract concepts. Previous work has shown that SD problems are difficult to solve with standard deep learning algorithms, especially in the few-shot case, despite the ability of such algorithms to learn abstract features. In this thesis, we propose a new method to solve SD problems with few training samples, in which same-different visual concepts can be recognized by examining similarities between Regions of Interest by using a same-different twins network. Our method achieves state-of-the-art …

Improving Neural Sequence Labelling Using Additional Linguistic Information, Muhammad Rifayat Samee

Electronic Thesis and Dissertation Repository

Sequence Labelling is the task of mapping sequential data from one domain to another domain. As we can interpret language as a sequence of words, sequence labelling is very common in the field of Natural Language Processing (NLP). In NLP, some fundamental sequence labelling tasks are Parts-of-Speech Tagging, Named Entity Recognition, Chunking, etc. Moreover, many NLP tasks can be modeled as sequence labelling or sequence to sequence labelling such as machine translation, information retrieval and question answering. An extensive amount of research has already been performed on sequence labelling. Most of the current high performing models are neural network models. …

Improving Deep Learning Image Recognition Performance Using Region Of Interest Localization Networks, Abdulwahab Kabani

Electronic Thesis and Dissertation Repository

Deep Learning has been gaining momentum and achieving the state-of-the-art results on many visual recognition problems. The roots of this field can be traced back to the 1940s of the 20th century. The field has recently started delivering some interesting results on many image understanding problems. This is mainly due to the availability of powerful hardware that can accelerate the training process. In addition, the growth of the Internet and imaging devices such as mobile phones and cameras has contributed to the increase in the amount of data that can be used to train neural networks. All of these factors …

Deep Learning Via Stacked Sparse Autoencoders For Automated Voxel-Wise Brain Parcellation Based On Functional Connectivity, Céline Gravelines

Electronic Thesis and Dissertation Repository

Functional brain parcellation – the delineation of brain regions based on functional connectivity – is an active research area lacking an ideal subject-specific solution independent of anatomical composition, manual feature engineering, or heavily labelled examples. Deep learning is a cutting-edge area of machine learning on the forefront of current artificial intelligence developments. Specifically, autoencoders are artificial neural networks which can be stacked to form hierarchical sparse deep models from which high-level features are compressed, organized, and extracted, without labelled training data, allowing for unsupervised learning. This thesis presents a novel application of stacked sparse autoencoders to the problem of parcellating …