Classification algorithms applied to structure formation simulations
This work addresses computational efficiency in cosmology simulations, though it appears incremental as it applies an existing machine learning method to a known domain problem.
The paper tackled the problem of predicting dark matter halo formation from initial conditions without running full cosmological simulations, using random forest classification to map initial density fields to halo labels with results showing the method is effective for this purpose.
Throughout cosmological simulations, the properties of the matter density field in the initial conditions have a decisive impact on the features of the structures formed today. In this paper we use a random-forest classification algorithm to infer whether or not dark matter particles, traced back to the initial conditions, would end up in dark matter halos whose masses are above some threshold. This problem might be posed as a binary classification task, where the initial conditions of the matter density field are mapped into classification labels provided by a halo finder program. Our results show that random forests are effective tools to predict the output of cosmological simulations without running the full process. These techniques might be used in the future to decrease the computational time and to explore more efficiently the effect of different dark matter/dark energy candidates on the formation of cosmological structures.