Computer Science

In this research project, predictive capabilities of Long Short-Term Memory (LSTM) models for identifying thermoacoustic instability within combustion systems are explored. LSTMs, a subset of Recurrent Neural Networks, are particularly effective in overcoming the vanishing gradient problem, thus enhancing their ability to learn from and classify time series data. As a part of this project, a machine learning model will ... Read More

The heightening demand for spectrum from wireless communication services has necessitated the development of more effective frequency allocation methods. Leveraging Dynamic Spectrum Allocation (DSA) techniques, we generated and modeled data from several allocation algorithms to understand trends in optimal performances. In wireless networks, transmitters with overlapping coverage must be assigned different frequencies to avoid harmful interference, posing complex frequency allocation ... Read More

Pancreatic cancer has less than a 10% survival rate after 5 years of treatment. Drug resistance, among other issues, is one of the main reasons for tumor relapse. One of the main mutations in pancreatic cancer is in the protein KRAS, which controls cell proliferation and when mutated sends signals continuously for the cell to divide. Recently, a new FDA-approved ... Read More

This study aims to address the pressing need for accessible and accurate detection of heart irregularities amidst the rising prevalence of cardiovascular diseases. Leveraging machine learning's capability to process extensive datasets, the research proposes the development of predictive models for identifying heart rhythm irregularities. However, a significant challenge in healthcare persists: ensuring the security and privacy of patient data. To ... Read More

Our project utilized various computational techniques, including machine learning, to address an existing issue with diagnosing esophageal diseases, specifically Gastroesophageal Reflux Disease (GERD) and Systemic Sclerosis (SSc). In the early stages of both diseases, symptoms such as chest pain, heartburn, and regurgitation are similar, potentially causing one to be mistaken for the other. Additionally, both of their current diagnoses are ... Read More

Rapid advancement of deep neural networks has significantly improved various tasks such as image and speech recognition. However, as complexity of the models increases, computational costs and the number of parameters increase, making it more difficult to be implemented on resource-limited devices. This paper proposes a novel memorization-based inference (MBI) model that is compute-free and size-agnostic. Our work capitalizes on ... Read More

Transformer-based large language models (LLMs) have gained prominence over the last few years, with their ability to generate human-like content. One of the biggest issues with LLMs is “hallucination” where they generate factually incorrect output in response to queries that don’t have much support from the data that was used to train the model. Previous methods for mitigating hallucinations, such ... Read More

With the introduction of Quantum Computers, multiple methods have arisen that focus on simulating Quantum Computers. The most common and popular way many of the largest Quantum Computers are built is with superconducting qubits, similar to those built at places such as Google and IBM. Neutral Atom Technology is a quantum computer that utilizes neutral atoms as qubits. By shining ... Read More

With text-to-image (“TTI”) models becoming increasingly popular, it is imperative that we ensure they are as unbiased as possible, an issue many text-to-image models are currently facing. For example, when we ran Google’s Image-FX text-to-image model to create an image of “a CEO from the Fortune 500 list” or “a CEO of the top 10 tech companies,” it always returned ... Read More

The purpose of this design experiment was to attempt to determine the accuracy/possibility of using extended Kalman filters (EKFs) to approximate a human’s shoulder’s location when occluded in a depth camera’s point of view. The project was conducted entirely through code and avoided any human involvement/error. The outcome of this research project may result in more accurate measurements for human-body ... Read More

This project explores the potential applications of millimeter-wave (mmWave) radar technology for object and activity recognition. Using a Texas Instruments radar wave card and several cases, the data retrieved demonstrates the capability of mmWave radar to distinguish between different objects. Mmwave is unique in the sense that it uses the reflection points of wifi-waves which are more adverse in their ... Read More

Many attempts have been made to use generative artificial intelligence—neural networks that create new text or images given inputs of the same type—to synthesize characters or entire fonts from a few characters. Previous studies have used glyph (individual strokes that make up characters) detection and conjoinment to create these characters but fell short in connecting the glyphs to reproduce characters. ... Read More

Self-assembling peptides, or chains of amino acids that form various structures in response to environmental conditions, have a variety of uses in material science as well as biomedicine. These uses include drug delivery, or as drugs themselves. In material sciences, self-assembling peptides can be used to create materials with a variety of properties. Our work hopes to utilize machine learning ... Read More

Quantum circuit gate and depth optimization is important for improving accuracy and reducing errors when running quantum computations on quantum computers. Current quantum optimization uses a number of algorithms, and heuristic brute force methods to find equivalent circuits for quantum gates Gate reduction minimizes the chance of quantum states deteriorating into mixed states, leading to better performance and fidelities. ZX-calculus ... Read More

Stroke rehabilitation faces significant hurdles in providing continuous, real-time care, particularly outside clinical settings. Traditional approaches, while beneficial, are hampered by their intermittent nature and lack of personalized, real-time monitoring. Addressing this gap, this research is the introduction of a wearable sensor technology, designed as a user-friendly bracelet. This device leverages advanced cross-modal deep learning techniques, including variational auto-encoders and ... Read More

Following the Hamas invasion of Israel on October 7th, 2023, there has been a significant increase in media coverage of the Israel/Hamas conflict. Given the risk of extreme polarization and the inherent unpredictability of the internet, the aim of this research is to perform a detailed examination of how media influences public perception and reaction in the context of a ... Read More

The main objective of this research was to explore the integration of machine learning algorithms, particularly the Random Forest Regressor model utilizing decision trees, in enhancing the Black-Scholes Model for options pricing within the financial industry. Machine learning is becoming more prevalent in the financial sector, so this research gives more insight into how it is directly applicable in the ... Read More

Age-related macular degeneration (AMD) is a progressive neurodegenerative eye disorder characterized by eventual degeneration of the retinal pigment epithelium (RPE) leading to permanent vision loss. Artificial intelligence (AI) and machine learning (ML) have revolutionized healthcare by advancing clinical diagnosis leveraging its ability to analyze vast amounts of patient data and accurately predict future outcomes. With no definitive treatment for AMD, ... Read More

This experiment developed the FlaviExplore platform with a main goal of making gathering West Nile virus (WNV) data simpler for researchers and extending phylogenetic knowledge about WNV. Currently, many researchers have to spend an excessive amount of time gathering sequences and converting them to fasta files, which could slow down the process of gaining new insights about the virus. This ... Read More

The surge of e-commerce demands innovative solutions to streamline last-mile delivery logistics. Autonomous delivery vehicles (ADVs) offer a promising avenue, also combatting the lack of delivery drivers. However, their success hinges on effectively managing the complexities arising from diverse delivery modes (e.g., aerial and ground-based) in obstructed or constrained environments. Traditional Multi-Agent Reinforcement Learning (MARL) approaches may not optimally coordinate ... Read More

Urban areas worldwide are bustling with activity and constantly changing. This perpetual motion is a testament to the vibrancy of city life and provides a rich basis for gathering and analyzing data. Visual analytics is crucial in urban settings, and it is supported by many frameworks to investigate urban data more effectively. I used one of these frameworks together with ... Read More

This research paper explores the application of Linear Feedback Shift Registers (LFSRs) to enhance the compression of neural networks. LFSRs, which employ a linear function to determine input bits based on previous states, are commonly used for generating bit sequences and pseudo-random numbers that can be used to generate pseudo-random weight approximations. Compressed neural networks offer a transformative solution by ... Read More

Natural Language Processing (NLP) has witnessed significant advancements with the emergence of large language models (LLM) capable of understanding and generating human-like text. However, there remains a critical need to explore and understand their efficiency and effectiveness, especially in processing languages beyond English. This study aims to evaluate the efficiency of various large language models in capturing semantic meaning across ... Read More

ImageJ is a software platform for open-source image analysis that has assisted researchers with various image analysis applications. Its success has been primarily attributed to supportive and cooperative developer and user communities. ImageJ is also useful for data segmentation, which assigns a label to every pixel in an image such that images with the same label share similar qualities. Segmenting ... Read More

The study of computer-human natural language interaction is the focus of the artificial intelligence field known as natural language processing (NLP). It entails creating models and algorithms that let computers comprehend, interpret, and produce human language. Language translation, emotion analysis, chatbots, speech recognition, and text summarization are just a few of the many uses for NLP. Natural language processing was ... Read More

3D reconstruction for images has several applications including mixed reality, game development, and film production. Recently, several 3D reconstruction algorithms have been proposed such as Neural Radiance Fields (NERF) and 3d Gaussian Splatting, which can synthesize realistic images from novel viewpoints. However, these algorithms do not model the physical interactions of objects in scenes, which are crucial for an immersive ... Read More

Beginning in February 2022, Russia’s invasion in Ukraine marked one of the largest invasions of a European country since World War II and impacted numerous people worldwide. Our project aims to study public opinions and topics of discussion during this international crisis by analyzing daily tweets and news articles from April 2022 to June 2023. For our analyses, we utilized ... Read More

As the number of deepfake image generators has skyrocketed, so has the number of deepfake detection models. However, explainability in these models remains underexplored, which is the key to building trust with users, and these are black box models. To address this issue, we studied how humans can understand the model’s justification behind its decision by integrating explainable AIs (XAIs) ... Read More

Self-determination theory (SDT) states that humans are motivated to pursue things that are intrinsically valuable to them, such as wellness or autonomy. Serious games implement different tactics such as identity, interactivity, agency or control, challenge, narrative, feedback, and immersion, to motivate the player to play the game and absorb its “serious” content, which is meant to improve the player’s expertise ... Read More

Entanglement is a fundamental concept in quantum mechanics that describes when two particles are intrinsically correlated and that one of the particles cannot be described without the others. The phenomenon of entanglement is not well understood by physicists. However, we may be able to better understand its behaviors by quantifying entanglement through entanglement measures. Entanglement measures aim to quantify the ... Read More

Human retinoblastoma is a pediatric cancer initiated by RB gene mutations in the developing retina. Originating in the retina, RB evolves in four separate stages. However, most patients do not have a distinct transition through these separate stages and these stages are not always preceded by a detectable preface state, making the cancer difficult to model. In this project, we ... Read More

This research presents an overview for search query management in StackOverflow, a popular platform for programming in which users can ask and answer questions about their code. With the use of Natural

Language Processing (NLP) techniques, and models including TF-IDF (Term Frequency – Inverse Document Frequency), GloVe (Global Vectors for Word Representation), OpenAI’s GPT, and Sentence-BERT (Bidirectional Encoder Representations from ... Read More

Mixed reality (XR) can be simply described as a fusion between augmented (AR) and virtual reality (VR), allowing interactions to occur between a virtual world and physical elements. With powerful applications in the educational, medical, and entertainment industries, our exploration of this vast technology primarily focuses on utilizing a virtual setting to track the location of objects in the physical ... Read More

Several machine learning models trained on electronic health records (EHR) data have been able to predict risk for Type 2 Diabetes accurately, but the efficacy in risk prediction for models trained on synthetic genotype data remains to be tested extensively. Using data gathered from Genome-Wide

Association Studies (GWAS) analyses, we identified several genes correlated with Type 2 Diabetes, each with ... Read More

Clinical Large-Language Models (LLMs) are essential to the biomedical industry: they can analyze and interpret physician notes, anonymize patient health information, and supplement diagnoses in various medical practices. Nowadays, most widely used clinical large-language models are trained on medical text that masks protected health information---corpora unlike the data that the models are used on in the field. Despite this, little ... Read More