Tasking Citizen Scientists from Galaxy Zoo to Model Galaxy Collisions

摘要

The huge data volumes across the sciences require us to consider new approaches to data analysis and simulations. In this project, we investigate how citizen scientists from CalaxyZoo.org can be tasked to address this data flood by providing the human input to classify simulations of galaxy interactions. The primary science goal in the galaxy collision project is developing numerical models for hundreds of interacting systems. With these simulations, we will determine if these collisions are unique locations in the multi-parameter phase space that defines the input parameters to our simulations. We will also correlate these dynamical parameters with the star formation and nuclear activity in these systems. Further self-consistent, high resolution models will follow from these simple matches to compare simulations directly with the observations. A secondary goal of the galaxy collision project is to create a set of human-selected matches between simulations and models which can he used as a training set for Machine learning algorithms. Analysis of this data set will allow the creation of a robust fitness function for evolutionary computing algorithms that can be used to model other interacting systems.