Silvia was born and raised in Summit, New Jersey, but currently resides in Palo Alto, California. She graduated from Henry M. Gunn High School and is currently working towards a Bachelor’s degree in Astrophysics and a Bachelor’s degree in Computer Science at UH Manoa. After graduating, she intends to pursue a career dedicated to research in General Relativity, Black Hole Cosmology, and Dark Matter. In her free time, Silvia loves to play the guitar, do anything outdoors, and go for a run.
Home Island: O‘ahu
High School: Henry M. Gunn High School
Institution when accepted: UH Manoa
Site: Gemini Observatory, Hilo, Hawai‘i
Mentors: Cicero Lu & Winston Wu
Project title:Â Comparing the Stellar Environments of Hot and Warm Jupiters: A Neural Network Analysis of Gaia Spectrophotometry
Project Abstract:
The discovery of thousands of exoplanets has revolutionized our view of planetary systems, but key questions remain about the stellar environments in which these planets form and evolve. Some of these planets, known as hot (<10-day orbital periods) and warm (10–200-day orbital periods) Jupiters because of their high temperatures and large radii, challenge Solar-System-based planet formation models due to their high mass and short orbital periods. To investigate how stellar properties influence giant planet formation, we analyzed a cross-matched sample of approximately 900,000 stars using data from the European Space Agency’s Gaia mission. In particular, we developed and compared three novel multimodal neural network architectures—a feedforward network, a convolutional network, and a transformer—to predict key stellar parameters: effective temperature, surface gravity, and metallicity. These parameters influence planet formation efficiency, as reflected in planetary occurrence rates. Each model’s multimodal architecture combined high- and low-resolution spectrophotometric data, including Gaia’s BP/RP (XP) and RVS spectra, to leverage complementary wavelength coverage and spectral resolution. We used the trained models to obtain improved stellar parameters for hot and warm Jupiters with varying orbital eccentricities and identified potential trends in stellar environments, such as metallicity enhancements or temperature differences, that correlate with different planet formation and migration histories. This work provides a homogeneous catalog of metallicities for all known hot and warm Jupiters discovered to date and deepens our understanding of giant planet formation.