Physical Sciences and Mathematics Commons^™

Do Pre-Trained Models Benefit Knowledge Graph Completion? A Reliable Evaluation And A Reasonable Approach, Xin Lv, Yankai Lin, Yixin Cao, Lei Hou, Juanzi Li, Zhiyuan Liu, Peng Li, Jie Zhou

Research Collection School Of Computing and Information Systems

In recent years, pre-trained language models (PLMs) have been shown to capture factual knowledge from massive texts, which encourages the proposal of PLM-based knowledge graph completion (KGC) models. However, these models are still quite behind the SOTA KGC models in terms of performance. In this work, we find two main reasons for the weak performance: (1) Inaccurate evaluation setting. The evaluation setting under the closed-world assumption (CWA) may underestimate the PLM-based KGC models since they introduce more external knowledge; (2) Inappropriate utilization of PLMs. Most PLM-based KGC models simply splice the labels of entities and relations as inputs, leading to …

Translate-Train Embracing Translationese Artifacts, Sicheng Yu, Qianru Sun, Hao Zhang, Jing Jiang

Research Collection School Of Computing and Information Systems

Translate-train is a general training approach to multilingual tasks. The key idea is to use the translator of the target language to generate training data to mitigate the gap between the source and target languages. However, its performance is often hampered by the artifacts in the translated texts (translationese). We discover that such artifacts have common patterns in different languages and can be modeled by deep learning, and subsequently propose an approach to conduct translate-train using Translationese Embracing the effect of Artifacts (TEA). TEA learns to mitigate such effect on the training data of a source language (whose original and …

Foundations Of Plasmas For Medical Applications, T. Von Woedtke, Mounir Laroussi, M. Gherardi

Electrical & Computer Engineering Faculty Publications

Plasma medicine refers to the application of nonequilibrium plasmas at approximately body temperature, for therapeutic purposes. Nonequilibrium plasmas are weakly ionized gases which contain charged and neutral species and electric fields, and emit radiation, particularly in the visible and ultraviolet range. Medically-relevant cold atmospheric pressure plasma (CAP) sources and devices are usually dielectric barrier discharges and nonequilibrium atmospheric pressure plasma jets. Plasma diagnostic methods and modelling approaches are used to characterize the densities and fluxes of active plasma species and their interaction with surrounding matter. In addition to the direct application of plasma onto living tissue, the treatment of liquids …

Graphcode2vec: Generic Code Embedding Via Lexical And Program Dependence Analyses, Wei Ma, Mengjie Zhao, Ezekiel Soremekun, Qiang Hu, Jie M. Zhang, Mike Papadakis, Maxime Cordy, Xiaofei Xie, Yves Le Traon

Research Collection School Of Computing and Information Systems

Code embedding is a keystone in the application of machine learning on several Software Engineering (SE) tasks. To effectively support a plethora of SE tasks, the embedding needs to capture program syntax and semantics in a way that is generic. To this end, we propose the first self-supervised pre-training approach (called Graphcode2vec) which produces task-agnostic embedding of lexical and program dependence features. Graphcode2vec achieves this via a synergistic combination of code analysis and Graph Neural Networks. Graphcode2vec is generic, it allows pre-training, and it is applicable to several SE downstream tasks. We evaluate the effectiveness of Graphcode2vec on four (4) …

Optimal In‐Place Suffix Sorting, Zhize Li, Jian Li, Hongwei Huo

Research Collection School Of Computing and Information Systems

The suffix array is a fundamental data structure for many applications that involve string searching and data compression. Designing time/space-efficient suffix array construction algorithms has attracted significant attention and considerable advances have been made for the past 20 years. We obtain the \emph{first} in-place suffix array construction algorithms that are optimal both in time and space for (read-only) integer alphabets. Concretely, we make the following contributions: 1. For integer alphabets, we obtain the first suffix sorting algorithm which takes linear time and uses only $O(1)$ workspace (the workspace is the total space needed beyond the input string and the output …

The Bracelet: An American Sign Language (Asl) Interpreting Wearable Device, Samuel Aba, Ahmadre Darrisaw, Pei Lin, Thomas Leonard

Chancellor’s Honors Program Projects

No abstract provided.

Sparse Model Selection Using Information Complexity, Yaojin Sun

Doctoral Dissertations

This dissertation studies and uses the application of information complexity to statistical model selection through three different projects. Specifically, we design statistical models that incorporate sparsity features to make the models more explanatory and computationally efficient.

In the first project, we propose a Sparse Bridge Regression model for variable selection when the number of variables is much greater than the number of observations if model misspecification occurs. The model is demonstrated to have excellent explanatory power in high-dimensional data analysis through numerical simulations and real-world data analysis.

The second project proposes a novel hybrid modeling method that utilizes a mixture …

Chandra Spectral Constraints Of The Low-Metallicity Collision Ring Cartwheel Galaxy, Chloë Benton

Physics Undergraduate Honors Theses

With the sophistication of the Chandra X-Ray Observatory, many observations of the behaviors of binary systems have been researched and studied. As an excellent example, for the low metallicity Cartwheel Galaxy, some fraction of the X-ray photons emitted from X-ray binaries have been intercepted by the detectors on the telescope and have been compiled into “events file.” When modeled properly, this data can produce insight into the behavior of the stellar populations in this object. The primary goal of this project is to use these events files to measure the X-ray properties of the X-ray binaries in the Cartwheel Galaxy …

Scale-Free Behavioral Dynamics Directly Linked With Scale-Free Cortical Dynamics, Sabrina Jones

Physics Undergraduate Honors Theses

In organisms, an interesting phenomenon occurs in both behavior and neuronal activity: organization with fractal, scale-free fluctuations over multiple spatiotemporal orders of magnitude (1,2). In regard to behavior, this sort of complex structure-- which manifests itself from small scale fidgeting to purposeful, full body movements-- may support goals such as foraging (3-6), visual search (4), and decision making (7,8). Likewise, the presence of this sort of structure in the cerebral cortex in the form of spatiotemporal cascades, coined “neuronal avalanches,” may offer optimal information transfer (9). Thus, when considering the functional relationship between the cerebral cortex and movements of the …

Ongoing Calculus In The Cerebral Cortex, Luke Long

Physics Undergraduate Honors Theses

Various modes of neuronal computations have long been theorized to be possible based on the structure and geometry of the brain. These computations also seem necessary for many of the integral functions of the brain, like information processing and regulatory processes in the body. However, experimental data directly supporting these claims have been rare.

In this study, data collected in mice from a large number of neurons over a long period of time provided the opportunity to search for some of these computations, specifically change detection and squaring calculations. Using Matlab, the goal of this analysis was to find statistically …

Synthesis, Characterization, And Functionality Of Novel 2d Material Sn3p2, Cory Stephenson

Physics Undergraduate Honors Theses

In recent years, two-dimensional (2D) materials have gained a lot of attention due to their potential applications in devices and their promise to revolutionize technology. In our group, we are capable of synthesizing such materials and study their properties. I have discovered a new material of the tin-phosphorus family that displays the 2D layered structure and therefore shows potential to be used in future devices due to this reduced dimensionality from the typical three-dimensional counterpart. This 2D structure may allow for new phenomena to emerge as there is no longer any interatomic interactions along the z- axis. Under this motivation, …

Monitoring The M-Dwarf Host Stars Of Tess Exoplanet Candidates: Stellar Flares And Habitability, Ashley Lieber

Physics Undergraduate Honors Theses

In the search for life beyond our solar system, the study of M-dwarfs has become increasingly important due to their unique characteristics including their small size, flaring capabilities, and long lifespans. Their small size allows for exoplanet detection due to observable gravitational interactions, and the stellar flares could potentially trigger prebiotic life on exoplanets in the system. Lastly, their long lifespans may provide the conditions necessary to foster prebiotic life and the development of more complex organisms over time. Flare rate is a critical factor in determining the habitability of the exoplanet due to its potential to damage or incubate …

Emergent Spectra Of Young X-Ray Emitting Populations Across Environments, Alex Siebenmorgen

Physics Undergraduate Honors Theses

We construct reasonably accurate models of the X-Ray spectra of a multitude of sources in M51. We construct both average and individual models for the sources, which are split into 16 groups as the counts per source increases. Then, we create a plot to show how the model-predicted values of column density (nH) and photon index (gamma) change with luminosity. These models will be used to create an accurate X-Ray stellar energy distribution (SED) for M51, and to better understand how the SED changes with environmental factors like metallicity and star formation rate (SFR).

Electrochromic Windows: Return On Investment Analysis, Thomas Michael Caprio

Construction Management

In recent years, sustainability in construction has become a topic of undeniable interest. With sustainability certifications becoming more prominent and desirable, members of the construction industry have begun to seek alternatives for standard systems to keep up with market trends. One of the main issues that the construction industry faces is accessibility of technology. While there are new products and technologies being released frequently, members of the construction industry lack knowledge on the specific details and implications of working with newer systems. One piece of technology that was recently introduced is electrochromic glass, or smart glass. Electrochromic windows analyze sunlight …

Beyond Accuracy In Machine Learning., Aneseh Alvanpour

Electronic Theses and Dissertations

Machine Learning (ML) algorithms are widely used in our daily lives. The need to increase the accuracy of ML models has led to building increasingly powerful and complex algorithms known as black-box models which do not provide any explanations about the reasons behind their output. On the other hand, there are white-box ML models which are inherently interpretable while having lower accuracy compared to black-box models. To have a productive and practical algorithmic decision system, precise predictions may not be sufficient. The system may need to have transparency and be able to provide explanations, especially in applications with safety-critical contexts …

Modeling And Debiasing Feedback Loops In Collaborative Filtering Recommender Systems., Sami Khenissi

Electronic Theses and Dissertations

Artificial Intelligence (AI)-driven recommender systems have been gaining increasing ubiquity and influence in our daily lives, especially during time spent online on the World Wide Web or smart devices. The influence of recommender systems on who and what we can find and discover, our choices, and our behavior, has thus never been more concrete. AI can now predict and anticipate, with varying degrees of accuracy, the news article we will read, the music we will listen to, the movies we will watch, the transactions we will make, the restaurants we will eat in, the online courses we will be interested …

Sustainable Nanocatalysis In Water For C–C And C–N Cross-Couplings., Tharique Ahammed Ansari Nalakath

Electronic Theses and Dissertations

The development and application of various micelle-enabled nanopalladium-catalyzed organic transformations are discussed in this dissertation. These methodologies primarily harness the “micellar effect” imparted by the custom-designed amphiphile PS-750-M. Modern spectroscopic and imaging techniques were employed to systematically study the metal-micelle interactions and answer the fundamental questions regarding the operational nature of micellar catalysis. Chapter 1 reviews the history and evolution of organic synthesis, providing a critical outlook regarding the alarming inefficiency of contemporary synthetic practices indicating an inevitable transition to more sustainable synthetic techniques. In this context, the development of micellar catalysis is discussed, emphasizing its basic principles, applications, and …

Structural, Charge Transport, Gas Sensing, Magnetic, Pseudocapacitive, And Electrocatalytic Properties Of Perovskite Oxides., Surendra Bahadur Karki

Electronic Theses and Dissertations

Perovskites are functional materials with the general formula ABO₃ (A = alkali, alkaline earth or lanthanoid cations and B = transition metal or main group cations). These materials are marked by a variety of crystal structures and interesting properties such as colossal magnetoresistance, ferroelectricity, multiferroicity, superconductivity, pseudocapacitance, gas sensing, charge transport, and electrocatalytic properties. The formula of perovskite can be written as AA’BB’O₆, when there is ordering between two cations over A and B-sites. Such compounds are called double perovskite oxides. Some amount of oxygen could be lost from crystal structure without decomposition of the phase. Such …

A Combinatorial Approach To Fairness Testing Of Machine Learning Models, Ankita Ramjibhai Patel

Computer Science and Engineering Theses

Machine Learning (ML) models could exhibit biased behavior, or algorithmic discrimination, resulting in unfair or discriminatory outcomes. The bias in the ML model could emanate from various factors such as the training dataset, the choice of the ML algorithm, or the hyperparameters used to train the ML model. In addition to evaluating the model’s correctness, it is essential to test ML models for fair and unbiased behavior. In this thesis, we present a combinatorial testing-based approach to perform fairness testing of ML models. Our approach is model agnostic and evaluates fairness violations of a pre-trained ML model in a two-step …

Human Behavior Modeling In Long Videos: Drowsiness Detection And Action Segmentation, Reza Ghoddoosian

Computer Science and Engineering Dissertations

"In this thesis we focus on two instances of human behavior modeling in long untrimmed videos: drowsiness detection, and action segmentation. In the first section, we focus on drowsiness detection. Specifically, we introduce a large and public real-life dataset and a baseline temporal model to classify drowsiness into three stages of alert, low vigilant, or drowsy. In the second section, we study action segmentation in instructional videos under weak supervision. In order to save time and cost, weakly supervised methods are trained based on only video-level action sequences as opposed to a fully supervised method which is trained using frame-level …

A Non-Contact Based System To Measure Spo2 And Systolic/Diastolic Blood Pressure Using Rgb-Nir Camera, Divya Saxena

Computer Science and Engineering Theses

In recent times, people have increasingly self-assessed their health using different devices on their bodies that monitor physiological attributes such as their oxygen level and blood pressure (BP) to monitor their health. One of the most popular health concerns that became prominent during the COVID-19 pandemic was the blood oxygen saturation (SPO2) level. It became increasingly important to monitor SPO2 in patients, time and again to determine whether the right amount of oxygen is in the blood. Low oxygen levels usually indicate there may be an issue with oxygen circulation or supply and thus informs diagnostic and treatment decisions such …

Learning Topology Preserving Embeddings For Speeding Up Nearest Neighbor Retrieval, Mason Lary

Computer Science and Engineering Theses

Given a database of objects and a query object, it’s possible to gather a number of the closest neighbors to the query object. This operation is important to a number of diverse fields such as computer vision, content- based information retrieval, and chemistry. However, distance measures used to determine neighbors can cause queries to be computationally expensive, either because the distance measure is complex or because it is nonmetric and prevents efficient indexing methods. This work presents novel methods of triplet mining that enable neural networks using triplet loss to learn the manifold that data resides in. These neural networks …

Distraction Detection And Intention Recognition For Gesture-Controlled Unmanned Aerial Vehicle Operation, Debayan Datta

Computer Science and Engineering Theses

Gesture Control as a way to replace more conventional remote-control operations has been pursued for a significant period of time with different levels of success. The use of gestures to control different types of human interfaces is today predominantly seen in the multimedia sector. People perform easy and intuitive gestures to control their televisions, to interact with multimedia, and to play games. Also, much research has been carried out to experiment with human interfaces to numerous augmented and virtual reality devices and tasks. The results of these experiments were so exciting that researchers started to expand the use of gestures …

Designing Large-Scale Key-Value Systems On High-Speed Storage Devices, Xingsheng Zhao

Computer Science and Engineering Dissertations

With the evolution of new technologies, such as edge computing, full self-driving, virtual reality, and multi-media streaming, the volume of data is growing at an accelerated speed. The global data volume could achieve 175~zettabytes by 2025. With this huge amount of data, the focus of data management has been shifted from traditional SQL databases to NoSQL databases, which provide higher performance and better scalability. Key-value (KV) stores are a common type of NoSQL database and are becoming a major storage infrastructure in various application domains. With the development of high-speed storage devices, such as NVMe SSD, Open-channel SSD, and non-volatile …

Robust Noise-Based Attacks Against Audio Event Detection Systems, Rodrigo Augusto Silva Dos Santos

Computer Science and Engineering Dissertations

The massive advances on the field of deep neural networks in the 2000 and 2010 decades led to an overwhelming adoption of these algorithms on all sorts of domains and applications. Under this widespread adoption scenario, it is natural that these neural networks have also been employed on safety-related use cases, bringing substantial improvements to the performance of existing as well as novel systems. Examples of these safety-inclined applications include scene recognition, object detection and tracking, speech recognition, audio event detection and classification, just to cite a few ones. Unfortunately, these neural network algorithms have been shown to be vulnerable …

“Strict Moderation?” The Impact Of Increased Moderation On Parler Content And User Behavior, Nihal Kumarswamy

Computer Science and Engineering Theses

Social media platforms have brought people from different backgrounds, ethnicity, race, gender, etc together to form a platform to share ideas and opinions and discuss news events among other social events. Unfortunately, these platforms have also been a safe haven for abusive users who harass, bully other users or spread misinformation and disinformation. Social media platforms have a huge incentive to police these abusive users and keep them in check to allow other genuine users to use their platform. Social media platforms employ several different content moderation techniques to perform this task. These techniques vary across platforms, for example, Parler …

Detection And Classification Of Object Presence And Characteristics In A Water Container Using High Frequency Ultrasound, Mehul Vishal Sadh

Computer Science and Engineering Theses

Detection and characterization of soluble, diffuse, and solid objects and their characteristics in water has important implications in various applications, including water quality assessment and incontinence monitoring for health applications. In particular in the latter task, it is essential to be able to non-intrusively detect the appearance, presence, and consistency of materials in the water without the need for special purpose instruments or a special purpose setting. Rather, it would be important that sensing could be performed int he context of existing toilet systems. To achieve this, this work investigates the potential use of high frequency sonar sensors retrofitted to …

Context-Aware Graph-Based Self-Supervised Learning Of Whole Slide Images, Milam Aryal, Nasim Yahyasoltani

Computer Science Faculty Research and Publications

The gigapixel resolution of a single whole slide image (WSI), and the lack of huge annotated datasets needed for computational pathology, makes cancer diagnosis and grading with WSIs a challenging task. Moreover, downsampling of WSIs might result in loss of information critical for cancer diagnosis. Motivated by the fact that context such as topological structures in the tumor environment may contain critical information in cancer grading and diagnosis, a novel two-stage learning approach is proposed. Self-supervised learning is applied to improve training through unlabled data and graph convolutional network (GCN) is deployed to incorporate context from tumor and surrounding tissues. …

Two Project On Information Systems Capabilities And Organizational Performance, Giridhar Reddy Bojja

Masters Theses & Doctoral Dissertations

Information systems (IS), as a multi-disciplinary research area, emphasizes the complementary relationship between people, organizations, and technology and has evolved dramatically over the years. IS and the underlying Information Technology (IT) application and research play a crucial role in transforming the business world and research within the management domain. Consistent with this evolution and transformation, I develop a two-project dissertation on Information systems capabilities and organizational outcomes.

Project 1 examines the role of hospital operational effectiveness on the link between information systems capabilities and hospital performance. This project examines the cross-lagged effects on a sample of 217 hospitals measured over …

New Accurate, Explainable, And Unbiased Machine Learning Models For Recommendation With Implicit Feedback., Khalil Damak

Electronic Theses and Dissertations

Recommender systems have become ubiquitous Artificial Intelligence (AI) tools that play an important role in filtering online information in our daily lives. Whether we are shopping, browsing movies, or listening to music online, AI recommender systems are working behind the scene to provide us with curated and personalized content, that has been predicted to be relevant to our interest. The increasing prevalence of recommender systems has challenged researchers to develop powerful algorithms that can deliver recommendations with increasing accuracy. In addition to the predictive accuracy of recommender systems, recent research has also started paying attention to their fairness, in particular …