Physical Sciences and Mathematics Commons^™

Generating Energy Data For Machine Learning With Recurrent Generative Adversarial Networks, Mohammad Navid Fekri, Ananda M. Ghosh, Katarina Grolinger

Electrical and Computer Engineering Publications

The smart grid employs computing and communication technologies to embed intelligence into the power grid and, consequently, make the grid more efficient. Machine learning (ML) has been applied for tasks that are important for smart grid operation including energy consumption and generation forecasting, anomaly detection, and state estimation. These ML solutions commonly require sufficient historical data; however, this data is often not readily available because of reasons such as data collection costs and concerns regarding security and privacy. This paper introduces a recurrent generative adversarial network (R-GAN) for generating realistic energy consumption data by learning from real data. Generativea adversarial …

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 …

An Environment For Developing Incremental Learning Applications For Data Streams, Farzin Sarvaramini

Electronic Thesis and Dissertation Repository

Smart cities look to leverage technology, particularly sensors, and software to provide improved services for its citizenry and enhanced operational efficiencies. Cities look to develop applications that can process data from sensors and other sources to gain insights into operation, enable them to improve operations and inform city leadership. Such applications often need to process streams of data from sensors or other sources to provide city staff with insights into city operations. However, cities are faced with limited budgets and limited staff. The development of applications by third parties can be extremely expensive. One alternative is to identify tools for …

High Multiplicity Strip Packing Problem With Three Rectangle Types, Andy Yu

Electronic Thesis and Dissertation Repository

The two-dimensional strip packing problem (2D-SPP) involves packing a set R = {r₁, ..., r_n} of n rectangular items into a strip of width 1 and unbounded height, where each rectangular item r_i has width 0 < w_i ≤ 1 and height 0 < h_i ≤ 1. The objective is to find a packing for all these items, without overlaps or rotations, that minimizes the total height of the strip used. 2D-SPP is strongly NP-hard and has practical applications including stock cutting, scheduling, and reducing peak power demand in smart-grids.

This thesis considers …

A New Algorithm For Primer Design, Debanjan Guha Roy

Electronic Thesis and Dissertation Repository

The Polymerase Chain Reaction (PCR) technology is widely used to create DNA copies. It has impacted many diverse fields including genetics, forensics, molecular paleontology, medical applications and environmental microbiology.

The main object in PCR is a primer, a short single strand of DNA, about 18-25 bases long, that serves as the starting point of DNA synthesis. Primers are essential for DNA replication because the enzymes that catalyze this process, DNA polymerases, can only add new nucleotides to an existing strand of DNA. The PCR starts at the 3’-end of the primer and copies the opposite strand.Designing good primers is essential …

Automatic Recall Of Software Lessons Learned For Software Project Managers, Tamer Mohamed Abdellatif Mohamed, Luiz Fernando Capretz, Danny Ho

Electrical and Computer Engineering Publications

Context: Lessons learned (LL) records constitute the software organization memory of successes and failures. LL are recorded within the organization repository for future reference to optimize planning, gain experience, and elevate market competitiveness. However, manually searching this repository is a daunting task, so it is often disregarded. This can lead to the repetition of previous mistakes or even missing potential opportunities. This, in turn, can negatively affect the organization’s profitability and competitiveness.

Objective: We aim to present a novel solution that provides an automatic process to recall relevant LL and to push those LL to project managers. This will dramatically …

Can We Rely On Smartphone Applications?, Sonia Meskini, Ali Bou Nassif, Luiz Fernando Capretz

Electrical and Computer Engineering Publications

Smartphones are becoming necessary tools in the daily lives of millions of users who rely on these devices and their applications. There are thousands of applications for smartphone devices such as the iPhone, Blackberry, and Android, thus their reliability has become paramount for their users. This work aims to answer two related questions: (1) Can we assess the reliability of mobile applications by using the traditional reliability models? (2) Can we model adequately the failure data collected from many users? Firstly, it has been proved that the three most used software reliability models have fallen short of the mark when …

Ml4iot: A Framework To Orchestrate Machine Learning Workflows On Internet Of Things Data, Jose Miguel Alves, Leonardo Honorio, Miriam A M Capretz

Electrical and Computer Engineering Publications

Internet of Things (IoT) applications generate vast amounts of real-time data. Temporal analysis of these data series to discover behavioural patterns may lead to qualified knowledge affecting a broad range of industries. Hence, the use of machine learning (ML) algorithms over IoT data has the potential to improve safety, economy, and performance in critical processes. However, creating ML workflows at scale is a challenging task that depends upon both production and specialized skills. Such tasks require investigation, understanding, selection, and implementation of specific ML workflows, which often lead to bottlenecks, production issues, and code management complexity and even then may …

Similarity-Based Chained Transfer Learning For Energy Forecasting With Big Data, Yifang Tian, Ljubisa Sehovac, Katarina Grolinger

Electrical and Computer Engineering Publications

Smart meter popularity has resulted in the ability to collect big energy data and has created opportunities for large-scale energy forecasting. Machine Learning (ML) techniques commonly used for forecasting, such as neural networks, involve computationally intensive training typically with data from a single building or a single aggregated load to predict future consumption for that same building or aggregated load. With hundreds of thousands of meters, it becomes impractical or even infeasible to individually train a model for each meter. Consequently, this paper proposes Similarity-Based Chained Transfer Learning (SBCTL), an approach for building neural network-based models for many meters by …

Vertical Ionization Energies From The Average Local Electron Energy Function, Amer Marwan El-Samman

Electronic Thesis and Dissertation Repository

It is a non-intuitive but well-established fact that the first and higher vertical ionization energies (VIE) of any N-electron system are encoded in the system's ground-state electronic wave function. This makes it possible to compute VIEs of any atom or molecule from its ground-state wave function directly, without performing calculations on the (N-1)-electron states. In practice, VIEs can be extracted from the wave function by using the (extended) Koopmans' theorem or by taking the asymptotic limit of certain wave-function-based quantities such as the ratio of kinetic energy density to the electron density. However, when the wave function is expanded in …

A Programming Model For Internetworked Things, Hao Jiang

Electronic Thesis and Dissertation Repository

The Internet of Things (IoT) emerges as a system paradigm that encompasses a wide spectrum of technologies and protocols related to Internetworking, services computing, and device connectivity. The main objective is to achieve an environment whereby physical devices and everyday objects can communicate and interact with each other over the Internet. The Internet of Things is heralded as the next generation Internet, and introduces significant opportunities for novel applications in many different domains. What is missing right now is a programming model whereby developers as well as end-users can specify any addressable resource at a higher level of abstraction, and …

High Multiplicity Strip Packing, Andrew Bloch-Hansen

Electronic Thesis and Dissertation Repository

In the two-dimensional high multiplicity strip packing problem (HMSPP), we are given k distinct rectangle types, where each rectangle type T_i has n_i rectangles each with width 0 < w_i and height 0 < h_i The goal is to pack these rectangles into a strip of width 1, without rotating or overlapping the rectangles, such that the total height of the packing is minimized.

Let OPT(I) be the optimal height of HMSPP on input I. In this thesis, we consider HMSPP for the case when k = 3 and present an OPT(I) + 5/3 polynomial time approximation algorithm for …

New Algorithms For Computing Field Of Vision Over 2d Grids, Evan Debenham

Electronic Thesis and Dissertation Repository

In many computer games checking whether one object is visible from another is very important. Field of Vision (FOV) refers to the set of locations that are visible from a specific position in a scene of a computer game. Once computed, an FOV can be used to quickly determine the visibility of multiple objects from a given position.

This thesis summarizes existing algorithms for FOV computation, describes their limitations, and presents new algorithms which aim to address these limitations. We first present an algorithm which makes use of spatial data structures in a way which is new for FOV calculation. …

A New Approach To Sequence Local Alignment: Normalization With Concave Functions, Qiang Zhou

Electronic Thesis and Dissertation Repository

Sequence local alignment is to find two subsequences from the input two sequences respectively, which can produce the highest similarity degree among all other pairs of subsequences. The Smith-Waterman algorithm is one of the most important technique in sequence local alignment, especially in computational molecular biology. This algorithm can guarantee that the optimal local alignment can be found with respect to the distance or similarity metric. However, the optimal solution obtained by Smith-Waterman is not biologically meaningful, since it may contain small pieces of irrelevant segments, but as long as they are not strong enough, the algorithm still take them …

Spatiotemporal Forecasting At Scale, Rafael Felipe Nascimento De Aguiar

Electronic Thesis and Dissertation Repository

Spatiotemporal forecasting can be described as predicting the future value of a variable given when and where it will happen. This type of forecasting task has the potential to aid many institutions and businesses in asking questions, such as how many people will visit a given hospital in the next hour. Answers to these questions have the potential to spur significant socioeconomic impact, providing privacy-friendly short-term forecasts about geolocated events, which in turn can help entities to plan and operate more efficiently. These seemingly simple questions, however, present complex challenges to forecasting systems. With more GPS-enabled devices connected every year, …

An Empirical Study Of User Support Tools In Open Source Software, Arif Raza, Luiz Fernando Capretz, Shuib Basri

Electrical and Computer Engineering Publications

End users’ positive response is essential for the success of any software. This is true for both commercial and Open Source Software (OSS). OSS is popular not only because of its availability, which is usually free but due to the user support it provides, generally through public platforms. The study model of this research establishes a relationship between OSS user support and available support tools. To conduct this research, we used a dataset of 100 OSS projects in different categories and examined five user support tools provided by different OSS projects. The results show that project trackers, user mailing lists, …

Algebraic Neural Architecture Representation, Evolutionary Neural Architecture Search, And Novelty Search In Deep Reinforcement Learning, Ethan C. Jackson

Electronic Thesis and Dissertation Repository

Evolutionary algorithms have recently re-emerged as powerful tools for machine learning and artificial intelligence, especially when combined with advances in deep learning developed over the last decade. In contrast to the use of fixed architectures and rigid learning algorithms, we leveraged the open-endedness of evolutionary algorithms to make both theoretical and methodological contributions to deep reinforcement learning. This thesis explores and develops two major areas at the intersection of evolutionary algorithms and deep reinforcement learning: generative network architectures and behaviour-based optimization. Over three distinct contributions, both theoretical and experimental methods were applied to deliver a novel mathematical framework and experimental …

Forecasting Building Energy Consumption With Deep Learning: A Sequence To Sequence Approach, Ljubisa Sehovac, Cornelius Nesen, Katarina Grolinger

Electrical and Computer Engineering Publications

Energy Consumption has been continuously increasing due to the rapid expansion of high-density cities, and growth in the industrial and commercial sectors. To reduce the negative impact on the environment and improve sustainability, it is crucial to efficiently manage energy consumption. Internet of Things (IoT) devices, including widely used smart meters, have created possibilities for energy monitoring as well as for sensor based energy forecasting. Machine learning algorithms commonly used for energy forecasting such as feedforward neural networks are not well-suited for interpreting the time dimensionality of a signal. Consequently, this paper uses Recurrent Neural Networks (RNN) to capture time …

An Adaptive Weighted Average (Wav) Reprojection Algorithm For Image Denoising, Halimah Alsurayhi

Electronic Thesis and Dissertation Repository

Patch-based denoising algorithms have an effective improvement in the image denoising domain. The Non-Local Means (NLM) algorithm is the most popular patch-based spatial domain denoising algorithm. Many variants of the NLM algorithm have proposed to improve its performance. Weighted Average (WAV) reprojection algorithm is one of the most effective improvements of the NLM denoising algorithm. Contrary to the NLM algorithm, all the pixels in the patch contribute into the averaging process in the WAV reprojection algorithm, which enhances the denoising performance. The key parameters in the WAV reprojection algorithm are kept fixed regardless of the image structure. In this thesis, …

Virtual Sensor Middleware: A Middleware For Managing Iot Data For The Fog-Cloud Platform, Fadi Almahamid

Electronic Thesis and Dissertation Repository

Internet of Things is a massively growing field where billions of devices are connected to the Internet using different protocols and produce an enormous amount of data. The produced data is consumed and processed by different applications to make operations more efficient. Application development is challenging, especially when applications access sensor data since IoT devices use different communication protocols.

The existing IoT architectures address some of these challenges. This thesis proposes an IoT Middleware that provides applications with the abstraction required of IoT devices while distributing the processing of sensor data to provide a real-time or near real-time response and …

Incorporating Figure Captions And Descriptive Text Into Mesh Term Indexing: A Deep Learning Approach, Xindi Wang

Electronic Thesis and Dissertation Repository

The exponential increase of available documents online makes document classification an important application in natural language processing. The goal of text classification is to automatically assign categories to documents. Traditional text classifiers depend on features, such as, vocabulary and user-specified information which mainly relies on prior knowledge. In contrast, deep learning automatically learns effective features from data instead of adopting human-designed features. In this thesis, we specifically focus on biomedical document classification. Beyond text information from abstract and title, we also consider image and table captions, as well as paragraphs associated with images and tables, which we demonstrate to be …

Studies On The Software Testing Profession, Luiz Fernando Capretz, Pradeep Waychal, Jingdong Jia, Daniel Varona, Yadira Tejeda Saldaña

Electrical and Computer Engineering Publications

This paper attempts to understand motivators and de-motivators that influence the decisions of software professionals to take up and sustain software testing careers across four different countries, i.e. Canada, China, Cuba, and India. The research question can be framed as “How many software professionals across different geographies are keen to take up testing careers, and what are the reasons for their choices?” Towards that, we developed a cross-sectional but simple survey-based instrument. In this study we investigated how software testers perceived and valued what they do and their environmental settings. The study pointed out the importance of visualizing software testing …

Comparing The Popularity Of Testing Career Among Canadian, Chinese, And Indian Students, Luiz Fernando Capretz, Pradeep Waychal, Jingdong Jia

Electrical and Computer Engineering Publications

Despite its importance, software testing is, arguably, the least understood part of the software life cycle and still the toughest to perform correctly. Many researchers and practitioners have been working to address the situation. However, most of the studies focus on the process and technology dimensions and only a few on the human dimension of testing, in spite of the reported relevance of human aspects of software testing. Testers need to understand various stakeholders’ explicit and implicit requirements, be aware of how developers work individually and in teams, and develop skills to report test results wisely to stakeholders. These multifaceted …

Design And Job Rotation In Software Engineering: Results From An Industrial Study, Ronnie Santos, Maria Teresa Baldassarre, Fabio Q. B. Silva Dr., Cleyton Magalhaes, Luiz Fernando Capretz, Jorge Correia-Neto

Electrical and Computer Engineering Publications

Job rotation is a managerial practice to be applied in the organizational environment to reduce job monotony, boredom, and exhaustion resulting from job simplification, specialization, and repetition. Previous studies have identified and discussed the use of project-to-project rotations in software practice, gathering empirical evidence from qualitative and field studies and pointing out set of work-related factors that can be positively or negatively affected by this practice. Goal: We aim to collect and discuss the use of job rotation in software organizations in order to identify the potential benefits and limitations of this practice supported by the statement of existing theories …

Machine Learning For Stock Prediction Based On Fundamental Analysis, Yuxuan Huang

Electronic Thesis and Dissertation Repository

Application of machine learning for stock prediction is attracting a lot of attention in recent years. A large amount of research has been conducted in this area and multiple existing results have shown that machine learning methods could be successfully used toward stock predicting using stocks’ historical data. Most of these existing approaches have focused on short term prediction using stocks’ historical price and technical indicators. In this thesis, we prepared 22 years’ worth of stock quarterly financial data and investigated three machine learning algorithms: Feed-forward Neural Network (FNN), Random Forest (RF) and Adaptive Neural Fuzzy Inference System (ANFIS) for …

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. …

Haptics-Enabled, Gpu Augmented Surgical Simulation Platform For Glenoid Reaming, Vlad Popa

Electronic Thesis and Dissertation Repository

Surgical simulators are technological platforms that provide virtual substitutes to the current cadaver-based medical training models. The advantages of exposure to these devices have been thoroughly studied, with enhanced surgical proficiency being one of the assets gained after extensive use. While simulators have already penetrated numerous medical domains, the field of orthopedics remains stagnant despite a demand for the ability to practice uncommon surgeries, such as total shoulder arthroplasty (TSA). Here we extrapolate the algorithms of an inhouse software engine revolving around glenoid reaming, a critical step of TSA. The project’s purpose is to provide efficient techniques for future simulators, …

Applicability Of Recurrent Neural Networks To Player Data Analysis In Freemium Video Games, Jonathan Tan

Electronic Thesis and Dissertation Repository

We demonstrate the applicability and practicality of recurrent neural networks (RNNs), a machine learning methodology suited for sequential data, on player data from the mobile video game My Singing Monsters. Since this data comes in as a stream of events, RNNs are a natural solution for analyzing this data with minimal preprocessing. We apply RNNs to monitor and forecast game metrics, predict player conversion, estimate lifetime player value, and cluster player behaviours. In each case, we discuss why the results are interesting, how the trained models can be applied in a business setting, and how the preliminary work can …

Approximation Algorithms For Problems In Makespan Minimization On Unrelated Parallel Machines, Daniel R. Page

Electronic Thesis and Dissertation Repository

A fundamental problem in scheduling is makespan minimization on unrelated parallel machines (R||C_max). Let there be a set J of jobs and a set M of parallel machines, where every job J_j ∈ J has processing time or length p_i,j ∈ ℚ⁺ on machine M_i ∈ M. The goal in R||C_max is to schedule the jobs non-preemptively on the machines so as to minimize the length of the schedule, the makespan. A ρ-approximation algorithm produces in polynomial time a feasible solution such that its objective value is within a multiplicative factor ρ of …

Local Search Approximation Algorithms For Clustering Problems, Nasim Samei

Electronic Thesis and Dissertation Repository

In this research we study the use of local search in the design of approximation algorithms for NP-hard optimization problems. For our study we have selected several well-known clustering problems: k-facility location problem, minimum mutliway cut problem, and constrained maximum k-cut problem.

We show that by careful use of the local optimality property of the solutions produced by local search algorithms it is possible to bound the ratio between solutions produced by local search approximation algorithms and optimum solutions. This ratio is known as the locality gap.

The locality gap of our algorithm for the k-uncapacitated facility …