Developing Algorithms to Reduce Human Bias in the Selection of Physical Parameters for Astrochemical Models

Presenter

Zachary M. Stomski

Campus

Worcester State University

Sponsor

Andrew Burkhardt, Department of Earth, Environment, and Physics, Worcester State University

Schedule

Session 1, 10:30 AM - 11:15 AM [Schedule by Time][Poster Grid for Time/Location]

Location

Poster Board A60, Campus Center Auditorium, Row 3 (A41-A60) [Poster Location Map]

Abstract

Astrochemistry is the study of how molecules form in space, and in the process of modeling these formations there is difficulty in determining the optical physical conditions for simulations. The sheer number of different parameters which need to be accounted for, along with how they're intertwined, makes it impossible for a human to get perfect results from just toying with the numbers manually. In this project, I created an algorithm that will find the best fit parameters for astrochemical kinetic simulations with the three-phase NAUTILUS code to reproduce interstellar molecular abundances from the GOTHAM (GBT Observations of TMC-1: Hunting for Aromatic Molecules) survey. TMC-1, or the Taurus Molecular Cloud, is extremely important in astrochemistry because of the cloud's vast richness. To give some scope into the matter, back in 2021, there were thirty-five new molecule detections reported; a staggering twenty-two of them were from TMC-1. To learn more about the formation of this celestial object, the neural network I made will take in human-inputted values for the abundances of different molecular species for a given observation, and then mathematically determine all the best fit conditions for its formation on its own. This was in the hope that my work can help reduce human bias and increase comparability between chemical models in publications. 

Keywords

Monte Carlo method, interstellar medium, astronomy, astrochemistry, kinetic simulations

Research Area

Astronomy, Cosmology, and Astrophysics

SIMILAR ABSTRACTS (BY KEYWORD)