Uncertainty Propagation within Chained Models for Machine Learning Reconstruction of Neutrino-LAr Interactions

TL;DR

This paper explores how to propagate uncertainty through chained machine learning models in scientific applications, focusing on neutrino-Argon interaction reconstruction to improve inference reliability.

Contribution

It introduces a method for adapting a model within a chain to incorporate input uncertainties, enhancing uncertainty quantification in neutrino interaction reconstruction.

Findings

Uncertainty-aware models outperform uncertainty-blinded models in inference quality.

Synthetic noise testing demonstrates the benefits of propagating upstream uncertainties.

The approach improves confidence estimation in neutrino physics experiments.

Abstract

Sequential or chained models are increasingly prevalent in machine learning for scientific applications, due to their flexibility and ease of development. Chained models are particularly useful when a task is separable into distinct steps with a hierarchy of meaningful intermediate representations. In reliability-critical tasks, it is important to quantify the confidence of model inferences. However, chained models pose an additional challenge for uncertainty quantification, especially when input uncertainties need to be propagated. In such cases, a fully uncertainty-aware chain of models is required, where each step accepts a probability distribution over the input space, and produces a probability distribution over the output space. In this work, we present a case study for adapting a single model within an existing chain, designed for reconstruction within neutrino-Argon interactions, developed for neutrino oscillation experiments such as MicroBooNE, ICARUS, and the future DUNE experiment. We test the performance of an input uncertainty-enabled model against an uncertainty-blinded model using a method for generating synthetic noise. By comparing these two, we assess the increase in inference quality achieved by exposing models to upstream uncertainty estimates.