Engineering Commons^™

Enhancing Robustness Of Graph Neural Network Against Adversarial Attacks By Balancing Local And Global Perspectives, Bibek Raj Joshi

Browse all Theses and Dissertations

Graph Neural Networks (GNNs) have increasingly gained popularity as tools for analyzing graph data in areas like biology, knowledge-graphs, social networks, biology, and recommendation systems. However, their vulnerability to adversarial attacks - small, targeted manipulations of graph structures or node features - raises serious concerns about their reliability in real-world applications. Existing defense strategies, such as adversarial training, edge filtering, low-rank approximations, and randomization-based methods, often suffer from high computational costs, scalability issues, or reduced clean-data performance. Unlike these methods, the proposed approach integrates multi-hop relationships, applies adaptive regularization, and maintains a balance between feature-based and structural embeddings, ensuring improved …

Meta-Learning-Based Model Stacking Framework For Hardware Trojan Detection In Fpga Systems, Mani Rupak Gurram

Browse all Theses and Dissertations

In today's technological landscape, hardware devices are integral to critical applications such as industrial automation, autonomous vehicles, and medical equipment, relying on advanced platforms like FPGAs for core functionalities. However, the multi-stage manufacturing process, often distributed across various foundries, introduces substantial security risks, notably the potential for hardware Trojan insertion. These malicious modifications compromise the reliability and safety of hardware systems. This research addresses the detection of hardware Trojans through side-channel analysis, utilizing power and electromagnetic signal data, combined with meta-learning techniques, specifically model stacking. By employing diverse base models and a meta-model to consolidate predictions, this non-invasive approach effectively …

Secure Similar Patients Query With Homomorphically Evaluated Thresholds, Mounika Pratapa, Aleksander Essex

Electrical and Computer Engineering Publications

Patient-centric precision medicine requires the analysis of large volumes of genomic data to tailor treatments and medications based on individual-level characteristics. Because the amount of data held by a single institution is limited, researchers may want access to genomic data held by other institutions. Owing to the inherent privacy implications of genomic data, performing comparisons on encrypted data is preferable in certain settings. The Similar patient query (SPQ) is an application that enables a secure search across genomic databases for patients with similar genetic makeup. Query results can be used to draw meaningful conclusions regarding suitable therapies.

However, existing protocols …

Deep Transfer Learning For Detection Of Upper And Lower Body Movements: Transformer With Convolutional Neural Network, Kyle Lacroix, Davoud Gholamiangonabadi, Ana Luisa Trejos, Katarina Grolinger

Electrical and Computer Engineering Publications

When humans repeat the same motion, the tendons, muscles, and nerves can be damaged, causing Repetitive Stress Injuries (RSI). If the repetitive motions that lead to RSI are recognized early, actions can be taken to prevent these injuries. As Human Activity Recognition (HAR) aims to identify activities employing wearable or environment sensors, HAR is the first step toward identifying repetitive motions. Deep learning models, such as Convolutional Neural Networks (CNNs), have seen great success in recognizing activities for participants whose data are used in the model training; however, their accuracy drops for new participants as people move in different ways. …

Federated Learning For Sentiment Analysis In Presence Of Non-Iid Data: Sensitivity Of Deep Learning Models, Davoud Gholamiangonabadi, Katarina Grolinger

Electrical and Computer Engineering Publications

In sentiment analysis, data are commonly distributed across many devices, and traditional machine learning requires transferring these data to a central location exposing data to security and privacy risks. Federated Learning (FL) avoids this transfer by training a model without requiring the clients/devices to share their local data; however, FL performance drops when data are not Independent and Identically Distributed (non-IID), such as when label distribution or data size vary across clients. Although techniques for non-IID data have been proposed primarily in the image domain, the sensitivity of various deep learning models to non-IID data needs to be examined. Consequently, …

Low-Cost Open-Source Melt Flow Index System For Distributed Recycling And Additive Manufacturing, Dawei Liu, Aditi Basdeo, Catalina Suescun Gonzalez, Alessia Romani, Hakim Boudaoud, Cécile Nouvel, Fabio A. Cruz Sanchez, Joshua M. Pearce

Electrical and Computer Engineering Publications

The increasing adoption of distributed recycling via additive manufacturing (DRAM) has facilitated the revalorization of materials derived from waste streams for additive manufacturing. Recycled materials frequently contain impurities and mixed polymers, which can degrade their properties over multiple cycles. This degradation, particularly in rheological properties, limits their applicability in 3D printing. Consequently, there is a critical need for a tool that enables the rapid assessment of the flowability of these recycled materials. This study presents the design, development, and manufacturing of an open-source melt flow index (MFI) apparatus. The open-source MFI was validated with tests on virgin polylactic acid pellets, …

การจัดลำดับความสำคัญของบทวิจารณ์ของผู้ใช้ซอฟต์แวร์โดยคำนึงถึงประสบการณ์ผู้ใช้โดยใช้การเรียนรู้ของเครื่อง, ลักษณ์สิปาง สาครวิจิตรี

Chulalongkorn University Theses and Dissertations (Chula ETD)

ปัจจุบันตลาดโมไบล์แอปพลิเคชันมีการแข่งขันสูง นักพัฒนาจึงจำเป็นต้องให้ความสำคัญกับบทวิจารณ์ของผู้ใช้ซึ่งบ่งบอกถึงความคิดเห็นและประสบการณ์ของผู้ใช้งานจริง เพื่อนำไปสู่การปรับปรุงและพัฒนาฟังก์ชันการทำงานให้ตรงตามความต้องการของผู้ใช้มากยิ่งขึ้น อย่างไรก็ตาม แอปพลิเคชันยอดนิยมมักมีบทวิจารณ์จำนวนมาก ทำให้เกิดข้อจำกัดในการที่นักพัฒนาจะสามารถอ่านและวิเคราะห์บทวิจารณ์ทั้งหมดได้อย่างมีประสิทธิภาพ เพื่อจัดการกับปัญหาดังกล่าว งานวิจัยฉบับนี้จึงนำเทคโนโลยีการเรียนรู้ของเครื่องเข้ามาช่วยในการจำแนกประเภทของบทวิจารณ์ โดยพิจารณาจากเนื้อหาในบทวิจารณ์ของผู้ใช้ซึ่งอาจสะท้อนถึงปัญหา ข้อบกพร่อง หรือข้อเสนอในการพัฒนาฟังก์ชันใหม่ ข้อมูลที่ได้จะถูกนำไปใช้จัดลำดับความสำคัญของการแก้ไขปรับปรุงตามผลกระทบที่มีต่อประสบการณ์ของผู้ใช้ งานวิจัยนี้ได้กำหนดหมวดหมู่ของบทวิจารณ์ไว้ทั้งหมด 5 ประเภท ได้แก่ ปัญหาด้านความจำเป็นพื้นฐาน, ปัญหาด้านการปฏิบัติ, ปัญหาด้านความเพลิดเพลิน, ปัญหาด้านความแปลกใหม่ และปัญหาอื่น ๆ ซึ่งสะท้อนลำดับความสำคัญของปัญหาจากมุมมองของผู้ใช้ ในการพัฒนาโมเดลการเรียนรู้ของเครื่องพบว่าโมเดลเบิร์ตมีประสิทธิภาพสูงกว่าโมเดลเอสวีเอ็ม แรนดอมฟอเรสต์ และโลจิสติกรีเกรสชัน โดยมีค่าความเที่ยงเป็น 0.72 ค่าเรียกกลับเป็น 0.719 ค่าเอฟวันเป็น 0.719 และค่าความแม่นเป็น 0.722 นอกจากนี้ งานวิจัยยังได้พัฒนาเว็บแอปพลิเคชันต้นแบบที่ใช้โมเดลเบิร์ตที่สร้างขึ้น เพื่อช่วยให้นักพัฒนาสามารถนำไปใช้วิเคราะห์และจัดลำดับความสำคัญของงานการบำรุงรักษาโมไบล์แอปพลิเคชัน โดยเว็บแอปพลิเคชันจะช่วยลดภาระในการอ่านบทวิจารณ์จำนวนมาก พร้อมทั้งสกัดข้อมูลที่สำคัญออกมาให้เห็นภาพรวมของความต้องการของผู้ใช้

Case Studies Of Restapi And Graphql Architectures, Janni Daniel Balraj, Gary Clynch

Academic Poster Collection

The adoption of microservices architecture has experienced significant growth, driven by its appeal in terms of modularity, scalability, and deployment ease. However, this proliferation of microservices has intensified the demand for efficient data exchange among them. While REST API has long been the standard protocol for microservices communication, its limitations in flexibility and performance have spurred the ascent of GraphQL as a more efficient alternative, as highlighted by studies comparing their performance. As the number of microservices continues to rise, traditional methods like REST APIs prove challenging, leading to issues such as over-fetching or under-fetching of data. GraphQL addresses these …

The Role Of Ai, Big Data And Predictive Analytics In Mitigating Unemployment Insurance Fraud, Siddikur Rahman, Md Abu Sayem, Shariar Emon Alve, Md Shahidul Islam, Muhammad Mahmudul Islam, Arifa Ahmed, Mohammed Kamruzzaman

Finance, Economics, and Data Analytics

The fraudulent claims for Unemployment Insurance (UI) have also risen massively in the United States especially during the onset of COVID-19 pandemic with billions of dollars that were lost. These approaches applied formerly in fraud detection and prevention have been challenged by new and advanced fraud systems. For this reason, AI, Big Data and Predictive Analytics are now crucial for improving fraud mitigation in UI programs. The aim of this research is to understand how far AI, Big Data and Predictive Analytics have been utilized, for how effective they are and the barriers they pose in tackling unemployment insurance fraud …

Ai And Workforce Development: A Comparative Analysis Of Skill Gaps And Training Needs In Emerging Economies, Gursahildeep Singh Sidhu, Md Abu Sayem, Nazifa Taslima, Ahmed Selim Anwar, Fariba Chowdhury, Manataka Rowshon

Finance, Economics, and Data Analytics

AI is developing quickly and offers considerable prospects for economic development; nevertheless, it presents crucial challenges for instance, skills development and its effect on employee's jobs in developing economies. This research will seek to establish the current state of AI related skills deficits and training in these economies to establish how governments and organizations can close the gaps by preparing the workforce for the future AI revolution. In line with the research questions, the following emerges as the research problem: To what extent have emerging market companies embraced AI? What key skills are scarce both in the internal and external …

Constructing An Interpretable Deep Learning Framework Utilizing Variational Autoencoder Latent Space For Part-Prototype Learning, Shiska Raut

Computer Science and Engineering Theses - Archive

What visual attributes do cats have in common, and what features set them apart from dogs? How are we able to tell the difference between the two? While we do not fully understand the mechanism humans use for object detection, one popular theory suggests that it boils down to identifying distinct visual features specific to each object. For example, all cats have vertical slit-shaped pupils when their eyes are constricted, which is something we do not see in dogs. These slit-shaped pupils are a feature ‘prototypical’ to cats. Object classification is a computer vision task that involves identifying and categorizing …

Nonlinear Guidance And Control Of Unmanned Aerial Manipulators For Delivering A Payload On A Moving Platform, Ravi Gyawali

Mechanical and Aerospace Engineering Theses - Archive

Unmanned Aerial manipulators (UAMs) are a class of Unmanned Aerial Vehicles (UAVs) equipped with a manipulator. By combining the aerial mobility of a UAV with a manipulator's dexterity, these hybrid systems can perform a wide range of complex tasks while reducing risks and costs. As a result, they are increasingly being utilized for military, industrial, and agricultural applications.

The thesis presents a novel approach for a multi-UAM system to collaboratively deliver a payload on a stationary or maneuvering platform. A sliding-mode-based guidance law, sourced from existing literature, is integrated with a combined control technique for the UAVs and their respective …

Adaptive Beyond Von-Neumann Computing Devices And Reconfigurable Architectures For Edge Computing Applications, Mousam Hossain

Graduate Thesis and Dissertation 2023-2024

The Von-Neumann bottleneck, a major challenge in computer architecture, results from significant data transfer delays between the processor and main memory. Crossbar arrays utilizing spin-based devices like Magnetoresistive Random Access Memory (MRAM) aim to overcome this bottleneck by offering advantages in area and performance, particularly for tasks requiring linear transformations. These arrays enable single-cycle and in-memory vector-matrix multiplication, reducing overheads, which is crucial for energy and area-constrained Internet of Things (IoT) sensors and embedded devices.

This dissertation focuses on designing, implementing, and evaluating reconfigurable computation platforms that leverage MRAM-based crossbar arrays and analog computation to support deep learning and error …

Addressing Challenges In Utilizing Gpus For Accelerating Privacy-Preserving Computation, Ardhi Wiratama Baskara Yudha

Graduate Thesis and Dissertation 2023-2024

Cloud computing increasingly handles confidential data, like private inference and query databases. Two strategies are used for secure computation: (1) employing CPU Trusted Execution Environments (TEEs) like AMD SEV, Intel SGX, or ARM TrustZone, and (2) utilizing emerging cryptographic methods like Fully Homomorphic Encryption (FHE) with libraries such as HElib, Microsoft SEAL, and PALISADE. To enhance computation, GPUs are often employed. However, using GPUs to accelerate secure computation introduces challenges addressed in three works.

In the first work, we tackle GPU acceleration for secure computation with CPU TEEs. While TEEs perform computations on confidential data, extending their capabilities to GPUs …

Internet-Of-Things Privacy In Wifi Networks: Side-Channel Leakage And Mitigations, Mnassar Alyami

Graduate Thesis and Dissertation 2023-2024

WiFi networks are susceptible to statistical traffic analysis attacks. Despite encryption, the metadata of encrypted traffic, such as packet inter-arrival time and size, remains visible. This visibility allows potential eavesdroppers to infer private information in the Internet of Things (IoT) environment. For example, it allows for the identification of sleep monitors and the inference of whether a user is awake or asleep.

WiFi eavesdropping theoretically enables the identification of IoT devices without the need to join the victim's network. This attack scenario is more realistic and much harder to defend against, thus posing a real threat to user privacy. However, …

Towards Performance Guarantee For Federated Computing, Stoddard A. Rosenkrantz

Computer Science and Engineering Dissertations - Archive

In federated computing environments, where multiple independent devices collaborate without centralized resource control, ensuring reliable and predictable performance is crucial for meeting Service Level Objectives (SLOs). This dissertation presents a novel framework for achieving SLOs by utilizing measured subtask response times, which are obtained and processed locally on the client devices. Each device generates a performance curve based on its subtask response times, and only the essential parameters of this curve are transmitted to a centralized client selector. The client selector then uses these parameters to determine which devices are best suited to meet predefined SLOs for specific tasks, ensuring …

A Few-Shot Multi-Modality Traversability Segmentation Framework For Indoor Robotic Navigation, Qiyuan An

Computer Science and Engineering Dissertations - Archive

Traversability in autonomous robotic navigation refers to the ability of an autonomous agent to safely navigate over a given terrain. It plays a critical role in enabling robots to navigate over unseen or unknown terrains. Traversability segmentation aims to find an arbitrary-shaped mask covering the traversable regions (termed free-space). Current research for traversability segmentation can be divided into two scenarios: outdoor and indoor environments. Compared to the outdoor environments mainly consists of paved roads, indoor environments present unique challenges for traversability segmentation, because of diverse lighting conditions, glass doors, various floor colors and textures, arbitrary shaped appliances and furniture, presence …

Exploratory Prompting Of Large Language Models To Act As Co-Pilots For Augmenting Business Process Work In Document Classification, Jose Ramon Ilagan, Joseph Benjamin R. Ilagan, Claire Louisse Basallo, Zachary Matthew Alabastro

Quantitative Methods and Information Technology Faculty Publications

Businesses deal with different types of documents containing unstructured documents. The data in these documents must be converted into digital forms other automated systems could only process. One generic use case is document classification, which usually involves manual transformation due to human understanding needed in the process. These documents go beyond those generated through regular business transactions and operations and also include web-based content such as online news, blogs, e-mails, and various digital libraries. Recent developments in robotic process automation (RPA) and artificial intelligence (AI) aim to automate the otherwise expensive, time-consuming, and repetitive manual steps. Through more powerful natural …

Effectiveness Of Generative Ai On The Development Of Graphic Software Artifacts, Gary Clynch, Ellen Mezera

Academic Poster Collection

Effectiveness of Generative AI on the Development of Graphic Software Artifacts

Algorithms For Safe And Robust Motion Control Of Autonomous Ground Robots And Vehicles, Manavendra Desai

Wayne State University Dissertations

This dissertation explores, investigates, develops and experimentally validates algorithms for safe and robust motion control of ground robots and vehicles. The emphasis is on design of controllers for ground autonomy with safety either embedded by design or enforced by appropriately wrapping a safety net around an existing nominal controller.

Ground robots have been considered for their importance in compensating for labour or personnel shortage for material handling in the manufacturing, logistics and healthcare sectors. On the other hand, ground vehicles have been considered for the existing potential of making driver assistance systems safer and more intelligent, particularly in at-the-limit driving …

Real-Time Detection Of Sea Turtles Using Uav And Neural Networks On Edge Devices, Jose A. Gonzalez Nunez, Jose G. Gonzalez Nunez, Mustafa I. Akbas, Patrick Currier, Nickolas D. Macchiarella

Journal of Aviation/Aerospace Education & Research

Sea turtle populations continue to diminish around the globe for various reasons. Therefore, the need for innovative solutions to monitor sea turtles has been increasing. This research paper focuses on an innovative application of artificial intelligence (AI) and machine learning (ML) together with unmanned aerial vehicles (UAV) to improve sea turtle conservation efforts. We outline the design, implementation, and evaluation of a system that deploys UAVs equipped with high-resolution cameras, coupled with a purpose-built neural network to recognize, classify, and monitor sea turtles. This project thus serves as a platform for understanding the wider applicability and limitations of this technology …

Post-Capture Synthesis Of Images Using Manipulable Integration Functions, Paul Eberhart

Theses and Dissertations--Computer Science

Traditional photographic practice, as dictated by the properties of photochemical emulsion film, mechanical apparatus, and human operators, largely treats the sensitivity (gain) and integration interval as coarsely parameterized constants for the entire scene, set no later than the time of exposure. This frame-at-a-time capture and processing model permeates digital cameras and computer image processing. Emerging imaging technologies, such as time domain continuous imaging (TDCI), quanta image sensors (QIS), event cameras, and conventional sensors augmented with computational processing and control, provide opportunities to break out of the frame-oriented paradigm and capture a stream of data describing changes to scene appearance over …

Manifold Learning In Robotics: A Tutorial And Survey, Marcus Hawkins

Computer Science and Engineering Theses - Archive

In this article, we hope to represent the current state of the art of manifold learning in an understandable and approachable way. The authors will present a general overview core algorithms associated with linear and nonlinear dimensionality reduction techniques, give rudimentary definitions from differential geometry, and tenets of robotic perception, manipulation and path planning. Some of the historical applications of these algorithms will be presented, as well as conjectures about future uses, through examples from peer-reviewed journals.

Exploring Machine Learning Techniques For Embedded Hardware, Neel R. Vora

Computer Science and Engineering Theses - Archive

This thesis delves into the intricate symbiosis between machine learning (ML) methodologies and embedded hardware systems, with a primary focus on augmenting efficiency and real-time processing capabilities across diverse application domains. It confronts the formidable challenge of deploying sophisticated ML algorithms on resource-constrained embedded hardware, aiming not only to optimize performance but also to minimize energy consumption. Innovative strategies are explored to tailor ML models for streamlined execution on embedded platforms, with validation conducted across various real-world application domains. Notable contributions include the development of a deep-learning framework leveraging a variational autoencoder (VAE) for compressing physiological signals from wearables while …

Stock Price Trend Prediction Using Emotion Analysis Of Financial Headlines With Distilled Llm Model, Rithesh H. Bhat

Computer Science and Engineering Theses - Archive

Capturing the volatility of stock prices helps individual traders, stock analysts, and institutions alike increase their returns in the stock market. Financial news headlines have been shown to have a significant effect on stock price mobility. Lately, many financial portals have restricted web scraping of stock prices and other related financial data of companies from their websites. In this study we demonstrate that emotion analysis of financial news headlines alone can be sufficient in predicting stock price movement, even in the absence of any financial data. We propose an approach that eliminates the need for web scraping of financial data. …

Development Of A Collaborative Research Platform For Efficient Data Management And Visualization Of Qubit Control, Devanshu Brahmbhatt

Computer Science and Engineering Theses - Archive

This thesis introduces QubiCSV, a pioneering open-source platform for quantum computing field. With an emphasis on collaborative research, QubiCSV addresses the critical need for specialized data management and visualization tools in qubit control. The platform is crafted to overcome the challenges posed by the high costs and complexities associated with quantum experimental setups. It emphasizes efficient utilization of resources through shared ideas, data, and implementation strategies. One of the primary obstacles in quantum computing research has been the ineffective management of extensive calibration data and the inability to visualize complex quantum experiment outcomes effectively. QubiCSV fills this gap by offering …

Adaptive Load-Aware Elastic Data Reduction And Re-Computation For Adaptive Mesh Refinement, Mengxiao Wang

Computer Science and Engineering Theses - Archive

The increasing performance gap between computation and I/O creates huge data management challenges for simulation-based scientific discovery. Data reduction, among others, is deemed to be a promising technique to bridge the gap through reducing the amount of data migrated to persistent storage. However, the reduction performance is still far from what is being demanded from production applications. To this end, we propose a new methodology that aggressively reduces data despite the substantial loss of information, and re-computes the original accuracy on-demand. As a result, our scheme creates an illusion of a fast and large storage medium with the availability of …

A Learning-Based Framework For Autonomous Robotic Operations In Resource-Denied Environments, Joseph M. Cloud

Computer Science and Engineering Dissertations - Archive

Establishing a sustained human presence beyond Earth necessitates the development of autonomous systems capable of extracting and utilizing local resources. On the Moon, in-situ resource utilization (ISRU) is essential to reduce the dependency on Earth-based supplies. Leveraging lunar resources such as water ice for life support and fuel production or regolith for surface construction will enable long-term lunar missions and future deep space exploration. NASA's Artemis program is targeting the lunar south pole (LSP), an area with raw, yet abundant resources available. The harsh environmental conditions present significant challenges for humans operating and installing surface infrastructure. Autonomous robotic systems are …

Web-Based Visualization Of Spatial And Spatio-Temporal Data Using Integrated Datasets, Mohammad Shaito

Computer Science and Engineering Dissertations - Archive

Currently, spatial geographic data can be collected for many applications that involve data on the planet earth. These collected data typically have coordinates (x,y), or longitude and latitude in map space, and thus can be located and displayed on maps. Data alone represents facts and has no meaning on its own but becomes meaningful when it is associated with application knowledge, such as elections, crimes, disease, etc. For example, there is no meaning behind those numbers (1, 23, 125, 355, . . .), yet they are data that can gain meaning when correlated with the total number of cases of …

Bringing "Virtual" To "Reality": Enhancing Security And Usability On Vr System And Applications, Huadi Zhu

Computer Science and Engineering Dissertations - Archive

With the rapid advancements in computer science, electronics, optics, and related fields, virtual reality (VR) gradually penetrates into our daily lives, and is predicted to become a core technology in the near future. Despite its potentials, however, existing designs and solutions for VR applications remain at the infant stage, introducing limited usability and efficiency for real-world users. Besides, the increasing prevalence of VR presents new security and privacy threats due to the vast amount of information stored in or accessible through VR devices. To bridge this gap, we exploit and combine techniques from computer science and human biology, as well …