# The Good, the Bad and the Ugly: Augmenting a black-box model with expert knowledge

## TL;DR

This paper explores augmenting black-box neural network models with expert knowledge at different levels to resolve parameter ambiguities in material science, benchmarking their performance against pure black-box models.

## Contribution

It introduces a method to incorporate expert knowledge into neural networks to improve parameter fitting in material science applications.

## Key findings

- Expert-augmented models outperform pure black-box models in resolving ambiguities.
- Different levels of expert knowledge provide varying degrees of improvement.
- Benchmark results demonstrate the effectiveness of the proposed augmentation approach.

## Abstract

We address a non-unique parameter fitting problem in the context of material science. In particular, we propose to resolve ambiguities in parameter space by augmenting a black-box artificial neural network (ANN) model with two different levels of expert knowledge and benchmark them against a pure black-box model.