YSP Alumni
Alumni Directory
2024 YSP Program Coordinators:
Lead Coordinator: Theodore Lourie
Co-Coordinators: Angel Le, Michael Marchev, Vicky Berry
>>2024 Video<<
Participating Labs
| Lab | YSP Students | Title | Abstract |
|---|---|---|---|
| Faculty: Alam, Md Noor E Mentors: Tianyu Yang | Rebecca Berberi, Ben Nguyen | Investigating the Impact of Short-Term Residential Opioid Use Disorder Rehabilitation on Treatment Completion for Individuals with Mental and Substance Use Disorders Presentation Poster | In the past 30 years, Opioid Use Disorder (OUD) has become a widespread health issue in the United States of America. In 2019, 1.6 million people in the U.S. were diagnosed with OUD. Of these people diagnosed with OUD, 39.1% were also reported to suffer from a mental illness. Patients with a diagnosed mental and substance use disorder are more likely to get opioid prescriptions, despite their increased risk of addiction and overdose, affecting their ability to complete OUD treatment. This study will examine the effectiveness of various OUD rehabilitation services on the completion of treatment for individuals who have co-occurring mental and substance use disorders in addition to OUD. Using the Substance Abuse and Mental Health Services Administration (SAMHSA) discharge data from 2021, this study will determine if there is a causal relationship between different types of rehabilitation and the OUD treatment completion. First, the data was processed to remove insignificant and unknown data points. Using statistical models from causal inference, the data was analyzed to select appropriate features used for matching and to predict treatment completion outcomes based on the covariates. This information was used to create causal conclusions between eight different types of services and the completion of OUD treatment. This study allows for a better understanding of treatment for individuals with OUD and co-occurring mental and substance use disorders. |
| Faculty: Alshawabkeh, Akram Mentors: Stephanie Sarrouf, Amanda Thomas, Muhammad Fahad Ehsan | Soline Fisher, Ahmed Othman | Effect of Granular Activated Carbon Pore Size on Electro-Oxidation Water Filtration Efficiency Presentation Poster | Water contamination is a global concern, but poses a greater threat to under resourced countries and those without efficient or affordable filtration resources. In Puerto Rico, groundwater contaminated from Superfund Sites has proven to be incredibly dangerous to the health of pregnant women and fetuses, young children, and elderly members of the community. This is due to the toxic nature of chemical contaminants at Superfund Sites, which have been identified globally as sites of high pollutant levels. Such pollutants from the sites diffuse into the environment and into drinking water sources, causing adverse health effects, even leading to long-term health issues, diseases including chorea, typhoid, and polio, miscarriages, and preterm births. These issues are prevalent within Puerto Rico, especially due to the scarcity of efficient, and sustainable, water filtration methods. The goal of Project 4 in PROTECT, an organization working to understand and improve the health and wellbeing of Puerto Rican locals, is to develop a portable, low-maintenance, and sustainable water filtration technology that will directly improve the quality of life for residents. To do so, an advanced electro-oxidative water filtration system using Granular Activated Carbon (GAC) as an adsorbent electrode was developed. This system relies on electrostatic interactions and oxidative reactions on the surface of the cathode to adsorb and degrade pollutants, thus purifying the water within the system. To improve adsorption, various parameters need to be optimized, some of which include; GAC pore size, amount of absorbent, pH of the system, and contact time. The focus of this study was to optimize system parameters for the adsorption of Congo Red, an industrial anionic dye, to improve adsorption. Throughout experimentation, it was found that a median amount of GAC along with pore sizes ranging on the smaller spectrum was most efficient in removal of pollutant from a pollutant of 10-40 ppm stock solution with an acidic pH, and it is hypothesized that GAC with smaller pore sizes and higher surface area combined with an optimized adsorbent mass will lead to further efficiencies. |
| Faculty: Amirabadi, Mahshid Mentors: Mojtaba Salehi | Gabriela Friel, Corbin Seidel | DC-DC Converter for Solar Powered Phone Chargers Presentation Poster | Photovoltaic cells (solar panels) have grown in prominence. However, a difficulty with using them is their inconsistent power production. Photovoltaic cells are composed of semiconductor cathodes and anodes that produce power when electrons are excited by solar energy. Unfortunately, the sun’s energy is often inconsistent, and PV cells’ efficiency, especially for small-scale, flexible panels, is only around 20%. This project aims to address the problem of optimal power delivery from photovoltaic cells. PV cells produce DC power in proportion to the sun’s intensity, necessitating the design of voltage regulator circuits that deliver a constant voltage output while maximizing power extracted from the cell. To do so, the ideal resistance level and voltage regulation efficiency must be determined. This research project aims to design a DC-to-DC voltage regulator that delivers a constant voltage and power output. We tested ideal load resistance and panel orientation, ultimately simulating and building two simple linear regulators for the panel. Additionally, we simulated a switching regulator to address photovoltaic power generation across a range of efficiency levels. |
| Faculty: Cassella, Cristian Mentors: Nicolas Casilli | Edi Ebong, Victor Lee | Subharmonic Tags for Localization Presentation Poster | We report on quasi-Harmonic Tags (qHT), which rely on parametric frequency dividers (PFDs) strongly coupled to electromechanical resonators, for localization and ranging applications. Unlike LiDAR, ultrasonic sensing, GPS, and inertial measurement units, qHTs offer a compact, low-profile localization platform that provides power-efficiency while mitigating ranging inaccuracies caused by multipath interference. qHTs perform localization through the fully passive generation of a frequency comb, which is activated once interrogated by a signal exceeding its power threshold. Because the comb-spacing of the inter-modulated signals depends on the power received by the qHT, the modulated output provides the basis for deriving the qHT’s displacement with respect to its interrogator. Such behavior permits to compute the exact location of a qHT when interrogated simultaneously by several interrogators at different positions. qHTs are posed as strong candidates to enable the quick, accurate, and power-efficient localization of lighter and smarter unmanned aerial vehicles (UAVs), revolutionizing surveying, agriculture monitoring, and search and rescue operations, especially in GPS-denied environments. Additionally, qHTs can also be applied to a plethora of other ranging applications such as “Find My” systems or video assistant referee technology. As a proof of concept, an experiment was conducted, demonstrating the dependency between the qHT’s comb spacing and its received interrogation power. Using this relationship, we created a simulation framework for accurate qHT-based UAV localization. |
| Faculty: Chakraborty, Srirupa Mentors: Natesan Mani, Jason Kantorow, Simran Pandey | Leo Murthy, Grace Niu | Enhancing Predictability in Antibody-Antigen Interaction Characterization Presentation Poster | Determining the structure of large, complex biomacromolecules by experimentation remains a major challenge, as methods such as X-ray crystallography and NMR spectroscopy are costly, time-consuming processes. Examining antibody-antigen (Ab-Ag) structures and interactions is particularly crucial to researching how antibodies serve as our frontline defense against viral infections. However, Ab-Ag binding is hindered by glycan shielding, which causes glycoproteins on the surfaces of viruses to block antibody binding, thus posing challenges in treating these viruses. Since glycans are difficult to study experimentally, our lab uses in silico modeling to provide higher accuracy atomistic models of glycoproteins. Google’s new AI-powered AlphaFold3 harnesses high potential in the study of Ab-Ag interactions because it can handle more molecules beyond amino acids, including glycans. AlphaFold3 applies advanced AI algorithms to precisely predict the structure of a protein given its sequence. Throughout our project, we evaluated how effectively AlphaFold3 can predict protein structure and Ab-Ag interactions compared to validated structures reported in literature. We then sought to curate a dataset in order to design a machine-learning based predictive algorithm for Ab-Ag interactions. We found that AlphaFold3 predicts highly accurate structures. This could help researchers save time and resources. |
| Faculty: Davidow, Juliet Mentors: Rebecca Hennessy | Harry Hall, Dayna Phan | Motivated Learning in Adolescence Presentation Poster | Adolescence is a life period defined by characteristics such as age, pubertal changes, and emotional maturity. Adolescence is a crucial period for brain development in which cognitive, social, and biological changes influence the way someone learns, makes decisions, and ultimately, thinks. To study adolescent brain development, we analyzed structural magnetic resonance imaging (MRI) data from 105 study participants (ages 13 – 30). Data were analyzed using Freesurfer, a neuroimaging analysis software that pre-processes the MRI data by performing skull-stripping, leaving only the brain tissue, and quantifying the gray and white matter portions of the brain. Freesurfer is prone to error, such as incorrectly classifying something or leaving bits of skull. To address these errors, we reviewed the data, rating the quality of each brain based on Freesurfer’s analysis, and manually added points of correction to the scans. This process ensures that the data are usable for further analysis in order to answer research questions regarding age-related changes and sex differences in adolescent brain development. Understanding how the adolescent brain develops can increase engagement and awareness surrounding motivated learning, education, and policy. |
| Faculty: Ebong, Eno Mentors: Nicholas O'Hare | Skyler Marnik, Celeste Nguyen | Identifying Key Regulators of Blood-Brain Barrier Permeability through Knock-Down Presentation Poster | Alzheimer’s disease (AD) often remains shrouded in mystery but has devastating social impacts to both the patient and their families. Since 2010, mortality rates from AD have risen 145%, creating a clear impetus for further research. Without any advancement, the onset of AD will increase by upwards of four times the current rate by 2050. We aim to demonstrate the connection between blood-brain barrier (BBB) dysfunction and endothelial glycocalyx (GCX) impairment. The GCX is believed to be degraded before late-stage AD events, thus making it a potential therapeutic target. This project specifically sought to validate the loss of the GCX core protein, CD44 and its associated glycosaminoglycan (GAG), hyaluronic acid (HA) through Western Blots and immunocytochemistry. Through validating the knockdown of CD44, functional assays may be performed to interrogate the link between GCX damage and BBB dysfunction. |
| Faculty: Jornet, Josep Mentors: Samar Elmaadawy | Isaac Chan, Kenneth Santizo | Thermal Effects of Terahertz-Band Radiation on Heart Tissue Presentation Poster | The terahertz (THz) frequency band has shown exceptional promise in wireless communication. Its superior data transfer speeds and non-invasive nature, among other advantages, have the potential to unlock vast possibilities in 6G and 7G technologies, including mobile devices, environmental monitoring, and healthcare. One particular application in the medical field is biomonitoring devices for the heart. We explore this application in our project, which aims to ensure the safety of THz radiation when interacting with human cardiac tissue. To do this, we extended an existing computational model in COMSOL Multiphysics© by developing both 2D and 3D models of the human heart to simulate the wave propagation and thermal effects of THz waves in cardiac tissue. These models help define safe limits for THz radiation in next-generation biomedical devices, accelerating the advancement of wireless networks. |
| Faculty: Kaeli, Dave | Vedi Raval | Data Analytics for Increasing Diversity in STEM Education Presentation Poster | The field of engineering has suffered from a lack of diversity, both in terms of race and gender. Some colleges have made progress in creating a more diverse learning environment for engineering students. This project is focused on understanding what factors are responsible for these positive changes so that they can be shared with the broader engineering community. Through our analysis of data from the Integrated Postsecondary Education Data System (IPEDS), we have identified schools that have made significant progress in terms of recruitment and retention of women in engineering. Our ongoing work is focused on identifying factors, such as federal funding grants, that influence these positive changes. |
| Faculty: Koppes, Abigail Mentors: Nolan Burson, Brent Buchinger | Leena Alshawabkeh, Tsering Shakya | Engineering Microfluidics for Human Health Applications Presentation Poster | A microfluidic chip is a device with geometries designed to mimic human physiology. We suspended cells in a hydrogel, allowing us to observe cell behavior as they grow into the platform. Optimizing hydrogel properties would then provide the most beneficial cell culture design. Specifically, this study investigated the influence that differing gelMA concentrations have on cell viability and morphology on F11 cells. Results include that higher concentrations—such as 15% and 30%—promoted neurite growth. This was contrasted by the 7.5% where little growth was seen. These findings showed more physiologically relevant systems with better optimized environments for cells. These results aid in the movement to reduce waste and the usage of animals in pharmaceutical research. |
| Faculty: Onabajo, Marvin Mentors: Yunfan Gao, Thomas Gourousis, Minghan Liu | Jacky Li, Emily Zhang | Automatic Calibration for Improved Suppression of Interference Signals in Wireless Receivers Presentation Poster | With the increasing number of wireless devices in our surroundings, it becomes increasingly important to design radio frequency receivers for wireless communications with enhanced capabilities to suppress unwanted interference signals. The ongoing research efforts involve the development of new devices, circuits and digital algorithms to improve the robustness of wireless receivers. As part of this goal, this research project focuses on the prototyping of an automatic calibration approach through the Arduino, based on signal power detection and the use of digital control circuits, particularly with the attenuator, to optimize the suppression of interference in analog radio frequency front-ends of wireless receivers. |
| Faculty: Shefelbine, Sandra Mentors: Lindsey Young | Leena Muntasser, Tyler Sacharow | Mechanoadaptation in Burrowing Mice Presentation Poster | The goal of this research is to characterize the differences in bone morphology between three different species of mice (P. Boylii, P. Californicus, & P. Maniculatus) based on their burrowing behavior. This research allows us to better understand how bones adapt to stress and behavior. We utilized Micro-CT imaging for accurate 3D printing of mice bones with additional refinement, then extracted data from those bones and performed the analysis. Specifically, we analyzed the Normalized Cross-Sectional Area (nCSA), the area of a slice of bone perpendicular to its length, which indicates the overall strength of the bone (the amount of force it can withstand). We also analyzed the Normalized Second Moment of Area (nIMax), which measures how the area of the bone is distributed around an axis, essentially defining the bone’s resistance to bending. After graphing the synopsis of each bone and species, we found that mouse species more intensely engaged in burrowing activity had greater nCSA and nIMax values on average and therefore showed greater bone adaptation. |
| Faculty: Su, Lili Mentors: Tongfei Guo | Rohit De, Ifechukwu Iwuchukwu | Evaluating the Effects of Uncertain Situations on Autonomous Vehicle Navigation Presentation Poster | Autonomous vehicles (AVs) are a new innovation and are capable of driving with little to no human input. They use trajectory prediction to predict the movement of objects around them, including other vehicles, using a series of sensors. However, these predictions can be jeopardized by uncertain situations, such as unorthodox (adversarial) behavior by other vehicles or changes in weather. It is crucial for AVs to be aware of these in an open-world environment in order to ensure safe navigation, but they sometimes have difficulty detecting adversarial examples. Our project intends to target this uncertainty through the creation of adversarial driving examples using the open-source CARLA simulator for autonomous driving. These examples disrupt the normal driving of AVs. They are then evaluated by a cumulative sum (CUSUM) changepoint detection algorithm, which determines if they are adversarial or not. Our project is meant to further highlight the importance of uncertainty awareness for AV navigation. |
| Faculty: Xiaoning, Jin Mentors: Guoyan Li, Zifeng Wang | Alexander Lee, Anthony Wang | Run to Run Quality Optimization for FDM Printing with Machine Learning for Manufacturing at Scale Presentation Poster | Traditional large scale manufacturing processes of plastic products inhibit customization and complexity, limiting the user experience and restricting product possibilities. Additive manufacturing, specifically Fused Deposition Modeling (FDM) printing, provides freedom in design and flexibility in fabrication–at the cost of longer production times. Within our research, we present a Bayesian optimization model that alters FDM process parameters to maximize output features of FDM processed parts within an acceptable range of efficiency, demonstrated by an optimization of the surface smoothness of a wall printed with Thermoplastic polyurethane (TPU). By demonstrating the viability of this algorithm, we hope to provide a starting point for the adoption of FDM manufacturing within the mass manufacture of plastic products in a smart manufacturing environment. |