Computer Sciences Commons^™

Message From The General Chairs, Falko Dressler, Sajal K. Das

Computer Science Faculty Research & Creative Works

No abstract provided.

The Impact Of Special Interest Groups On The Federal Dietary Guidelines: Consequences For American Health, Dory Mcmillan

School of Professional Studies

This research paper explores the impact of relationships between lobbyists and both the USDA and HSS, and the impact these relationships have on the Dietary Guidelines for Americans that the agencies work together to create. The paper focuses specifically on the information the guidelines present in regard to red meat consumption, and the impacts this may have on American health, and healthcare costs associated. It was hypothesized that a relationship would be found between special interest groups and the U.S. Department of Health and Human Services and/or the U.S. Department of Agriculture. Research found there was a relationship between special …

The Killingly Mascot Case Study, Jordan Lumpkins

School of Professional Studies

In the summer of 2019, in Killingly, Connecticut the local Board of Education voted to retire the "Redmen" mascot name it had used for nearly a century. This legislation was widely opposed and received extensive media coverage. Within a few months, the town experienced a massive political referendum where several local Board of Education members and Councilmen were replaced by single issue politicians promising to reinstate the "Redmen" name. Now holding a majority on the Board of Education, these Board members made Killingly the first school in U.S. history to reinstate a mascot after being deemed "derogatory."

It is the …

Higher Education Responses To Crisis: A Case Study Of Clark University And The Pandemic Of 2020, Lisa Gillingham

School of Professional Studies

The COVID-19 pandemic has delivered an existential challenge to universities and other academic institutions at a time when they are already grappling with other weighty issues that may alter the fabric of higher education. COVID-19 has forced these institutions to consider and employ new ways of conducting its work with a sense urgency that is unprecedented in the recent history of the academy. The rate of learning around these models is rapid, and Higher Education is ripe for change.

Clark University has addressed the pandemic with a plan to protect and pivot using strategies that support the continuation of its …

Special Issue On 6g Wireless Systems, Periklis Chatzimisios, David Soldani, Abbas Jamalipour, Antonio Manzalini, Sajal K. Das

Computer Science Faculty Research & Creative Works

No abstract provided.

A Distance Based Multisample Test For High-Dimensional Compositional Data With Applications To The Human Microbiome, Qingyang Zhang, Thy Dao

Mathematical Sciences Faculty Publications and Presentations

Background

Compositional data refer to the data that lie on a simplex, which are common in many scientific domains such as genomics, geology and economics. As the components in a composition must sum to one, traditional tests based on unconstrained data become inappropriate, and new statistical methods are needed to analyze this special type of data.

Results

In this paper, we consider a general problem of testing for the compositional difference between K populations. Motivated by microbiome and metagenomics studies, where the data are often over-dispersed and high-dimensional, we formulate a well-posed hypothesis from a Bayesian point of view and …

Adaptive Discounting In Reinforcement Learning, Milan Zinzuvadiya

Master's Theses

In Markov Decision Process (MDP) models of sequential decision-making, it is common practice to account for temporal discounting by incorporating a constant discount factor. While the effectiveness of fixed-rate discounting in various Reinforcement Learning (RL) settings is well-established, the efficiency of this scheme has been questioned in recent studies. Another notable shortcoming of fixed-rate discounting stems from abstracting away the experiential information of the agent, which is shown to be a significant component of delay discounting in human cognition. To address this issue, this thesis proposes a novel method for adaptive discounting entitled State-wise Adaptive Discounting from Experience (SADE). This …

Fundamentals Of Human-Centric Artificial Intelligence (A.I.): Comparative Analysis Of Europe And The U. S. Landscape, Torré A. Williams

Cybersecurity Undergraduate Research Showcase

This research is a comparative analysis of human-centric Artificial Intelligence (A.I.) in Europe and the U.S. This research establishes fundamentals that are critical to what makes A.I. human-centric. This research contains eight phases: 1) Lawful A.I.; 2) Robust A.I.; 3) Ethical A.I.; 4) Human-centric A.I.; 5) Current State of A.I.; 6) A.I. in Europe; 7) A.I. in the U.S.; 9) Importance of Human-centric A.I. This research shows that there are still ongoing changes with having a human-centric A.I. and why it is very important to society. This research is the beginning of the making of a successful and reliable human-centric …

Generating Adversarial Examples For Recruitment Ranking Algorithms, Anahita Samadi

Computer Science and Engineering Theses - Archive

There is no doubt that recruitment process plays an important role for both employers and applicants. Based on huge number of job candidates and open vacancies, recruitment process is expensive, time consuming and stressful for both applicants and companies. In today’s world so many recruitment processes are based on machine learning techniques. Therefore, it is very important to ensure security of these algorithms. Adversarial examples are proposed to examine vulnerability of machine leaning algorithms. Many research studies have been done on evaluating the resistance of artificial intelligence-based systems, in computer vision and text classification, against adversarial examples. However, to the …

Semi-Automatic Hand Pose Estimation Using A Single Depth Camera, Giffy Jerald Chris

Computer Science and Engineering Theses - Archive

This paper addresses the problem of 3D hand pose annotations using a single depth camera. Although hand pose estimation methods rely critically on accurate 3D training data, creating such reliable training data is challenging and labor intensive. We propose a semi-automatic method for efficiently and accurately labeling the 3D hand key-points in a hand depth video. The process starts by selecting a subset of frames that are representative of all the frames in the dataset and the annotator only provides an estimate of the 2D hand key-points in these selected frames. We use this information to infer the 3D location …

A Survey On Ddos Attacks In Edge Servers, Iftakhar Ahmad

Computer Science and Engineering Theses - Archive

In modern times, the need for latency sensitive applications is growing rapidly. Cloud computing infrastructure is unable to provide support to such delay sensitive applications. Therefore, a new paradigm called edge computing has emerged. In edge computing various paradigms like Fog, Cloudlet, Mobile Edge Computing, etc. provide real-time, location aware services to users. As a result number of requests are generated for processing in the edge servers. If these edge servers for some reason become unavailable for providing service, users will not be able to perform their delay sensitive or location aware operations. Like other servers in the network, edge …

Incomplete Time Series Forecasting Using Generative Neural Networks, Harshit Tarun Shah

Computer Science and Engineering Theses - Archive

Dealing with missing data is a long pervading problem and it becomes more challenging when forecasting time series data because of the complex relationships between data and time, which is why incomplete data can lead to unreliable results. While some general-purpose methods like mean, zero, or median imputation can be employed to alleviate the problem, they might disrupt the inherent structure and the underlying data distributions. Another problem associated with conventional time series forecasting methods whose goal is to predict mean values is that they might sometimes overlook the variance or fluctuations in the input data and eventually lead to …

Link Prediction Based Face Clustering Using Variational Attentional Graph Autoencoder, Harish Deepak Verlekar

Computer Science and Engineering Theses - Archive

In this work, we address the problem of clustering faces according to their individual identities present inherently in the dataset.The current clustering frameworks are either based on some heuristic method or require labelled data for training the models,also some of them make assumptions on data distribution or shape of the clusters.We have framed the problem of forming clusters to that of link prediction on graphs and learn how to do that in a completely unsupervised way by proposing to use Variational Graph Autoencoders and use Graph Attentional Network as the Encoder. We call this network as Variational Attentional Graph Autoencoder(VAGAE).Our …

Early Detection Of Glaucoma Using Modified Residual U-Net Convolutional Neural Network, Balasubramaniam Theetharappan

Computer Science and Engineering Theses - Archive

Glaucoma is the second leading cause of blindness all over the world, with apparently 75 million cases reported worldwide in 2018. If it’s not diagnosed at an early stage, glaucoma may cause irreversible damage to the optic nerve which results in blindness. The Optic head examination is the widely used structured diagnosis approach in the current medical field for Glaucoma detection which involves measuring the Optic Cup-to-Disc ratio from the fundus image. Estimation of Optic Cup-to-Disc requires accurate segmentation of the Optic Cup and Optic Disc from the fundus which is a tedious and time-consuming task even for the experienced …

Internet Of Things (Iot): Cybersecurity Risks In Healthcare, Ruhi Patel

Cybersecurity Undergraduate Research Showcase

The rapid growth and investment in the Internet of Things (IoT) has significantly impacted how individuals and industries operate. The Internet of Things (IoT) refers to a network of physical, technology-embedded objects that communicate, detect, and interact with their external environment or internal state (Hung, 2017). According to Tankovska (2020), IoT devices are estimated to reach 21.5 billion units by 2025. This technological boom is leading various industrial sectors to notice a quick increase in cybersecurity risks and threats. One industrial sector has been particularly vulnerable to numerous cyber threats across the globe: healthcare. Oliver Noble (2020), a data encryption …

Vision-Based Analytics For Improved Ai-Driven Iot Applications, Amit Sharma

Dissertations and Theses Collection (Open Access)

Proliferation of Internet of Things (IoT) sensor systems, primarily driven by cheaper embedded hardware platforms and wide availability of light-weight software platforms, has opened up doors for large-scale data collection opportunities. The availability of massive amount of data has in-turn given way to rapidly growing machine learning models e.g. You Only Look Once (YOLO), Single-Shot-Detectors (SSD) and so on. There has been a growing trend of applying machine learning techniques, e.g., object detection, image classification, face detection etc., on data collected from camera sensors and therefore enabling plethora of vision-sensing applications namely self-driving cars, automatic crowd monitoring, traffic-flow analysis, occupancy …

Changes In R 3.6–4.0, Tomas Kalibera, Sebastian Meyer, Kurt Hornik

The R Journal

We give a selection of the most important changes in R 4.0.0 and in the R 3.6 release series. Some statistics on source code commits and bug tracking activities are also provided.

Analyzing Basket Trials Under Multisource Exchangeability Assumptions, Michael J. Kane, Nan Chen, Alexander M. Kaizer, Xun Jiang, H Amy Xia, Brian P. Hobbs

The R Journal

Basket designs are prospective clinical trials that are devised with the hypothesis that the presence of selected molecular features determine a patient’s subsequent response to a particular “targeted” treatment strategy. Basket trials are designed to enroll multiple clinical subpopulations to which it is assumed that the therapy in question offers beneficial efficacy in the presence of the targeted molecular profile. The treatment, however, may not offer acceptable efficacy to all subpopulations enrolled. Moreover, for rare disease settings, such as oncology wherein these trials have become popular, marginal measures of statistical evidence are difficult to interpret for sparsely enrolled subpopulations. Consequently, …

Motbfs: An R Package For Learning Hybrid Bayesian Networks Using Mixtures Of Truncated Basis Functions, Inmaculada Pérez-Bernabé, Ana D. Maldonado, Antonio Salmerón, Thomas D. Nielsen

The R Journal

This paper introduces MoTBFs, an R package for manipulating mixtures of truncated basis functions. This class of functions allows the representation of joint probability distributions involving discrete and continuous variables simultaneously, and includes mixtures of truncated exponentials and mixtures of polynomials as special cases. The package implements functions for learning the parameters of univariate, multivariate, and conditional distributions, and provides support for parameter learning in Bayesian networks with both discrete and continuous variables. Probabilistic inference using forward sampling is also implemented. Part of the functionality of the MoTBFs package relies on the bnlearn package, which includes functions for learning the …

A Graphical Eda Tool With Ggplot2: Brinton, Pere Millán-Martínez, Ramon Oller

The R Journal

We present brinton package, which we developed for graphical exploratory data analysis in R. Based on ggplot2, gridExtra and rmarkdown, brinton package introduces wideplot() graphics for exploring the structure of a dataset through a grid of variables and graphic types. It also introduces longplot() graphics, which present the entire catalog of available graphics for representing a particular variable using a grid of graphic types and variations on these types. Finally, it introduces the plotup() function, which complements the previous two functions in that it presents a particular graphic for a specific variable of a dataset. This set of functions is …

Nts: An R Package For Nonlinear Time Series Analysis, Xialu Liu, Rong Chen, Ruey Tsay

The R Journal

Linear time series models are commonly used in analyzing dependent data and in forecasting. On the other hand, real phenomena often exhibit nonlinear behavior and the observed data show nonlinear dynamics. This paper introduces the R package NTS that offers various computational tools and nonlinear models for analyzing nonlinear dependent data. The package fills the gaps of several outstanding R packages for nonlinear time series analysis. Specifically, the NTS package covers the implementation of threshold autoregressive (TAR) models, autoregressive conditional mean models with exogenous variables (ACMx), functional autoregressive models, and state-space models. Users can also evaluate and compare the performance …

Aquadtree: An R Package For Quadtree Anonymization Of Point Data, Raymond Lagonigro, Ramon Oller, Joan Carles Martori

The R Journal

The demand for precise data for analytical purposes grows rapidly among the research community and decision makers as more geographic information is being collected. Laws protecting data privacy are being enforced to prevent data disclosure. Statistical institutes and agencies need methods to preserve confidentiality while maintaining accuracy when disclosing geographic data. In this paper we present the AQuadtree package, a software intended to produce and deal with official spatial data making data privacy and accuracy compatible. The lack of specific methods in R to anonymize spatial data motivated the development of this package, providing an automatic aggregation tool to anonymize …

Kspm: A Package For Kernel Semi-Parametric Models, Catherine Schramm, Sébastien Jacquemont, Karim Oualkacha, Aurélie Labbe, Celia M. T. Greenwood

The R Journal

Kernel semi-parametric models and their equivalence with linear mixed models provide analysts with the flexibility of machine learning methods and a foundation for inference and tests of hypothesis. These models are not impacted by the number of predictor variables, since the kernel trick transforms them to a kernel matrix whose size only depends on the number of subjects. Hence, methods based on this model are appealing and numerous, however only a few R programs are available and none includes a complete set of features. Here, we present the KSPM package to fit the kernel semi-parametric model and its extensions in …

Ordinalclust: An R Package To Analyze Ordinal Data, Margot Selosse, Julien Jacques, Christophe Biernacki

The R Journal

Ordinal data are used in many domains, especially when measurements are collected from people through observations, tests, or questionnaires. ordinalClust is an innovative R package dedicated to ordinal data that provides tools for modeling, clustering, co-clustering and classifying such data. Ordinal data are modeled using the BOS distribution, which is a model with two meaningful parameters referred to as "position" and "precision". The former indicates the mode of the distribution and the latter describes how scattered the data are around the mode: the user is able to easily interpret the distribution of their data when given these two parameters. The …

A Fast And Scalable Implementation Method For Competing Risks Data With The R Package Fastcmprsk, Eric S. Kawaguchi, Jenny I. Shen, Gang Li, Marc A. Suchard

The R Journal

Advancements in medical informatics tools and high-throughput biological experimentation make large-scale biomedical data routinely accessible to researchers. Competing risks data are typical in biomedical studies where individuals are at risk to more than one cause (type of event) which can preclude the others from happening. The Fine and Gray (1999) proportional subdistribution hazards model is a popular and well-appreciated model for competing risks data and is currently implemented in a number of statistical software packages. However, current implementations are not computationally scalable for large-scale competing risks data. We have developed an R package, fastcmprsk, that uses a novel forward-backward scan …

Six Years Of Shiny In Research: Collaborative Development Of Web Tools In R, Peter Kasprzak, Lachlan Mitchell, Olena Kravchuk, Andy Timmins

The R Journal

The use of Shiny in research publications is investigated over the six and a half years since the appearance of this popular web application framework for R, which has been utilised in many varied research areas. While it is demonstrated that the complexity of Shiny applications is limited by the background architecture, and real security concerns exist for novice app developers, the collaborative benefits are worth attention from the wider research community. Shiny simplifies the use of complex methodologies for people of different specialities, at the level of proficiency appropriate for the end user. This enables a diverse community of …

Assembling Pharmacometric Datasets In R: The Puzzle Package, Mario González-Sales, Olivier Barrière, Pierre Olivier Tremblay, Guillaume Bonnefois, Julie Desrochers, Fahima Nekka

The R Journal

Pharmacometric analyses are integral components of the drug development process. The core of each pharmacometric analysis is a dataset. The time required to construct a pharmacometrics dataset can sometimes be higher than the effort required for the modeling per se. To simplify the process, the puzzle R package has been developed aimed at simplifying and facilitating the time consuming and error prone task of assembling pharmacometrics datasets.

Puzzle consist of a series of functions written in R. These functions create, from tabulated files, datasets that are compatible with the formatting requirements of the gold standard non-linear mixed effects modeling …

Fitzroy: An R Package To Encourage Reproducible Sports Analysis, Robert Nguyen, James Day, David Warton, Oscar Lane

The R Journal

The importance of reproducibility, and the related issue of open access to data, has received a lot of recent attention. Momentum on these issues is gathering in the sports analytics community. While Australian Rules football (AFL) is the leading commercial sport in Australia, unlike popular international sports, there has been no mechanism for the public to access comprehensive statistics on players and teams. Expert commentary currently relies heavily on data that isn’t made readily accessible and this produces an unnecessary barrier for the development of an inclusive sports analytics community. We present the R package fitzRoy to provide easy access …

Comparing Multiple Survival Functions With Crossing Hazards In R, Hsin-Wen Chang, Pei-Yuan Tsai, Jen-Tse Kao, Guo-You Lan

The R Journal

It is frequently of interest in time-to-event analysis to compare multiple survival functions nonparametrically. However, when the hazard functions cross, tests in existing R packages do not perform well. To address the issue, we introduce the package survELtest, which provides tests for comparing multiple survival functions with possibly crossing hazards. Due to its powerful likelihood ratio formulation, this is the only R package to date that works when the hazard functions cross. We illustrate the use of the procedures in survELtest by applying them to data from randomized clinical trials and simulated datasets. We show that these methods lead …

The Biglasso Package: A Memory- And Computation-Efficient Solver For Lasso Model Fitting With Big Data In R, Yaohui Zeng, Patrick Breheny

The R Journal

Penalized regression models such as the lasso have been extensively applied to analyzing high-dimensional data sets. However, due to memory limitations, existing R packages like glmnet and ncvreg are not capable of fitting lasso-type models for ultrahigh-dimensional, multi-gigabyte data sets that are increasingly seen in many areas such as genetics, genomics, biomedical imaging, and high-frequency finance. In this research, we implement an R package called biglasso that tackles this challenge. biglasso utilizes memory-mapped files to store the massive data on the disk, only reading data into memory when necessary during model fitting, and is thus able to handle out-of-core computation …