Characterizing out-of-distribution generalization of neural networks: application to the disordered Su–Schrieffer–Heeger model
Open Access
- 22 January 2025
- journal article
- research article
- Published by IOP Publishing in Machine Learning: Science and Technology
- Vol. 6 (1) , 015014
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, and the analysis of the latent representation of the data with the principal component analysis 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 (OOD) 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 also discuss the characteristics of the data representation learned by a network that are predictors of its good OOD generalization. We show this on an example of the topological Su–Schrieffer–Heeger 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.Keywords
All Related Versions
Funding Information
- Institut de Ciències Fotòniques (Internal “QuantumGaudi” project)
- Polish Ministry of Science and Higher Education (Excellence initiative – research university)
- 'la Caixa’ Foundation (ID 100010434)
- Narodowe Centrum Nauki (2020/36/T/ST2/00588)
- European Union (101113690)
- Ministry for Digital Transformation and of Civil Service of the Spanish Government (QUANTUM ENIA project call - Quantum Spain project)
- Fundació Mir-Puig
- Fundació Cellex
- Infrastruktura PL-Grid (PLG/2023/016878)
- HORIZON EUROPE European Innovation Council (No 101080086 NeQSTGrant Agreement 101080086 — Ne)
- H2020 European Research Council (AdG NOQIA)
- Barcelona Supercomputing Center MareNostrum (FI-2023-3-0024)
- Agència de Gestió d’Ajuts Universitaris i de Recerca (AGAUR Grant No. 2021 SGR 01452)
- QuantERA (DYNAMITE PCI2022-132919)
- Simons Foundation (Flatiron Institute)
- Agencia Estatal de Investigación (CEX2019-000910-S/10.13039/501100011033)
This publication has 100 references indexed in Scilit:
- Unsupervised learning of interacting topological phases from experimental observablesFundamental Research, 2023
- Generalization in quantum machine learning from few training dataNature Communications, 2022
- Towards robust explanations for deep neural networksPattern Recognition, 2022
- Correlator convolutional neural networks as an interpretable architecture for image-like quantum matter dataNature Communications, 2021
- Topological phases of a dimerized Fermi–Hubbard model for semiconductor nano-latticesnpj Quantum Information, 2020
- Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based LocalizationInternational Journal of Computer Vision, 2019
- Identifying topological order through unsupervised machine learningNature Physics, 2019
- Applications of neural networks to the studies of phase transitions of two-dimensional Potts modelsAnnals of Physics, 2018
- Learning phase transitions by confusionNature Physics, 2017
- Machine learning phases of matterNature Physics, 2017