Improvements to Inference Compilation for Probabilistic Programming in Large-Scale Scientific Simulators
This addresses inference challenges for scientists using large-scale, non-differentiable simulators in fields like particle physics, though it appears incremental as an extension of prior work.
The paper tackles Bayesian inference in complex scientific simulators with intractable likelihoods by extending inference compilation methods to large-scale applications, introducing a C++ library called CPProb and demonstrating its integration with the SHERPA particle physics code-base.
We consider the problem of Bayesian inference in the family of probabilistic models implicitly defined by stochastic generative models of data. In scientific fields ranging from population biology to cosmology, low-level mechanistic components are composed to create complex generative models. These models lead to intractable likelihoods and are typically non-differentiable, which poses challenges for traditional approaches to inference. We extend previous work in "inference compilation", which combines universal probabilistic programming and deep learning methods, to large-scale scientific simulators, and introduce a C++ based probabilistic programming library called CPProb. We successfully use CPProb to interface with SHERPA, a large code-base used in particle physics. Here we describe the technical innovations realized and planned for this library.