Our research is focused on determining the dynamical parameters associated with galaxy collisions that cannot be obtained via direct observation, such as orbital velocities, orientations, and mass-ratios between the galaxies. Beyond understanding how dynamics affects star formation, galaxy evolution, and AGN activity, we wish to understand how tightly the dynamical history of interacting systems can be constrained based on observable quantities. We are accomplishing this by fitting numerical models of galactic interactions with astronomical observations of colliding galaxies and measuring the uncertainty in the parameter space of these models. To optimize our numerical models, our research group is working on two tasks. First, we develop computational methods for scoring the similarity between a galactic model and observed target galaxies. Second, we find the best models by optimizing the parameters that characterize this nonlinear system using the machine scoring method above. This poster will focus on how we compare our numerical models to observed colliding galaxies. To do so, gravitational n-body models are generated that capture the tidal distortions from galaxy collisions. The final particle positions are then processed to create model images that have realistic intensity profiles and resolutions. We compare the model images with the astronomical images to obtain an objective machine score that should reflect the fitness of the model. In order to test our scoring methods, we then compare our machine scores to the human scores obtained from the citizen science project Galaxy Zoo: Mergers.