Deep learning insights into cosmological structure formation

While the evolution of linear initial conditions present in the early universe into extended halos of dark matter at late times can be computed using cosmological simulations, a theoretical understanding of this complex process remains elusive. Here, we build a deep learning framework to learn this non-linear relationship, and develop techniques to physically interpret the learnt mapping. A three-dimensional convolutional neural network (CNN) is trained to predict the mass of dark matter halos from the initial conditions. N-body simulations follow the microphysical laws of gravity, whereas the CNN model provides a simplified description of halo collapse where features are extracted from the initial conditions through convolutions and combined in a non-linear way to provide a halo mass prediction. We find no significant change in the predictive accuracy of the model if we retrain it removing anisotropic information from the inputs, suggesting that the features learnt by the CNN are equivalent to spherical averages over the initial conditions. Despite including all possible feature combinations that can be extracted by convolutions in the model, the final halo mass predictions do not depend on anisotropic aspects of the initial conditions. Our results indicate that deep learning frameworks can provide a powerful tool for extracting physical insight into cosmological structure formation.