Physical Sciences and Mathematics Commons^™

Attribution Robustness Of Neural Networks, Sunanda Gamage

Electronic Thesis and Dissertation Repository

While deep neural networks have demonstrated excellent learning capabilities, explainability of model predictions remains a challenge due to their black box nature. Attributions or feature significance methods are tools for explaining model predictions, facilitating model debugging, human-machine collaborative decision making, and establishing trust and compliance in critical applications. Recent work has shown that attributions of neural networks can be distorted by imperceptible adversarial input perturbations, which makes attributions unreliable as an explainability method. This thesis addresses the research problem of attribution robustness of neural networks and introduces novel techniques that enable robust training at scale.

Firstly, a novel generic framework …

Ai Applications On Planetary Rovers, Alexis David Pascual

Electronic Thesis and Dissertation Repository

The rise in the number of robotic missions to space is paving the way for the use of artificial intelligence and machine learning in the autonomy and augmentation of rover operations. For one, more rovers mean more images, and more images mean more data bandwidth required for downlinking as well as more mental bandwidth for analyzing the images. On the other hand, light-weight, low-powered microrover platforms are being developed to accommodate the drive for planetary exploration. As a result of the mass and power constraints, these microrover platforms will not carry typical navigational instruments like a stereocamera or a laser …

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 …

Leveraging Machine Learning Techniques Towards Intelligent Networking Automation, Cesar A. Gomez

Electronic Thesis and Dissertation Repository

In this thesis, we address some of the challenges that the Intelligent Networking Automation (INA) paradigm poses. Our goal is to design schemes leveraging Machine Learning (ML) techniques to cope with situations that involve hard decision-making actions. The proposed solutions are data-driven and consist of an agent that operates at network elements such as routers, switches, or network servers. The data are gathered from realistic scenarios, either actual network deployments or emulated environments. To evaluate the enhancements that the designed schemes provide, we compare our solutions to non-intelligent ones. Additionally, we assess the trade-off between the obtained improvements and the …

Optimized Machine Learning Models Towards Intelligent Systems, Mohammadnoor Ahmad Mohammad Injadat

Electronic Thesis and Dissertation Repository

The rapid growth of the Internet and related technologies has led to the collection of large amounts of data by individuals, organizations, and society in general [1]. However, this often leads to information overload which occurs when the amount of input (e.g. data) a human is trying to process exceeds their cognitive capacities [2]. Machine learning (ML) has been proposed as one potential methodology capable of extracting useful information from large sets of data [1]. This thesis focuses on two applications. The first is education, namely e-Learning environments. Within this field, this thesis proposes different optimized ML ensemble models to …

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 …

A Visual Analytics System For Making Sense Of Real-Time Twitter Streams, Amir Haghighatimaleki

Electronic Thesis and Dissertation Repository

Through social media platforms, massive amounts of data are being produced. Twitter, as one such platform, enables users to post “tweets” on an unprecedented scale. Once analyzed by machine learning (ML) techniques and in aggregate, Twitter data can be an invaluable resource for gaining insight. However, when applied to real-time data streams, due to covariate shifts in the data (i.e., changes in the distributions of the inputs of ML algorithms), existing ML approaches result in different types of biases and provide uncertain outputs. This thesis describes a visual analytics system (i.e., a tool that combines data visualization, human-data interaction, and …

Secured Data Masking Framework And Technique For Preserving Privacy In A Business Intelligence Analytics Platform, Osama Ali

Electronic Thesis and Dissertation Repository

The main concept behind business intelligence (BI) is how to use integrated data across different business systems within an enterprise to make strategic decisions. It is difficult to map internal and external BI’s users to subsets of the enterprise’s data warehouse (DW), resulting that protecting the privacy of this data while maintaining its utility is a challenging task. Today, such DW systems constitute one of the most serious privacy breach threats that an enterprise might face when many internal users of different security levels have access to BI components. This thesis proposes a data masking framework (iMaskU: Identify, Map, Apply, …

Machs: Mitigating The Achilles Heel Of The Cloud Through High Availability And Performance-Aware Solutions, Manar Jammal

Electronic Thesis and Dissertation Repository

Cloud computing is continuously growing as a business model for hosting information and communication technology applications. However, many concerns arise regarding the quality of service (QoS) offered by the cloud. One major challenge is the high availability (HA) of cloud-based applications. The key to achieving availability requirements is to develop an approach that is immune to cloud failures while minimizing the service level agreement (SLA) violations. To this end, this thesis addresses the HA of cloud-based applications from different perspectives. First, the thesis proposes a component’s HA-ware scheduler (CHASE) to manage the deployments of carrier-grade cloud applications while maximizing their …

Metafork: A Compilation Framework For Concurrency Models Targeting Hardware Accelerators, Xiaohui Chen

Electronic Thesis and Dissertation Repository

Parallel programming is gaining ground in various domains due to the tremendous computational power that it brings; however, it also requires a substantial code crafting effort to achieve performance improvement. Unfortunately, in most cases, performance tuning has to be accomplished manually by programmers. We argue that automated tuning is necessary due to the combination of the following factors. First, code optimization is machine-dependent. That is, optimization preferred on one machine may be not suitable for another machine. Second, as the possible optimization search space increases, manually finding an optimized configuration is hard. Therefore, developing new compiler techniques for optimizing applications …

Analysis Of 3d Cone-Beam Ct Image Reconstruction Performance On A Fpga, Devin Held

Electronic Thesis and Dissertation Repository

Efficient and accurate tomographic image reconstruction has been an intensive topic of research due to the increasing everyday usage in areas such as radiology, biology, and materials science. Computed tomography (CT) scans are used to analyze internal structures through capture of x-ray images. Cone-beam CT scans project a cone-shaped x-ray to capture 2D image data from a single focal point, rotating around the object. CT scans are prone to multiple artifacts, including motion blur, streaks, and pixel irregularities, therefore must be run through image reconstruction software to reduce visual artifacts. The most common algorithm used is the Feldkamp, Davis, and …

Climbing Up Cloud Nine: Performance Enhancement Techniques For Cloud Computing Environments, Mohamed Abusharkh

Electronic Thesis and Dissertation Repository

With the transformation of cloud computing technologies from an attractive trend to a business reality, the need is more pressing than ever for efficient cloud service management tools and techniques. As cloud technologies continue to mature, the service model, resource allocation methodologies, energy efficiency models and general service management schemes are not yet saturated. The burden of making this all tick perfectly falls on cloud providers. Surely, economy of scale revenues and leveraging existing infrastructure and giant workforce are there as positives, but it is far from straightforward operation from that point. Performance and service delivery will still depend on …

Automated Image Interpretation For Science Autonomy In Robotic Planetary Exploration, Raymond Francis

Electronic Thesis and Dissertation Repository

Advances in the capabilities of robotic planetary exploration missions have increased the wealth of scientific data they produce, presenting challenges for mission science and operations imposed by the limits of interplanetary radio communications. These data budget pressures can be relieved by increased robotic autonomy, both for onboard operations tasks and for decision- making in response to science data.

This thesis presents new techniques in automated image interpretation for natural scenes of relevance to planetary science and exploration, and elaborates autonomy scenarios under which they could be used to extend the reach and performance of exploration missions on planetary surfaces.

Two …

On The Interoperability Of Programming Languages Based On The Fork-Join Parallelism Model, Sushek Shekar

Electronic Thesis and Dissertation Repository

This thesis describes the implementation of MetaFork, a meta-language for concurrency platforms targeting multicore architectures. First of all, MetaFork is a multithreaded language based on the fork-join model of concurrency: it allows the programmer to express parallel algorithms assuming that tasks are dynamically scheduled at run-time. While MetaFork makes no assumption about the run-time system, it formally defines the serial C-elision of a MetaFork program. In addition, MetaFork is a suite of source-to-source compilers permitting the automatic translation of multithreaded programs between programming languages based on the fork-join model. Currently, this compilation framework supports the OpenMP and CilkPlus concurrency platforms. …

Disaster Data Management In Cloud Environments, Katarina Grolinger

Electronic Thesis and Dissertation Repository

Facilitating decision-making in a vital discipline such as disaster management requires information gathering, sharing, and integration on a global scale and across governments, industries, communities, and academia. A large quantity of immensely heterogeneous disaster-related data is available; however, current data management solutions offer few or no integration capabilities and limited potential for collaboration. Moreover, recent advances in cloud computing, Big Data, and NoSQL have opened the door for new solutions in disaster data management.

In this thesis, a Knowledge as a Service (KaaS) framework is proposed for disaster cloud data management (Disaster-CDM) with the objectives of 1) facilitating information gathering …

Application Of Computer Algebra In List Decoding, Muhammad Foizul Islam Chowdhury

Electronic Thesis and Dissertation Repository

The amount of data that we use in everyday life (social media, stock analysis, satellite communication etc.) are increasing day by day. As a result, the amount of data needs to be traverse through electronic media as well as to store are rapidly growing and there exist several environmental effects that can damage these important data during travelling or while in storage devices. To recover correct information from noisy data, we do use error correcting codes. The most challenging work in this area is to have a decoding algorithm that can decode the code quite fast, in addition with the …

Collaborative Policy-Based Autonomic Management In Iaas Clouds, Omid Mola

Electronic Thesis and Dissertation Repository

With the increasing number of "machines" (either virtual or physical) in a computing environment, it is becoming harder to monitor and manage these resources. Relying on human administrators, even with tools, is expensive and the growing complexity makes management even harder. The alternative is to look for automated approaches that can monitor and manage computing resources in real time with no human intervention. One of the approaches to this problem is policy-based autonomic management. However, in large systems having one single autonomic manager to manage everything is almost impossible. Therefore, multiple autonomic managers will be needed and these will need …

Reliability Models Applied To Smartphone Applications, Sonia Meskini

Electronic Thesis and Dissertation Repository

Smartphones have become the most used electronic devices. They carry out most of the functionalities of desktops, offering various useful applications that suit the user’s needs. Therefore, instead of the operator, the user has been the main controller of the device and its applications, therefore its reliability has become an emergent requirement. As a first step, based on collected smartphone applications failure data, we investigated and evaluated the efficacy of Software Reliability Growth Models (SRGMs) when applied to these smartphone data in order to check whether they achieve the same accuracy as in the desktop/laptop area. None of the selected …