An informal seminar series to highlight work from early career researchers from institutions around the world.
Title: COSMIC: Characterization of Star clusters using Machine-learning Inference and Clustering
Abstract: COSMIC (Characterization of Star clusters using Machine-Learning Inference and Clustering) is a software tool developed to semi-automate the analysis of open clusters by leveraging machine learning and Bayesian inference. It employs HDBSCAN for clustering and PyMC with the No-U-Turn Sampler (NUTS), an advanced Hamiltonian Monte Carlo extension, for Bayesian modeling. The software is designed to determine membership, kinematic, dynamic, and structural parameters of open clusters using the extensive datasets from Gaia DR3. The creation of COSMIC aims to facilitate the handling of large data volumes from Gaia and other observatories such as the incoming Vera Rubin Observatory. Currently, COSMIC is in the development and validation stage, being tested on both well-characterized open clusters and synthetic clusters, and is applied to the analysis of the young open cluster NGC 6383. As an open-source Python code, COSMIC allows for continuous improvement by the scientific community.
Zoom link: https://uwmadison.zoom.us/j/96573094371?pwd=ZkYzemwwQjFTcHYyTzd0eDVUUVl5Zz09
Organized by: Leon Trapman, Dan Rybarczyk, Nickolas Pingel