research, programs, and other explorations I’ve been working on lately!
▸G.L.O.W.: Novel HNN Glioblastoma Localization
We developed the “Glioblastoma Localization and Optimization Workbench (G.L.O.W)” a novel neural conduit that integrates deep-learning and histone modification factors via ChIP-seq genomics and QUEST MRI, an MRI designed to visualize oxidative redox, to model cancer through a spatial/temporal approach. G.L.O.W. deciphers the spatial distribution of carcinogenic biomarkers (e.g., glycolysis enzymes for energy production) and their genetic precursors to predict the precise location of the tumor. Our model is 98% accurate for localization and carcinogenesis identification for growths scaled upwards from 13 micrometers. G.L.O.W. predicted tumor growth over 50 weeks using longitudinal data, achieving F1 scores from 0.75 to 0.91, surpassing all other clinical diagnostics. Our model also achieved mutation identification of critical tumor-related genes including causative factors for targeted therapies. G.L.O.W. enables us to locate minuscule tumors for early diagnosis while being cost-effective and time-efficient.
▸ConnectEd
Founded and led an education technology and social media startup geared toward helping students network, garnering over 300 social media followers and 10k+ users. Overseeing all operations, focused on managing finances, funding, and marketing. Check it out at useconnected.com.
▸Modeling optimal observation points for Halley's comet via a geographic approach
Using Wolfram’s built in AstronomicalData, I modeled Halley’s distance to earth in the project timeframe of its arrival by both day and minute to determine the exact time of its closest approach to Earth. Then, we will find Halley’s exact astronomical coordinates at that time and convert that to the corresponding earth coordinates perpendicularly below Halley—the closest direct distance.
▸Machine Learning Based 1muNp Selections in ICARUS/NuMI cross-section
Working with the Colorado State University High Energy Physics Laboratory under Dr. Mooney, in collaboration with Fermilab and the Stanford Linear Accelerator Center on producing Monte Carlo and data-informed selections of 1 muon N proton events in the ICARUS detector at Fermilab in the search for the hypothetical sterile neutrino.
▸Visualizing Particle Decay Chains and Analyzing Regge Trajectories
In the first half, I successfully created decay chain plots for all particles in the Wolfram database via an interconnected plot of all particles showing their cyclical decay relationships and flat directed decay graphs for each individual particle. I also implemented hypergraphs to better visualize particle decay within subgroups of products, including multi-layer decays. The effective visualizations of these decay pathways are helpful for mapping these processes, simplifying tracing particle transformations to better identify shared decay products and uncover underlying patterns in particle behavior.
In the second half, I investigated subatomic properties and quantum numbers (inferred from non-elementary particles’ quark contents) and their relationship to Regge trajectories (plots of particle mass-squared vs. spin). We were reliably informed that due to potential misnomers of particles, we could hypothesize to find additional particles in the gaps of our linear potential trajectories with the missing spin values. Within the three particle families analyzed, we did not find external additions to the Regge trajectories for ρ mesons, but potentially for Δ baryons. Preparing research for publication as well as contributing original functions to the
Wolfram Function Repository.
▸Applying transformer-based machine learning to condensed matter systems
Using neural networks to solve the Schrödinger Equation for time independent solutions of lower energy states to minimize the energy Ansatz. Applying partial differential equations for the eigenvalues of the lowest energy state to solve Fermionic wavefunctions with the Devakul Lab at Stanford University.
▸Stochastic Modeling of H2 Formation on Carbonaceous Grains
Applying novel stochastic Kinetic Monte Carlo (KMC) methods to model the formation of H2 in the early universe. Simulating formation rates, efficiency, and constants informed with observational constraints and UV-driven processes for insights into small-scale molecular interactions in an astronomical context.