Poster Abstracts

professors

Melt rheology in plastic upcycling

Dr. Tanyaradzwa Muzata, Assistant Professor of Chemical Engineering, Hampton University

The Polymer Processing Lab under Dr. Muzata is currently developing antimicrobial polymer coatings, which can potentially be used to develop PPE for healthcare workers in the fight against COVID-19 and other future pandemics. The antimicrobial polymeric coatings can also be used in packaging applications. The lab is also working on developing polymer-based composite materials that attenuate electromagnetic radiation pollution. The rapid increase in the production of different numbers of electronic gadgets has greatly benefited humanity. However, one of the shortfalls in developing these gadgets is the production of electromagnetic pollution, leading to electromagnetic interference (EMI). Electromagnetic pollution causes an interruption in the normal functioning of different electronic devices, and it has been reported that it has adverse effects on human health. Designing and developing polymer-based EMI shielding materials made of polymeric materials derived from virgin and post-consumer recycled plastics is paramount to mitigate this pollution.

postdoctoral researchers

Tracing Accelerated Galaxy Evolution in Protoclusters with Cosmological Simulations

Dr. Devontae Baxter, Postdoctoral Researcher in astronomy & astrophysics, University of California San Diego

The protocluster stage of galaxy cluster assembly remains an underexplored frontier. This is evident in the literature, where only ~60 spectroscopically confirmed protoclusters have been identified. In addition to being sparse, this sample suffers from diverse selection functions, resulting in a highly heterogeneous dataset that evades meaningful statistical analysis. Fundamental questions — such as the role of the protocluster environment in driving galaxy evolution — remain challenging to test without large, homogeneously selected samples. Fortunately, facilities such as Euclid and JWST, along with upcoming observatories like Roman and LSST, will together provide the necessary field of view, deep photometry, and high-resolution spectroscopy to generate large, uniformly selected protocluster samples. In preparation for this data, I present an investigation of protocluster environments and their impact on galaxy evolution using the TNG-Cluster simulation, the highest-resolution large-box cosmological simulation of its kind. Specifically, I analyze a population of 352 protoclusters to determine whether the densest regions of galaxy protoclusters are associated with accelerated galaxy evolution, as some observations suggest. I also demonstrate that most spectroscopically confirmed protoclusters probe only a fraction of the total volume occupied by protoclusters. Furthermore, I test whether observationally selected protocluster galaxies — limited by SFR and/or stellar mass — can reliably identify the true highest-density regions as characterized by the full protocluster galaxy population. Finally, I establish the completeness limits required to identify these regions. Beyond shedding light on how protocluster environments influence galaxy evolution, this work lays the foundation for interpreting current protocluster samples and making predictions for future galaxy protocluster observations.

Advancing Stellar Activity and Binary Evolution Science with Red Giant Branch Stars

Dr. Don Dixon, Postdoctoral Researcher in astronomy, Vanderbilt University

The rotational dependence of dynamo-powered stellar activity has been thoroughly investigated for main-sequence stars, but much less so for stars evolved onto the giant branch. In addition to being bigger and brighter, giant stars are generally observed to have negligible rotation rates compared to main-sequence stars and consequently tend to have comparatively weak dynamo action and global magnetic fields. This is generally expected due to angular momentum being carried away by stellar winds before ascending the red giant branch, but if a companion star orbits in close enough proximity, a red giant star can be spun up by tidal synchronization or mass accretion events. As a result, rotationally active giants are unique systems for furthering the understanding of evolved stellar activity and binary evolution processes that impart angular momentum for rotational spin-up. To characterize giant branch rotational activity, we derive an empirical rotation-activity relation using projected rotational velocities and excess near-ultraviolet emission produced by magnetic heating of the chromosphere. We find our relation is similar in structure to those fitted to main-sequence stars, which may suggest a similar dynamo mechanism. Additionally, we find that the most active giants have near-ultraviolet emission that exceeds expectations given their rotation rates and that they are all in synchronized binaries with orbital periods less than twenty days. Lastly, we find tentative evidence that giants approaching breakup rotational velocities are likely formed from planetary engulfment or stellar merger processes, both of which are still not well understood.

graduate students

Quantum free will… but at what cost?

Jae Eniraiyetan, PhD Student in Philosophy, University of California San Diego

It is often accepted that Quantum Mechanics (QM) describes an indeterministic world, particularly in situations describing entangled states. QM describes a world where the laws of nature and past events do not uniquely determine the outcome. But does quantum indeterminism have any consequences for the famous free will debate, and in particular, the position known as compatibility, which holds that free will and determinism are compatible? This paper considers the potential implications of quantum indeterminism on the problem of free will. Centring on Arthur Fine’s 1993 paper “Indeterminism and the Freedom of the Will”, I investigate how Fine’s distinct approach to the discussion of quantum free will introduces an innovative concept, which I call ‘local libertarian’. This new concept defines free will, where the libertarian requirement of having the ‘option to do otherwise’ is facilitated by entangled states, where entities do not have definite states. The strength and philosophical challenges of local libertarianism are considered, and it is shown that a local libertarian view of free will comes at a high cost to our preconceived notions of free will and measurement in an indeterministic world governed by QM.

Development of a Machine Learning-Based Predictive Modeling Framework for MoltenSalt Reactor Sensor Data Analysis

Omobolade Odedoyin, MSE Candidate in Data Science, Johns Hopkins University

Molten Salt Reactors (MSRs) offer groundbreaking advantages in nuclear energy including enhanced thermal efficiency and safety. However the extreme chemical and operationalenvironments of MSRs present challenges such as sensor degradation and corrosion whichhinder accurate monitoring of electrochemical behavior. Addressing these challenges is critical for advancing MSR technology ensuring safe reactor operations and supporting applications such as medical isotope production. Can machine learning (ML) models combined with advanced feature engineering and optimization improve the predictive accuracy of key electrochemical parameters such as peak current derived from MSR sensor data? In thepresent study we analyzed Cyclic Voltammetry (CV) and Open-Circuit Potentiometry (OCP)data for predictive model development focusing on features such as cycle direction scan rateapplied voltage resulting current and molten salt solute concentration. Qualitative featuressuch as compound names of reactor electrodes and molten salts were transformed into unifiednumeric values using a function that integrates normalized molar mass and ionization energy.Bayesian optimization identified the optimal weight combinations for these transformationsmaximizing correlations with the target variable (current). Recursive Feature Elimination withCross-Validation (RFECV) was applied for feature selection while nested cross-validationoptimized regressors and hyperparameters. The resulting supervised models were trained andevaluated using mean absolute error (MAE) and correlation metrics. Our approach identified optimal weight combinations to transform qualitative features and selected the most relevant features yielding predictive models with significantly improved accuracy compared to baseline methods. The best-performing configuration used a Random Forest Regressor with optimizedhyperparameters and optimal weight combinations for transforming qualitative features. Theoptimal feature selection included all features available with the optimal weight for qualitativefeature transformation. This best model achieved a test MAE of 0.00441 a significantimprovement over the baseline mean MAE of 0.02559. Improved correlations between engineered features and peak current validated the methodology. These results demonstratethe potential of ML-driven frameworks to reliably predict electrochemical behavior underextreme MSR conditions. This research demonstrates the effectiveness of integrating MLtechniques advanced feature engineering and Bayesian optimization to address challenges inmonitoring electrochemical behavior and MSR sensor data analysis. The robust predictive modeling framework significantly enhances accuracy and reliability in predicting electrochemical behaviors supporting safer reactor operations and paving the way for broader applications such as medical isotope production and sustainable nuclear energy development.These findings underscore the potential of ML-driven approaches to revolutionize monitoring and optimization in high-stakes nuclear systems.

The nematicide effect of chalcones in Caenorhabditis elegans do not involve the oxidative stress or neuropeptide pathways of the nematode

Keyora Wharry, MS Candidate in Biotechnology, California State University Fresno

Plant-parasitic nematodes (PPNs) pose a significant obstacle to global farming, and the current solutions to control PPNs are very harmful pesticides. Dr. Calderón-Urrea’s laboratory developed an environmentally sustainable method to control PPNs using chalcones; Chalcone 17, Chalcone 25, and Chalcone 30 were found to be effective nematicides at concentration of 10-4 M. Furthermore, they are likely to induce death in C. elegans by targeting different, yet unknown, molecular pathways. Two possibilities are that these chalcones exert their influence via the oxidative stress or the neuropeptide pathways. These are the two pathways that other nematicides exert their effects on nematodes; my research focuses on testing if these chalcones exert their effects through either of these two pathways. I use Reverse Transcription-Quantitative Polymerase Chain Reaction (RT-qPCR) to monitor the gene expression of the following oxidative stress pathway genes skn-1, sod-3, daf-16, ctl-1, and gcs-1, and the following neuropeptide signaling pathway genes egl-3, egl-21, and kpc-1. RNA was extracted from two life stages of the nematodes (L1 and L4 larvae) to study gene expression at different developmental stages, and the RNA was converted into cDNA to analyze gene expression in nematodes exposed to the three chalcones for 1 hour, 2 hours, and 3 hours. Our results show that the daf-16 and ctl-1 are affected by the exposure to chalcones, although the effect in the ctl-1 gene expression is minor. Although the daf-16 gene plays a minor role in oxidative stress, it is involved in several processes such as regulation of dauer larval development and regulation of metabolic processes, and it is expressed in several structures, including germ cells, gonads, hypodermis, neurons, and somatic cell. Our results combined point to the lack of involvement of the oxidative stress and the neuropeptide pathways as a mechanism of action (MOA) of the chalcones.

From Stories to Action: Liberatory Computing and Youth Data Activism for Social Justice

Raechel Walker, PhD Student in Media Arts and Sciences, Massachusetts Institute of Technology

Computing curricula often inadvertently reinforce a harmful, singular narrative about African American communities, focusing solely on narratives that emphasize crime prediction and policing. This reinforces the harmful stereotype that African American communities are primarily sites of criminal activity rather than centers of innovation, creativity, and resilience. In contrast, “liberatory computing” offers a framework that can be integrated into computing curricula precisely to counter these cliches. Composed of five pillars–a sound racial identity, critical consciousness, collective obligation, a liberation-centered academic identity, and activism skills—liberatory computing empowers students to challenge and mitigate systemic oppression through informatics. This research applies this framework as a way to empower African American students to address embedded racism through data analysis. We particularly focus our curriculum on data activism, evident in two related projects. The first taught students how to use data science to support minoritized communities, while the second incorporated collaboration with community organizers, increasing the inclusion of desire-based research.

Our first program included 12 high school students of color, while the second program engaged 24 high school students of African American descent. In the second program, these students partnered with community organizations in the Greater Boston area for a range of data activism initiatives. These projects encompassed data analysis, geospatial analysis, qualitative analysis, surveys, interviews, and artistic expression. The student surveys revealed heightened awareness of data science's role in combating racism and enhanced proficiency in promoting racial justice. Interviews with the students further revealed that mitigating systemic oppression through their data activism projects was a pivotal aspect that motivated them to persist in integrating data activism into their future pursuits. Integrating community researchers into the technical curriculum empowers students to infuse data science projects with personal narratives, breaking away from the conventional singular narrative. The expertise of the community partners in topics related to justice, such as housing insecurity and environmental injustice, allowed them to express a nuanced understanding of their experiences and identities. The community partners, in turn, intend to use the students' projects for advocacy purposes, such as advocating for policies addressing flooding in African American and low-income Boston communities using data visualizations. The implications of this research demonstrate how African American students can be empowered to utilize data science in order to catalyze societal transformation, helping to fill the gap in the availability of computing curricula tailored to empower African American students to apply their computing skills for the betterment of their communities. This is achieved by fostering opportunities for them to apply their data science skills to tangible real-life issues through collaborations with community organizations addressing systemic challenges.

Modeling Joint Cartilage Regeneration in Zebrafish

Oluchi Ofoegbu, PhD Candidate in Stem Cell Biology and Regenerative Medicine,

University of Southern California

Osteoarthritis (OA) is the leading cause of disability in the US, affecting 58 million people. OA is characterized by the irreversible deterioration of the specialized cartilage lining joints. Mammals, including humans, cannot regenerate their joint cartilage. Zebrafish are a common vertebrate model system with many skeletal development mechanisms conserved between humans but maintain regenerative capabilities into adult stages. The Crump lab has recently shown that adult zebrafish can regenerate the synovial jaw joint after injury. Uncovering the mechanism of joint cartilage regeneration in zebrafish may aid in understanding the lack of regenerative ability in mammals. Through single-cell multiomics, we have identified an enhancer that drives joint-specific transgene expression in zebrafish. Utilizing this enhancer, I have driven the joint-specific expression of the bacterial nitroreductase enzyme (NTR) to selectively ablate the articular cartilage and observe regeneration. When the prodrug metronidazole (MTZ) is added to the tank water, NTR-expressing cells convert MTZ to a toxin, leading to cell-specific ablation. Preliminarily, I find that the addition of MTZ to these fish ablates a subset of joint chondrocytes which then rapidly regenerate. I am currently working to optimize my system for complete ablation of joint chondrocytes, as well as to create new lineage tracing zebrafish lines to determine the cell populations mediating regeneration. The results of this study will give insight into the progenitor populations that mediate joint cartilage regeneration in fish, which will inform future efforts to stimulate the natural regeneration of joints in osteoarthritic patients.

Electrochemical Hydrosilylation of Olefins using First-Row

Transition Metals

Khamyl Cooksey, PhD Student in Chemistry and Biochemistry, University of California Merced

Hydrosilylation is a pivotal transformation in synthetic chemistry, enabling the addition of siliconhydrogen (Si–H) bonds across unsaturated substrates. This reaction is widely utilized in the synthesis of organosilicon compounds, which serve as key intermediates in pharmaceuticals, agrochemicals, and materials science. Traditionally, hydrosilylation relies on metal catalysts, most commonly third-row transition metals such as platinum and rhodium, due to their efficiency, regioselectivity, and functional group tolerance. However, the widespread application of these catalysts is often hindered by the high cost, limited availability, and toxicity of noble metals, as well as the requirement for harsh reaction conditions, including high temperatures and excess reducing agents. As the demand for more sustainable and environmentally benign methodologies grows, electrochemical hydrosilylation has emerged as a promising alternative. This strategy harnesses electrical energy as a driving force, reducing the need for stoichiometric chemical reductants and enabling milder reaction conditions. Moreover, the use of first-row transition metal catalysts, which are more earth-abundant and cost-effective, offers a greener and more scalable solution for industrial applications. In this work, we demonstrate the electrochemical hydrosilylation of olefins using a first-row transition metal catalyst and an organosilane under ambient conditions. By leveraging electrochemical techniques, we achieve selective and efficient Si–H bond addition without the drawbacks associated with conventional methods. Our findings contribute to the growing field of electrocatalysis and highlight the potential of electrochemical strategies in advancing sustainable synthetic methodologies. This approach not only provides a viable alternative to traditional hydrosilylation but also opens new avenues for electrochemically driven transformations in organic synthesis.

Morphological And phylogenetic Analysis of A Paleogene Typotheria from Cachapoal

Preston Dozie Rukewe Willard Itie, Paleontologist, Industry Professional

South America for most of the Cenozoic period was isolated. Due to its isolation, this led to the mammals of this period being endemic. This period has been referred to as the South American Land Mammal Ages or SALMA. A recently established SALMA, the Tinguirirican may possibly have a bigger geological range. A river valley that approaches the Tinguirirican,the Cachapoal River Valley, has been hypothesized to be part of the SALMA and could extend its geological range. To answer this question, I describe a fossil from the Cachapoal River Valley to determine whether the Cachapoal River Valley is Tinguirirican in age. Using my description, I compared the characters I identified with other specimens and used phylogenetics to determine the species. The results of my studies indicate that the fossil is a sister group with Protypotherium Australe and Miocochilius. The geological ranges of these taxa make them younger than the Tinguirirican and indicate that the Cachapoal River Valley is not of the Tinguirirican age.

undergraduate students

Scalable Community Mentorship:

A Vision for Engineering Literacy & Access

Samira Shirazy, Undergraduate in Human Centered Design and Eng., University of Washington

Black and Hispanic/Latine communities remain underrepresented in engineering careers due to persistent economic, social justice, and human rights issues. Consequently, students from these communities face unique challenges when entering into engineering fields. Barriers stem from economic disenfranchisement, racial profiling, lack of community, and an absence of cultural competence. However, many programs that aim to tackle these issues are not holistic nor scalable to the degree needed for broader impact. To address the lack of holistic and scalable mentorship programs for Black and Hispanic/Latine students, we analyzed AVELA - A Vision for Engineering Literacy & Access' (AVELA) holistic student-led STEM engagement model. This model leverages multi-tier near-peer mentorship, mentor-embodied community representation, culturally responsive experiential & service-based learning, and compensated student-led community engagement. Thus far, AVELA has supported 361 university student instructors in teaching to 4,412 secondary school students across 213 classrooms. To evaluate members' motivations for participating in this mentorship program, we conducted 24 semi-structured interviews with undergraduate and graduate AVELA members. We find AVELA's holistic student-led STEM engagement model improves Black and Hispanic/Latine engineering students' experience in post-secondary education. Reduced power dynamics, community involvement, soft skill development, as well as financial support and technical skill building were indicated as key motivators for students' participation in the program. Students expressed that the depth, applicability, and scale of their mentorship relationships cultivated a reliable student community that empowered them to build a robust professional skill set.

Auditing Diversity in LLM (Language Learning Models) and

TTI (Text-to-Image) Systems

ndeye diop, Undergraduate in informatics, University of washington

This research project focuses on developing and enhancing an AI auditing system to assess diversity and fairness in large language modeling (LLMs) systems. By replicating an existing Python-based audit framework, originally created by my Principal Investigator (PI), this study extends its functionality to specifically evaluate how race and ethnicity are represented in AI-generated outputs related to professional occupations. The enhanced auditing system cross-references race and ethnicity data with job positions to identify potential biases, providing a deeper understanding of whether AI systems (specifically GPT-4) disproportionately associate certain ethnic groups with specific professions. These findings contribute to the ongoing discourse on fairness in AI, offering insights into how LLM models may perpetuate or mitigate biases in career representation. This research is critical for the development of more equitable AI systems that reflect diversity across various social and professional contexts, highlighting the importance of fairness in the deployment and usage of AI technology.

Assessing associations between race, discrimination, and emergency room use in California Adults

Jaylene Gardner-Perales, Undergraduate in Biological Sciences, University of California Irvine

Avoidable emergency department utilization results in ED overcrowding, increased medical bills, and poorer quality of care. This issue disproportionately affects BIPOC patients, as they have a higher prevalence of visiting the ED for preventable cases. Our study seeks to investigate potential factors that may influence patients to visit the ED for these cases, rather than visiting their primary care offices for preventative care. In our study, approximately 1300 patients from the UCI Emergency Department completed questionnaires. Specific questions pertained to race, insurance coverage, and provider referrals. These factors will be analyzed to discover their influence on the type of emergency department visit -- preventable or non-preventable.

The Role of Computational Models in Understanding Human Perception:

AI vs. Human Vision

Joyce Nimely, Undergraduate in computer science, berea college

The ability to perceive and interpret visual information is a fundamental aspect of human cognition. In recent years, artificial intelligence (AI) has made significant strides in replicating aspects of human vision through computational models, particularly convolutional neural networks (CNNs) and vision transformers. However, despite their advancements, these models differ fundamentally from biological vision in terms of context awareness, generalization, and robustness.

This literature review examines the key mechanisms of human vision, including feature detection, depth perception, and attention, and compares them to computational approaches in AI. Through an analysis of existing research in neuroscience, computer vision, and artificial intelligence, this review explores the strengths and limitations of AI-driven perception models. Key topics include the differences in learning processes, the impact of adversarial attacks on AI vision, and the ethical concerns surrounding bias in AI-powered visual recognition systems.

By synthesizing findings from interdisciplinary studies, this review highlights the gaps between human and artificial perception and discusses potential pathways for bridging these differences through advancements in neuromorphic computing and interdisciplinary research. This work contributes to the ongoing discourse on the intersection of AI and human cognition, emphasizing the importance of incorporating insights from neuroscience to improve the next generation of AI-driven vision systems.

The Long-COVID Transcriptomic Landscape: Uncovering Genetic Bases

Grace Nwankwo, Undergraduate in Health & Human Sciences, University of Southern California

Long COVID, characterized by symptoms like shortness of breath, pulmonary scarring, muscle aches, and cough persisting three months post-COVID-19 infection, shares similar clinical characteristics with interstitial lung disease (ILD) despite different etiologies. Transcriptomics, the study of RNA and active gene expression, has proven valuable in identifying genetic factors in disease progression via previous studies looking into patterns in gene expression driving the development of ILD symptoms. However, the specific genetic risk markers for long COVID remain largely unexplored. This translational research study investigated the transcriptomic profiles of patients diagnosed with long COVID, comparing them with the profiles of patients diagnosed with ILD to discover if any differential gene expression was linked to developing long COVID. Using transcriptomic analysis via blood samples from participants, seven upregulated genes and four downregulated genes were identified to be differentially expressed in patients with long COVID. The most differentially expressed gene, MYOM2, and the most significantly upregulated gene, SRP9P1, are linked to disruptions in muscle cell repair and sustained inflammation in the lung and heart, telltale of long COVID symptoms. Additionally, the downregulated genes were associated with improved alveolar structure repair, suggesting less severe fibrotic damage in long COVID compared to ILD. These findings bring to light a potential link between genetic risk factors and long COVID symptoms, offering insights into predicting, diagnosing, and developing treatments for long COVID, given our knowledge about its genetic landscape within the transcriptome.

Evaluation of the Colorfastness of Metal-Complexable Monoazo

Dyes for Human Hair

Kennedy Walker, Undergraduate in Chemical Engineering, Pennsylvania State University

Hair dyes are widely used products among consumers desiring to change the color of their hair, with permanent hair dyes being the most popular and ideal choice. However, the chemical components of these dyes can pose risks to the health of both humans and the environment. Key precursors such as para-phenylenediamine (PPD) and resorcinol are potent skin allergens and can infiltrate aquatic ecosystems. These precursors undergo oxidative coupling, forming high-molecular-weight dye molecules within the hair. This makes it difficult for them to desorb from the hair shaft during washing- hence their classification as “permanent” dyes. To explore more sustainable alternatives to conventional permanent hair dyes, this study evaluated the colorfastness (color retention) of metal complexable arylide and arylazonaphthol monoazo dyes These dyes can be complexed within hair fibers at 40°C using environmentally benign ions (Al3+ and Fe2+). Using dyed human hair samples, a series of color fastness tests were conducted to evaluate the color properties of dyed hair samples when exposed to daily external factors that human hair fibers endure which include washing, UV light exposure, and abrasion. The samples were assessed via visual color assessments utilizing the AATCC Gray Scales for Color Change and Staining, as well as quantitative color measurements utilizing a UV-Vis spectrophotometer. Results indicated that hair samples treated with complexed dyes exhibited superior wash, light, and rubbing fastness, whereas those dyed with uncomplexed dyes showed a lower degree of color retention. Among all tested dyes, hair samples treated with Fe-arylazonaphthol dye exhibited the highest resistance to color change, highlighting its potential as a viable alternative to conventional permanent hair dyes.

Acknowledgement – This project was funded by the TECS (Textile Engineering, Chemistry, and Science) REU program hosted by the Wilson College of Textiles at NC State University.

Page updated

Google Sites

Report abuse