• Publication: Submitted to the Publications of the Astronomical Society of the Pacific (PASP) under peer-review; Supervisors: Prof. Christopher A. Theissen

Spectral typing has been providing invaluable insights into the intrinsic properties of stars. With the advent of the Gaia mission, we have been presented with a treasure trove of data, which when synergized with existing datasets, can offer unprecedented clarity and depth. This project sought to harness the rich data from Gaia DR3, marrying it with the established benchmarks from the Sloan Digital Sky Survey, to forge a novel set of low-resolution spectral templates These templates encompass a sample of 37,945 stellar spectra after filtering. 

Specifically, in the range of 3,360 - 10,200 Angstroms, the collection includes templates for main-sequence stars across the OBAFGKM spectrum, each with subcategories from 0 to 9. Additionally, templates for white dwarfs (ranging from DA0.5 to DA7.0) and carbon stars (including dCG, dCK, and dCM types) are provided in the same wavelength range. Designed with precision and depth, these templates are aimed at supporting the astronomical community in their stellar classification endeavors, ensuring both accuracy and comprehensiveness in spectral analyses and related tasks.

• Conducted extensive spectral analysis on 219 million low-resolution spectra from Gaia DR3. Developed and calibrated 201 main-sequence spectral templates across diverse categories, grounded in spectral standards from the Sloan Digital Sky Survey (SDSS).

• Prepared and purified a substantial sample of 117,966 stellar spectra, ensuring sample purity and resulting in a cleaned and refined sample comprising 37,945 spectra.

• Innovated a weighted median approach to generate spectral templates, expanding the application scope of spectral typing methodologies. Provided foundational tools for precise spectral typing and stellar property analysis, advancing future research in astrophysics.

Github Repo   Poster 

• Journal: A Prestigious Astrophysical Journal under peer-review

• Supervisors: Tianxing Zhou and Prof. Christopher A. Theissen, etc.

During my senior year, I had the privilege of contributing to a pivotal research project in the realm of astrophysics, which culminated in the development of the study "Classifying Cool Dwarfs: A Comprehensive Study of Field and Peculiar Dwarfs Using Machine Learning." This paper is slated for submission to a reputable peer-reviewed journal in the field of astrophysics, underscoring its academic rigor and contribution to the scientific community.  This research project introduces a groundbreaking Machine-Learning model to classify very-low-mass stars and brown dwarfs, specifically those within the spectral types M0 and later. These objects are integral to the comprehensive understanding of stellar and planetary characteristics and demographics within the near-infrared wavelength range (0.85-2.45 microns).

The overarching goal of this research was to illuminate the relative importance of diverse spectral regions in ascertaining gravity and age, thereby paving the path for more refined and precise classification methodologies in the field of astrophysics. The insights derived from this research have the potential to significantly impact the broader scientific community, providing a robust framework for future investigative endeavors in the field.