Exploiting Network Topology for Accelerated Bayesian Inference of Grain Surface Reaction Networks

TL;DR

This paper introduces topology-based methods to efficiently perform Bayesian inference on complex grain-surface reaction networks in interstellar chemistry, significantly reducing computational costs while maintaining accurate reaction rate estimates.

Contribution

The authors present two novel topology-exploiting methods to accelerate Bayesian inference in large reaction networks, enabling analysis of complex interstellar molecules.

Findings

Methods recover maximum-posterior reaction rates with minimal bias.

Network separation into sub-networks facilitates computational efficiency.

Applicable to large, complex organic molecule networks.

Abstract

In the study of grain-surface chemistry in the interstellar medium, there exists much uncertainty regarding the reaction mechanisms with few constraints on the abundances of grain-surface molecules. Bayesian inference can be performed to determine the likely reaction rates. In this work, we consider methods for reducing the computational expense of performing Bayesian inference on a reaction network by looking at the geometry of the network. Two methods of exploiting the topology of the reaction network are presented. One involves reducing a reaction network to just the reaction chains with constraints on them. After this, new constraints are added to the reaction network and it is shown that one can separate this new reaction network into sub-networks. The fact that networks can be separated into sub-networks is particularly important for the reaction networks of interstellar complex organic molecules, whose surface reaction networks may have hundreds of reactions. Both methods allow the maximum-posterior reaction rate to be recovered with minimal bias.