Abstract by Jacob Nuttall
Physics and Astronomy
Improving Methods of Understanding Complex Models
Recent discoveries in predictive modeling have resulted in powerful theories to optimize and understand the space of models of systems. One of these theories, the Model Boundary Approximation Method (MBAM), involves representing a model as a mapping from parameters to predictions and traversing the surface of the model's topology to find simpler models with reduced parameters. However, the process of applying MBAM leads to the problem of how to organize and interpret the resulting information in order to explore the space of these models more completely. We discuss computational tools that we anticipate will provide solutions to this problem.