Computer Sciences Commons^™

Artificial Neural Network Models For Pattern Discovery From Ecg Time Series, Mehakpreet Kaur

College of Graduate Studies: Theses & Dissertations

Artificial Neural Network (ANN) models have recently become de facto models for deep learning with a wide range of applications spanning from scientific fields such as computer vision, physics, biology, medicine to social life (suggesting preferred movies, shopping lists, etc.). Due to advancements in computer technology and the increased practice of Artificial Intelligence (AI) in medicine and biological research, ANNs have been extensively applied not only to provide quick information about diseases, but also to make diagnostics accurate and cost-effective. We propose an ANN-based model to analyze a patient's electrocardiogram (ECG) data and produce accurate diagnostics regarding possible heart diseases …

Analyze Informant-Based Questionnaire For The Early Diagnosis Of Senile Dementia Using Deep Learning, Fubao Zhu, Xiaonan Li, Daniel Mcgonigle, Haipeng Tang, Zhuo He, Chaoyang Zhang, Guang-Uei Hung, Pai-Yi Chiu, Weihua Zhou

Michigan Tech Publications, Part 1

OBJECTIVE: This paper proposes a multiclass deep learning method for the classification of dementia using an informant-based questionnaire.

METHODS: A deep neural network classification model based on Keras framework is proposed in this paper. To evaluate the advantages of our proposed method, we compared the performance of our model with industry-standard machine learning approaches. We enrolled 6,701 individuals, which were randomly divided into training data sets (6030 participants) and test data sets (671 participants). We evaluated each diagnostic model in the test set using accuracy, precision, recall, and F1-Score.

RESULTS: Compared with the seven conventional machine learning algorithms, the DNN …

Technological Challenges And Innovations In Cybersecurity And Networking Technology Program, Syed R. Zaidi, Ajaz Sana, Aparicio Carranza

Publications and Research

This era is posing a unique challenge to the Cybersecurity and related Engineering Technology areas, stimulated by the multifaceted technological boom expressed in accelerated globalization, digital transformation, the cloud, mobile access apps, and the Internet of Things (IoT)—where more and more devices are connected to the Internet every day. As the use of new Internet-based technologies increase; so does the risk of theft and misuse of sensitive information. This demands the awareness of cyber-criminality and the need for cyber hygiene in corporations, small businesses, and the government. As the need for experienced cybersecurity specialists has skyrocketed in recent years and …

Identifying And Disentangling Interleaved Activities Of Daily Living From Sensor Data, Eoin Rogers

Doctoral

Activity discovery (AD) refers to the unsupervised extraction of structured activity data from a stream of sensor readings in a real-world or virtual environment. Activity discovery is part of the broader topic of activity recognition, which has potential uses in fields as varied as social work and elder care, psychology and intrusion detection. Since activity recognition datasets are both hard to come by, and very time consuming to label, the development of reliable activity discovery systems could be of significant utility to the researchers and developers working in the field, as well as to the wider machine learning community.

This …

Enhancing Learning With Primitive-Decomposed Cognitive Representations, Jamie C. Macbeth

Computer Science: Faculty Publications

This paper proposes work that applies insights from meaning representation systems for in-depth natural language understanding to representations for self-supervised learning systems, which show promise in developing complex, deeply-nested symbolic structures through self-motivated exploration of their environments. The core of the representation system transforms language inputs into language-free structures that are complex combinations of conceptual primitives, forming a substrate for human-like understanding and common-sense reasoning. We focus on decomposing representations of expectation, intention, planning, and decision-making which are essential to a self-motivated learner. These meaning representations may enhance learning by enabling a rich array of mappings between new experiences and …

Image Instance Segmentation: Using The Cirsy System To Identify Small Objects In Low Resolution Images, Orghomisan William Omatsone

Dissertations

The CIRSY system (or Chick Instance Recognition System) is am image processing system developed as part of this research to detect images of chicks in highly-populated images that uses the leading algorithm in instance segmentation tasks, called the Mask R-CNN. It extends on the Faster R-CNN framework used in object detection tasks, and this extension adds a branch to predict the mask of an object along with the bounding box prediction. Mask R-CNN has proven to be effective ininstance segmentation and object de-tection tasks after outperforming all existing models on evaluation of the Microsoft Common Objects in Context (MS COCO) …

Content-Based Filtering Recommendation Approach To Label Irish Legal Judgements, Sandesh Gangadhar

Dissertations

Machine learning approaches are applied across several domains to either simplify or automate tasks which directly result in saved time or cost. Text document labelling is one such task that requires immense human knowledge about the domain and efforts to review, understand and label the documents. The company Stare Decisis summarises legal judgements and labels them as they are made available on Irish public legal source www.courts.ie. This research presents a recommendation-based approach to reduce the time for solicitors at Stare Decisis by reducing many numbers of available labels to pick from to a concentrated few that potentially contains the …

Machine Learning Assisted Gait Analysis For The Determination Of Handedness In Able-Bodied People, Hugh Gallagher

Dissertations

This study has investigated the potential application of machine learning for video analysis, with a view to creating a system which can determine a person’s hand laterality (handedness) from the way that they walk (their gait). To this end, the convolutional neural network model VGG16 underwent transfer learning in order to classify videos under two ‘activities’: “walking left-handed” and “walking right-handed”. This saw varying degrees of success across five transfer learning trained models: Everything – the entire dataset; FiftyFifty – the dataset with enough right-handed samples removed to produce a set with parity between activities; Female – only the female …

An Examination Of The Smote And Other Smote-Based Techniques That Use Synthetic Data To Oversample The Minority Class In The Context Of Credit-Card Fraud Classification, Eduardo Parkinson De Castro

Dissertations

This research project seeks to investigate some of the different sampling techniques that generate and use synthetic data to oversample the minority class as a means of handling the imbalanced distribution between non-fraudulent (majority class) and fraudulent (minority class) classes in a credit-card fraud dataset. The purpose of the research project is to assess the effectiveness of these techniques in the context of fraud detection which is a highly imbalanced and cost-sensitive dataset. Machine learning tasks that require learning from datasets that are highly unbalanced have difficulty learning since many of the traditional learning algorithms are not designed to cope …

Transformer Neural Networks For Automated Story Generation, Kemal Araz

Dissertations

Towards the last two-decade Artificial Intelligence (AI) proved its use on tasks such as image recognition, natural language processing, automated driving. As discussed in the Moore’s law the computational power increased rapidly over the few decades (Moore, 1965) and made it possible to use the techniques which were computationally expensive. These techniques include Deep Learning (DL) changed the field of AI and outperformed other models in a lot of fields some of which mentioned above. However, in natural language generation especially for creative tasks that needs the artificial intelligent models to have not only a precise understanding of the given …

Identifying Online Sexual Predators Using Support Vector Machine, Yifan Li

Dissertations

A two-stage classification model is built in the research for online sexual predator identification. The first stage identifies the suspicious conversations that have predator participants. The second stage identifies the predators in suspicious conversations. Support vector machines are used with word and character n-grams, combined with behavioural features of the authors to train the final classifier. The unbalanced dataset is downsampled to test the performance of re-balancing an unbalanced dataset. An age group classification model is also constructed to test the feasibility of extracting the age profile of the authors, which can be used as features for classifier training. The …

Classification Of Animal Sound Using Convolutional Neural Network, Neha Singh

Dissertations

Recently, labeling of acoustic events has emerged as an active topic covering a wide range of applications. High-level semantic inference can be conducted based on main audioeffects to facilitate various content-based applications for analysis, efficient recovery and content management. This paper proposes a flexible Convolutional neural network-based framework for animal audio classification. The work takes inspiration from various deep neural network developed for multimedia classification recently. The model is driven by the ideology of identifying the animal sound in the audio file by forcing the network to pay attention to core audio effect present in the audio to generate Mel-spectrogram. …

Framework For Examination Of Software Quality Characteristics In Conflict: A Security And Usability Exemplar, Bilal Naqvi, Ahmed Seffah, Alain Abran

All Works

© 2020, © 2020 The Author(s). This open access article is distributed under a Creative Commons Attribution (CC-BY) 4.0 license. Standards and best practices for software quality guide on handling each quality characteristic individually, but not when two or more characteristics come into conflict such as security and usability. The objectives of this paper are twofold: (a) to argue on the importance of handling the conflicts between quality characteristics in general; (b) to formulate a framework for conflict examination of the software quality characteristics, we do so while considering the specific case of security and usability. In line with the …

Bodacious-Instance Coverage Mechanism For Wireless Sensor Network, Shahzad Ashraf, Omar Alfandi, Arshad Ahmad, Asad Masood Khattak, Bashir Hayat, Kyong Hoon Kim, Ayaz Ullah

All Works

Copyright © 2020 Shahzad Ashraf et al. Due to unavoidable environmental factors, wireless sensor networks are facing numerous tribulations regarding network coverage. These arose due to the uncouth deployment of the sensor nodes in the wireless coverage area that ultimately degrades the performance and confines the coverage range. In order to enhance the network coverage range, an instance (node) redeployment-based Bodacious-instance Coverage Mechanism (BiCM) is proposed. The proposed mechanism creates new instance positions in the coverage area. It operates in two stages; in the first stage, it locates the intended instance position through the Dissimilitude Enhancement Scheme (DES) and moves …

Designing Shared Control Strategies For Teleoperated Robots Across Intrinsic User Qualities, Nancy Pham

School of Computing: Dissertations, Theses, and Student Research

Accounting for variance in human behavior is an integral part of interacting with robotic systems that share control between users and robots in order to reduce errors, improve performance, and maintain safety. In this work we focus on the shared control of a telepresence robot and how individual user traits may affect a person's performance while navigating the robot. This requires understanding which user qualities impact performance and cause conflicts -- with the ultimate goal of building shared controllers that adapt to those qualities. Toward this goal, we develop novel adaptive shared controllers and integrate the study of intrinsic user …

Pretraining Deep Learning Models For Natural Language Understanding, Han Shao

Honors Papers

Since the first bidirectional deep learn- ing model for natural language understanding, BERT, emerged in 2018, researchers have started to study and use pretrained bidirectional autoencoding or autoregressive models to solve language problems. In this project, I conducted research to fully understand BERT and XLNet and applied their pretrained models to two language tasks: reading comprehension (RACE) and part-of-speech tagging (The Penn Treebank). After experimenting with those released models, I implemented my own version of ELECTRA, a pretrained text encoder as a discriminator instead of a generator to improve compute-efficiency, with BERT as its underlying architecture. To reduce the number …

Courcelle's Theorem: Overview And Applications, Samuel Frederic Barr

Honors Papers

Courcelle's Theorem states that any graph property expressible in monadic second order logic can be decidedin O(f(k)n) for graphs of treewidth k. This paper gives a broad overview of how this theorem is proved and outlines tools available to help express graph properties in monadic second order logic.

Asking Questions Is Easy, Asking Great Questions Is Hard: Constructing Effective Stack Overflow Questions, Jane W. Hsieh

Honors Papers

This paper explores and seeks to improve the ways in which Stack Overflow question posts can elicit answers. Using statistical data analysis approaches and reviews of existing literature, we pin- point three key factors that are found in many previously success- ful/answerable questions. We then present a prototypical sidebar for the ask page that leverages these factors to dynamically (1) evaluate the quality of questions in construction (2) display answer previews of relevant questions and (3) scaffold the identified factors to subsequent askers during their question development processes.

Machine Learning? In My Election? It's More Likely Than You Think: Voting Rules Via Neural Networks, Daniel Firebanks-Quevedo

Honors Papers

Impossibility theorems in social choice have represented a barrier in the creation of universal, non-dictatorial, and non-manipulable voting rules, highlighting a key trade-off between social welfare and strategy-proofness. However, a social planner may be concerned with only a particular preference distribution and wonder whether it is possible to better optimize this trade-off. To address this problem, we propose an end-to-end, machine learning-based framework that creates voting rules according to a social planner's constraints, for any type of preference distribution. After experimenting with rank-based social choice rules, we find that automatically-designed rules are less susceptible to manipulation than most existing rules, …

Towards An Evaluation Visualization With Color, Megan H. Varnum, Kate M.B. Spencer, Alicia M. Grubb

Computer Science: Faculty Publications

Goal models help stakeholders understand project scenarios and make decisions. In prior work, we used Tropos evaluation semantics to allow for automated analysis over time; however, formal evaluation labels (e.g., (F, ⊥)) are difficult for users to interpret across a large model. In this paper, we present our work towards understanding the extent to which using colors in goal modeling affects users’ ability to make decisions. Specifically, we are interested in studying if coloring intentions with evaluation information allows for better comparisons of initial states and simulations of future paths. To address this question, we developed a color visualization extension …

A Preliminary Investigation Of The Utility Of Goal Model Construction, Naomi Cebula, Lily Diao, Alicia M. Grubb

Computer Science: Faculty Publications

Goal models have long been used in the literature to model and reason about stakeholders’ intentions. Prior work proposed several studies aimed at investigating what utility stakeholders derive from constructing and analyzing goal models. We designed and conducted an initial empirical study that explores the construction stage of goal modeling, asking whether stakeholders benefit from manually drawing their own model. We recruited eight qualified participants and asked each to create a goal model for a decision they were considering while talking out loud. Half of the participants in this study used BloomingLeaf, while the remaining participants drew goal models by …

The Artificial University: Decision Support For Universities In The Covid-19 Era, Wesley J. Wildman, Saikou Y. Diallo, George Hodulik, Andrew Page, Andreas Tolk, Neha Gondal

VMASC Publications

Operating universities under pandemic conditions is a complex undertaking. The Artificial University (TAU) responds to this need. TAU is a configurable, open-source computer simulation of a university using a contact network based on publicly available information about university classes, residences, and activities. This study evaluates health outcomes for an array of interventions and testing protocols in an artificial university of 6,500 students, faculty, and staff. Findings suggest that physical distancing and centralized contact tracing are most effective at reducing infections, but there is a tipping point for compliance below which physical distancing is less effective. If student compliance is anything …

Security In 5g-Enabled Internet Of Things Communication: Issues: Challenges, And Future Research Roadmap, Mohammad Wazid, Ashok Kumar Das, Sachin Shetty, Prosanta Gope, Joel J.P.C. Rodrigues

VMASC Publications

5G mobile communication systems promote the mobile network to not only interconnect people, but also interconnect and control the machine and other devices. 5G-enabled Internet of Things (IoT) communication environment supports a wide-variety of applications, such as remote surgery, self-driving car, virtual reality, flying IoT drones, security and surveillance and many more. These applications help and assist the routine works of the community. In such communication environment, all the devices and users communicate through the Internet. Therefore, this communication agonizes from different types of security and privacy issues. It is also vulnerable to different types of possible attacks (for example, …

Ethics, Privacy And Data Collection: A Complex Intersection, Matthew S. Brown

Honors Theses

The technology around us enables incredible abilities such as high-resolution video calls and the ability to stay connected with everyone we care about through social media. This technology also comes with a hidden cost in the form of data collection.

This work explores what privacy means and how users understand what data social media companies collect and monetize. This thesis also proposes a more ethical business model that addresses privacy concerns from an individual perspective.

Score Following With Hidden Tempo Using A Switching State-Space Model, Yucong Jiang, Chris Raphael

Department of Math & Statistics Faculty Publications

A score-following program traces the notes in a musical score during a performance. This capability is essential to many meaningful applications that synchronize audio with a score in an on-line fashion. Existing algorithms often stumble on certain difficult cases, one of which is piano music. This paper presents a new method to tackle such cases. The method treats tempo as a variable rather than a constant (with constraints), allowing the program to adapt to live performance variations. This is first expressed by a Kalman filter model at the note level, and then by an almost equivalent switching state-space model at …

Toward Multi-Label Sentiment Analysis: A Transfer Learning Based Approach, Jie Tao, Xing Fang

Faculty Publications - Information Technology

Sentiment analysis is recognized as one of the most important sub-areas in Natural Language Processing (NLP) research, where understanding implicit or explicit sentiments expressed in social media contents is valuable to customers, business owners, and other stakeholders. Researchers have recognized that the generic sentiments extracted from the textual contents are inadequate, thus, Aspect Based Sentiment Analysis (ABSA) was coined to capture aspect sentiments expressed toward specific review aspects. Existing ABSA methods not only treat the analytical problem as single-label classification that requires a fairly large amount of labelled data for model training purposes, but also underestimate the entity aspects …

A Single-Source Shortest Path Algorithm For Dynamic Graphs, Muteb Alshammari, Abdelmounaam Rezgui

Faculty Publications - Information Technology

Graphs are mathematical structures used in many applications. In recent years, many applications emerged that require the processing of large dynamic graphs where the graph’s structure and properties change constantly over time. Examples include social networks, communication networks, transportation networks, etc. One of the most challenging problems in large scale dynamic graphs is the single-source shortest path (SSSP) problem. Traditional solutions (based on Dijkstra’s algorithms) to the SSSP problem do not scale to large dynamic graphs with a high change frequency. In this paper, we propose an efficient SSSP algorithm for large dynamic graphs. We first present our algorithm and …

Landing Throttleable Hybrid Rockets With Hierarchical Reinforcement Learning In A Simulated Environment, Francesco Alessandro Stefano Mikulis-Borsoi

Honors Theses and Capstones

In this paper, I develop a hierarchical Markov Decision Process (MDP) structure for completing the task of vertical rocket landing. I start by covering the background of this problem, and formally defining its constraints. In order to reduce mistakes while formulating different MDPs, I define and develop the criteria for a standardized MDP definition format. I then decompose the problem into several sub-problems of vertical landing, namely velocity control and vertical stability control. By exploiting MDP coupling and symmetrical properties, I am able to significantly reduce the size of the state space compared to a unified MDP formulation. This paper …

Automatic Target Recognition With Deep Metric Learning., Abdelhamid Bouzid

Electronic Theses and Dissertations

An Automatic Target Recognizer (ATR) is a real or near-real time understanding system where its input (images, signals) are obtained from sensors and its output is the detected and recognized target. ATR is an important task in many civilian and military computer vision applications. The used sensors, such as infrared (IR) imagery, enlarge our knowledge of the surrounding environment, especially at night as they provide continuous surveillance. However, ATR based on IR faces major challenges such as meteorological conditions, scale and viewpoint invariance. In this thesis, we propose solutions that are based on Deep Metric Learning (DML). DML is a …

3d Localization Of Defects In Facility Inspections, Nicholas Califano

Electronic Theses and Dissertations, 2020-2023

Wind tunnels are crucial facilities that support the aerospace industry. However, these facilities are large, complex, and pose unique maintenance and inspection requirements. Manual inspections to identify defects such as cracks, missing fasteners, leaks, and foreign objects are important but labor and schedule intensive. The goal of this thesis is to utilize small Unmanned Aircraft Systems with onboard cameras and computer vision-based analysis to automate the inspection of the interior and exterior of NASA's critical wind tunnel facilities. Missing fasteners are detected as the defect class, and existing fasteners are detected to provide potential future missing fastener sites for preventative …