Characterizing out-of-distribution generalization of neural networks: application to the disordered Su-Schrieffer-Heeger model

TL;DR

This paper demonstrates that using interpretability techniques like CAM and PCA can enhance the out-of-distribution generalization of neural networks in classifying quantum phases, specifically applied to the disordered SSH model.

Contribution

It introduces a systematic approach combining interpretability methods to improve neural network generalization in complex quantum phase classification tasks.

Findings

CAM and PCA increase trust in neural network predictions.

Choosing neural networks that learn known phase characteristics improves out-of-distribution performance.

Disordered SSH model presents challenges for supervised neural network classification.

Abstract

Machine learning (ML) is a promising tool for the detection of phases of matter. However, ML models are also known for their black-box construction, which hinders understanding of what they learn from the data and makes their application to novel data risky. Moreover, the central challenge of ML is to ensure its good generalization abilities, i.e., good performance on data outside the training set. Here, we show how the informed use of an interpretability method called class activation mapping (CAM), and the analysis of the latent representation of the data with the principal component analysis (PCA) can increase trust in predictions of a neural network (NN) trained to classify quantum phases. In particular, we show that we can ensure better out-of-distribution generalization in the complex classification problem by choosing such an NN that, in the simplified version of the problem, learns a known characteristic of the phase. We show this on an example of the topological Su-Schrieffer-Heeger (SSH) model with and without disorder, which turned out to be surprisingly challenging for NNs trained in a supervised way. This work is an example of how the systematic use of interpretability methods can improve the performance of NNs in scientific problems.