Variance Tolerance Factors For Interpreting ALL Neural Networks
This work addresses the need for interpretability in neural networks for applications like laboratory experiments and model defense, though it appears incremental by building on existing concepts like influence functions and Rashomon sets.
The authors tackled the problem of interpreting black-box neural networks by proposing a variance tolerance factor (VTF) to rank feature importance, and they applied it to synthetic, benchmark, and real-world datasets, including predicting noncrystalline gold nanoparticle formation and chemical toxicity in aromatic compounds.
Black box models only provide results for deep learning tasks, and lack informative details about how these results were obtained. Knowing how input variables are related to outputs, in addition to why they are related, can be critical to translating predictions into laboratory experiments, or defending a model prediction under scrutiny. In this paper, we propose a general theory that defines a variance tolerance factor (VTF) inspired by influence function, to interpret features in the context of black box neural networks by ranking the importance of features, and construct a novel architecture consisting of a base model and feature model to explore the feature importance in a Rashomon set that contains all well-performing neural networks. Two feature importance ranking methods in the Rashomon set and a feature selection method based on the VTF are created and explored. A thorough evaluation on synthetic and benchmark datasets is provided, and the method is applied to two real world examples predicting the formation of noncrystalline gold nanoparticles and the chemical toxicity 1793 aromatic compounds exposed to a protozoan ciliate for 40 hours.