Computer Engineering Commons^™

Sequential Memory Generation For Cognitive Models, Eben Miles Sherwood

Master's Theses

Understanding the process of memory formation in neural systems is of great interest in the field of neuroscience. Valiant’s Neuroidal Model poses a plausible theory for how memories are created within a computational context. Previously, the algorithm JOIN has been used to show how the brain could perform conjunctive and disjunctive coding to store memories. A limitation of JOIN is that it does not consider the coding of temporal information in a meaningful manner. We propose SeqMem, a similar algorithmic primitive that is designed to encode a series of items within a random graph model. We investigate the feasibility of …

Optimal False Data Injection (Fdi) In Simulated Cooperative Adaptive Cruise Control (Cacc) Systems, Lovro Dukic

Master's Theses

In the rapidly advancing field of autonomous vehicles, ensuring the security and reliability of self-driving systems is crucial. Autonomous vehicle systems, such as cooperative adaptive cruise control (CACC), must undergo significant research and testing before their integration into commercial intelligent transportation systems. CACC considers multiple vehicles in close proximity as a single entity, or platoon, with each vehicle equipped with a controller that uses sensor-based measurements and vehicle-to-vehicle (V2V) communication to control inter-vehicle spacing. While this system offers numerous potential benefits for traffic safety and efficiency, it is also susceptible to False Data Injection (FDI) attacks, which can cause the …

Anomaly Detection In Heterogeneous Iot Systems: Leveraging Symbolic Encoding Of Performance Metrics For Anomaly Classification, Maanav Patel

Master's Theses

Anomaly detection in Internet of Things (IoT) systems has become an increasingly popular field of research as the number of IoT devices proliferate year over year. Recent research often relies on machine learning algorithms to classify sensor readings directly. However, this approach leads to solutions being non-portable and unable to be applied to varying IoT platform infrastructure, as they are trained with sensor data specific to one configuration. Moreover, sensors generate varying amounts of non-standard data which complicates model training and limits generalization. This research focuses on addressing these problems in three ways a) the creation of an IoT Testbed …

A Federation Of Sentries: Secure And Efficient Trusted Hardware Element Communication, Blake A. Ward

Master's Theses

Previous work introduced TrustGuard, a design for a containment architecture that allows only the result of the correct execution of approved software to be outputted. A containment architecture prevents results from malicious hardware or software from being communicated externally. At the core of TrustGuard is a trusted, pluggable device that sits on the path between an untrusted processor and the outside world. This device, called the Sentry, is responsible for validating the correctness of all communication before it leaves the system. This thesis seeks to leverage the correctness guarantees that the Sentry provides to enable efficient secure communication between two …

A Study On Privacy Over Security And Privacy Enhancing Networks, Everett Lee Conway

Master's Theses

With rapid developments in communication technologies and awareness of security and privacy risks online, Security and Privacy Enhancing Networks (SPENs) have become increasingly popular. Especially during the COVID-19 pandemic, workplaces encouraged employees to take additional security measures, such as VPNs. In this work, we conduct a comprehensive study on website fingerprinting attacks. A comprehensive system model and threat model based on two types of SPENs (Virtual Private Networks and Tor Networks) are presented. Moreover, we demonstrate a website fingerprinting attack by ethically collecting website fetch data and analyzing the collected data using five different machine learning classification models including k …

Generalized Model To Enable Zero-Shot Imitation Learning For Versatile Robots, Yongshuai Wu

Master's Theses

The rapid advancement in Deep Learning (DL), especially in Reinforcement Learning (RL) and Imitation Learning (IL), has positioned it as a promising approach for a multitude of autonomous robotic systems. However, the current methodologies are predominantly constrained to singular setups, necessitating substantial data and extensive training periods. Moreover, these methods have exhibited suboptimal performance in tasks requiring long-horizontal maneuvers, such as Radio Frequency Identification (RFID) inventory, where a robot requires thousands of steps to complete.

In this thesis, we address the aforementioned challenges by presenting the Cross-modal Reasoning Model (CMRM), a novel zero-shot Imitation Learning policy, to tackle long-horizontal robotic …

Evaluating The Effect Of Noise On Secure Quantum Networks, Karthick Anbalagan

Master's Theses

This thesis focuses on examining the resilience of secure quantum networks to environmental noise. Specifically, we evaluate the effectiveness of two well-known quantum key distribution (QKD) protocols: the Coherent One-Way (COW) protocol and Kak’s Three-Stage protocol (Kak06). The thesis systematically evaluates these protocols in terms of their efficiency, operational feasibility, and resistance to noise, thereby contributing to the progress of secure quantum communications. Using simulations, this study evaluates the protocols in realistic scenarios that include factors such as noise and decoherence. The results illustrate each protocol’s relative benefits and limitations, highlighting the three-stage protocol’s superior security characteristics, resistance to interference, …

Diegetic Sonification For Low Vision Gamers, Jhané Dawes

Master's Theses

There are not many games designed for all players that provide accommodations for low vision users. This means that low vision users may not get to engage with the gaming community in the same way as their sighted peers. In this thesis, I explore how diegetic sonification can be used as a tool to support these low vision gamers in the typical gaming environment. I asked low vision players to engage with a prototype game level with two diegetic sonification techniques applied, without the use of their corrective lenses. I found that participants had more enjoyment and experienced less difficulty …

Deep Learning Using Vision And Lidar For Global Robot Localization, Brett E. Gowling

Master's Theses

As the field of mobile robotics rapidly expands, precise understanding of a robot’s position and orientation becomes critical for autonomous navigation and efficient task performance. In this thesis, we present a snapshot-based global localization machine learning model for a mobile robot, the e-puck, in a simulated environment. Our model uses multimodal data to predict both position and orientation using the robot’s on-board cameras and LiDAR sensor. In an effort to minimize localization error, we explore different sensor configurations by varying the number of cameras and LiDAR layers used. Additionally, we investigate the performance benefits of different multimodal fusion strategies while …

Building Software At Scale: Understanding Productivity As A Product Of Software Engineering Intrinsic Factors, Gauthier Ingende Wa Boway

Master's Theses

During our education at KSU, we have learned about various factors that affect productivity such as schedule, budget, and risks, but those are often controlled outside of what we could learn as software engineering principles, patterns, or practices. On top of that, other off-work factors such as health conditions, emotional distress, or political climate, just to name a few, could drastically affect the productivity of a software engineering team. We see a demarcation between those factors that affect productivity in software engineering but are not inherent to the discipline itself, which we call resistance factors, and the factors that are …

A Study Of Random Partitions Vs. Patient-Based Partitions In Breast Cancer Tumor Detection Using Convolutional Neural Networks, Joshua N. Ramos

Master's Theses

Breast cancer is one of the deadliest cancers for women. In the US, 1 in 8 women will be diagnosed with breast cancer within their lifetimes. Detection and diagnosis play an important role in saving lives. To this end, many classifiers with varying structures have been designed to classify breast cancer histopathological images. However, randomly partitioning data, like many previous works have done, can lead to artificially inflated accuracies and classifiers that do not generalize. Data leakage occurs when researchers assume that every image in a dataset is independent of each other, which is often not the case for medical …

Insights Into Cellular Evolution: Temporal Deep Learning Models And Analysis For Cell Image Classification, Xinran Zhao

Master's Theses

Understanding the temporal evolution of cells poses a significant challenge in developmental biology. This study embarks on a comparative analysis of various machine-learning techniques to classify cell colony images across different timestamps, thereby aiming to capture dynamic transitions of cellular states. By performing Transfer Learning with state-of-the-art classification networks, we achieve high accuracy in categorizing single-timestamp images. Furthermore, this research introduces the integration of temporal models, notably LSTM (Long Short Term Memory Network), R-Transformer (Recurrent Neural Network enhanced Transformer) and ViViT (Video Vision Transformer), to undertake this classification task to verify the effectiveness of incorporating temporal features into the classification …

Data-Driven Control Of Acoustic Waves Using Movable And Flexible Scatterers, Noam Smilovich

Master's Theses

Partial Differential Equations (PDEs) serve as fundamental tools in scientific and engineering disciplines, modeling phenomena ranging from material design to climate dynamics. Developing robotic systems capable of controlling PDE-governed systems, particularly when these phenomena are only partially observable, has the potential to drive significant technological advancements. This work presents a framework that leverages physics-informed machine learning (ML) for the control of PDEs. At the core of this approach is an agent equipped with sensors that generate a low-dimensional, physics-informed representation of the environment, enabling the derivation of optimal sparse control policies for the agent’s actuators. The focus is on manipulating …

Unrealvision: A Synthetic Dataset Generator For Human-Pose Estimation And Behavior Analysis, Thinh Lu

Master's Theses

For over a decade, computer vision (CV) has become an indispensable component of numerous camera surveillance applications as well as intelligent autonomous systems. Thanks to new advances in AI, Edge Computing, and IoT technologies, there is now a rapidly growing number of smart camera devices, industrial and consumer robots that are using computer vision for various applications - from human tracking and analysis, object classification, to visual inspection and anomaly detection. For most common use cases, building vision-based applications can be a straightforward and affordable task thanks to the increasing number of publicly accessible datasets and research publications. However, it …

Uncovering Weaknesses In Autonomous Driving: A Formal Approach To Adversarial Scenario Generation And Falsification, Carlos O. Hernandez

Master's Theses

Autonomous vehicles utilize advanced safety features like proactive driving assistance and pre-collision alerts to minimize the risk of accidents. However, evaluating the correct functionality of these systems is complex. First, safety systems are highly sophisticated, integrating software, networking, and hardware components, many of which rely on advanced artificial intelligence and machine learning algorithms. Second, an array of dynamic factors, including numerous actors and physical variables, can influence the performance of safety mechanisms during critical scenarios. Each actor’s unique behavior introduces unpredictability, making it difficult to anticipate future states and outcomes. This thesis presents a comprehensive framework for testing autonomous vehicle …

Brunet: Disruption-Tolerant Tcp And Decentralized Wi-Fi For Small Systems Of Vehicles, Nicholas Brunet

Master's Theses

Reliable wireless communication is essential for small systems of vehicles. However, for small-scale robotics projects where communication is not the primary goal, programmers frequently choose to use TCP with Wi-Fi because of their familiarity with the sockets API and the widespread availability of Wi-Fi hardware. However, neither of these technologies are suitable in their default configurations for highly mobile vehicles that experience frequent, extended disruptions. BRUNET (BRUNET Really Useful NETwork) provides a two-tier software solution that enhances the communication capabilities for Linux-based systems. An ad-hoc Wi-Fi network permits decentralized peer-to-peer and multi-hop connectivity without the need for dedicated network infrastructure. …

Decentralized Machine Learning On Blockchain: Developing A Federated Learning Based System, Nikhil Sridhar

Master's Theses

Traditional Machine Learning (ML) methods usually rely on a central server to per-
form ML tasks. However, these methods have problems like security risks, data
storage issues, and high computational demands. Federated Learning (FL), on the
other hand, spreads out the ML process. It trains models on local devices and then
combines them centrally. While FL improves computing and customization, it still
faces the same challenges as centralized ML in security and data storage.

This thesis introduces a new approach combining Federated Learning and Decen-
tralized Machine Learning (DML), which operates on an Ethereum Virtual Machine
(EVM) compatible blockchain. The …

A Study On Rapidly Exploring Random Tree Algorithms For Robot Path Planning, Sahil Sharma

Master's Theses

Robot path planning is a critical feature of autonomous systems. Rapidly-exploring Random Trees (RRT) is a path planning technique that randomly samples the robot configuration space to find a path between the start and end point. This thesis studies and compares the performance of four important RRT algorithms, namely, the original RRT, the optimal RRT (also termed RRT*), RRT*-Smart, and Informed RRT* for six different environments. The performance measures include the final path length (which is also the shortest path length found by each algorithm), time to find the first path, run time (of 1000 iterations) for each algorithm, total …

Characterization And Estimation Of Musculoskeletal Pain Using Machine Learning, Boluwatife Faremi

Master's Theses

Traditional scales utilized for recording pain are known to be highly subjective and biased due to inaccuracies in recollecting actual pain intensities. As a result, machine learning (ML) models that are trained using these scores as ground truth are reported to have low performance for objective pain classification because of the huge disparity between what was felt in moments of pain and the scores recorded afterward.

In the present study, two devices were designed for gathering real-time, continuous in-session subjective pain scores and the recording of the autonomic nervous system (ANS) altered endodermal (EDA) activity. 24 participants were recruited to …

Contextually Dynamic Quest Generation Using In-Session Player Information In Mmorpg, Shangwei Lin

Master's Theses

Massively multiplayer online role-playing games (MMORPGs) are one of the most

popular genres in video games that combine massively multiplayer online genres with

role-playing gameplay. MMORPGs’ featured social interaction and forms of level pro-

gression through quest completion are the core for gaining players’ attention. Varied

and challenging quests play an essential part in retaining that attention. However,

well-crafted content takes much longer to develop with human efforts than it does to

consume, and the dominant procedural content generation models for quests suffer

from the drawback of being incompatible with dynamic world changes and the feeling

of repetition over time. …

Effects Of Concussion And Visuomotor Metrics On Nhl Performance: An Explainable Ai Approach, Michael T. Moschitto

Master's Theses

Cognitive motor integration (CMI), the simultaneous coordination between cerebral function and motor output, is known to deteriorate following a mild traumatic brain injury (mTBI). This thesis explores the relationship between mTBI, CMI, and the performance of elite athletes in the National Hockey League (NHL). The approach focuses on examining the predictive value of various supervised Machine Learning (ML) models with an emphasis on Explainable Artificial Intelligence (XAI) models. Since the ML solution is intended to complement human scouting decisions, we evaluate the experiments based on both interpretability and accuracy on a limited class imbalanced dataset. The contributions of this research …

Neural Compression Inference Accelerator: A Cost And Energy-Effective Alternative To Conventional Machine Learning Inference Methods, Matthew Wallace

Master's Theses

Recent developments in machine learning and artificial intelligence have sparked an influx of workloads that require specialized computer hardware for cloud services. The hardware running machine learning models predominantly consists of graphics processing units (GPUs) and tensor processing units (TPUs). However, these com- ponents are expensive for cloud services to purchase, costly for customers to rent, prone to price spikes, and energy-intensive. In this research we show that both cloud services and customers would benefit from utilizing field programmable gate arrays (FPGAs) to alleviate the aforementioned challenges. An FPGA can be configured as a machine learning accelerator, operating similarly to …

Assessing The Resilience Of Mycorrhizal Networks Following Central Tree Removal, Deon Lillo

Master's Theses

Mycorrhizal networks (MNs), or the networks of fungal mycelia that connect plants to each other, are vital in contributing to the well-being of ecosystems. They not only assist in the transport of nutrients across an ecosystem, but also help protect an ecosystem from disease and adverse conditions. However, more research into these networks is needed and modelling these networks as graphs can help us achieve this. By applying centrality analysis and performing k-core partitioning on these networks, we are able to identify the trees that are most important and central to a MN and observe the effects of removing these …

Predicting Suicide Risk Among Youths Using Machine Learning Methods, Saswati Bhattacharjee

Master's Theses

Suicide is the second leading cause of death among youths in the USA. Although machine learning approaches have provided great potential for predicting suicide risk using survey data, prediction accuracy may not meet the need for clinical diagnosis due to the intrinsic characteristics of datasets. In this study, I perform a comparative study of six classification algorithms including naïve Bayes (NB), logistic regression (LR), multilayer perceptron (MLP), AdaBoost (Ada), random forest (RF), and bagging using YRBSS dataset and investigate the effectiveness of several data handling techniques to improve the overall performance of suicide risk prediction.

The dataset consists of 76 …

Shelfaware: Accelerating Collaborative Awareness With Shelf Crdt, John C. Waidhofer

Master's Theses

Collaboration has become a key feature of modern software, allowing teams to work together effectively in real-time while in different locations. In order for a user to communicate their intention to several distributed peers, computing devices must exchange high-frequency updates with transient metadata like mouse position, text range highlights, and temporary comments. Current peer-to-peer awareness solutions have high time and space complexity due to the ever-expanding logs that each client must maintain in order to ensure robust collaboration in eventually consistent environments. This paper proposes an awareness Conflict-Free Replicated Data Type (CRDT) library that provides the tooling to support an …

Racketframes: A Dataframe Implementation For The Racket Programming Language, Shubham Kahal

Master's Theses

The DataFrame is a powerful table-like data structure used frequently in Data Science, the in-demand and innovative field focused on the extraction of valuable insights from data. Typically, datasets are not perfect upon collection and need to be prepared so that the resulting dataset is useful for statistical analysis. A DataFrame API supports optimized methods such as, selecting, aggregating and filtering rows, columns, and cells as well as renaming row and column labels. It also supports methods for normalizing data, merging data, adding new columns and labelling missing data among numerous other features. An API to work with tabular data …

Analysis And Usage Of Natural Language Features In Success Prediction Of Legislative Testimonies, Marine Cossoul

Master's Theses

Committee meetings are a fundamental part of the legislative process in which
constituents, lobbyists, and legislators alike can speak on proposed bills at the
local and state level. Oftentimes, unspoken “rules” or standards are at play in
political processes that can influence the trajectory of a bill, leaving constituents
without a political background at an inherent disadvantage when engaging with
the legislative process. The work done in this thesis aims to explore the extent to
which the language and phraseology of a general public testimony can influence a
vote, and examine how this information can be used to promote civic …

Deep Learning In Ai Medical Imaging For Stroke Diagnosis, James Mario Guzman

Master's Theses

Enhancing medical imaging stroke diagnosis applications with artificial intelligence (AI) tools to determine lesion volume, location and clinical metadata is vital toward guiding patient treatment and procedure. A major hardship in developing stroke diagnosis AI tools is the scarcity of publicly available clinical 3D stroke datasets. Through working with Johns Hopkins University, University of Michigan’s ICPSR data repository and SJSU research, we gained access to potentially the largest 3D MRI stroke dataset with clinical metadata annotated by neuroradiologists known as ICPSR 38464. With the ICPSR 38464 dataset recently being available through institutional review board (IRB) approval or exemption, we were …

Group-Invariant Reinforcement Learning, Fnu Ankur

Master's Theses

Our work introduces a way to learn an optimal reinforcement learning agent accompanied by intrinsic properties of the environment. The extracted properties helps the agent to extrapolate the learning to unseen states efficiently. Out of all the various types of properties, we are intrigued towards equivariant and invariant properties, which essentially translates to symmetry. Contrary to many approaches, we do not assume the symmetry, rather learn them, making the approach agnostic to the environment and the property. The learned properties offers multiple perspective of the environment to exploit it to benefit decision making while interacting with the environment. By building …

Automatic Presentation Slide Generation Using Llms, Tanya Gupta

Master's Theses

Presentation slides are widely used for conveying information in academic and professional contexts. However, manual slide creation can be time-consuming. Our research focuses on automated slide generation, specifically for scientific research papers. Automating the creation of presentation slides for scientific documents is a rather novel task and hence, there’s limited training data available and there also exists the token constraints of language models like BERT, with a maximum sequence length of 512 tokens. In this study, we fine-tune large language models, including Longformer-Encoder-Decoder (supporting sequences up to 16,834 tokens) and BIGBIRD-Pegasus (supporting sequences up to 4,096 tokens). We tackle this …