Poster
in
Workshop: AI for Science: Scaling in AI for Scientific Discovery
Accelerating statistical inferences in astrophysics with Neural Networks and Hamiltonian Monte Carlo
Diego Gonzalez-Hernandez · Molly Wolfson · Joseph Hennawi
Keywords: [ Astrophysics ] [ emulators ] [ Hamiltonian Monte Carlo ] [ Epoch of Reionization ]
We present an approach to accelerate statistical inferences in astrophysics by using a combination of neural networks and Hamiltonian Monte Carlo. The neural networks are used to create high-fidelity surrogates of computationally expensive models, while Hamiltonian Monte Carlo accelerates the inferences by more efficiently exploring the parameter space. We demonstrate the potential of this approach by applying it to a realistic model for the Epoch of Reionization.