Interpretable and unsupervised phase classification
Fully automated classification methods that provide direct physical insights into phase diagrams are of current interest. Interpretable, i.e., fully explainable, methods are desired for which we understand why they yield a given phase classification. Ideally, phase classification methods should also...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
American Physical Society
2021-07-01
|
Series: | Physical Review Research |
Online Access: | http://doi.org/10.1103/PhysRevResearch.3.033052 |
_version_ | 1797210983749386240 |
---|---|
author | Julian Arnold Frank Schäfer Martin Žonda Axel U. J. Lode |
author_facet | Julian Arnold Frank Schäfer Martin Žonda Axel U. J. Lode |
author_sort | Julian Arnold |
collection | DOAJ |
description | Fully automated classification methods that provide direct physical insights into phase diagrams are of current interest. Interpretable, i.e., fully explainable, methods are desired for which we understand why they yield a given phase classification. Ideally, phase classification methods should also be unsupervised. That is, they should not require prior labeling or knowledge of the phases of matter to be characterized. Here, we demonstrate an unsupervised machine-learning method for phase classification, which is rendered interpretable via an analytical derivation of the functional relationship between its optimal predictions and the input data. Based on these findings, we propose and apply an alternative, physically-motivated, data-driven scheme, which relies on the difference between mean input features. This mean-based method does not rely on any predictive model and is thus computationally cheap and directly explainable. As an example, we consider the physically rich ground-state phase diagram of the spinless Falicov-Kimball model. |
first_indexed | 2024-04-24T10:19:16Z |
format | Article |
id | doaj.art-3c61cce813644c48bf7970f06234a918 |
institution | Directory Open Access Journal |
issn | 2643-1564 |
language | English |
last_indexed | 2024-04-24T10:19:16Z |
publishDate | 2021-07-01 |
publisher | American Physical Society |
record_format | Article |
series | Physical Review Research |
spelling | doaj.art-3c61cce813644c48bf7970f06234a9182024-04-12T17:11:49ZengAmerican Physical SocietyPhysical Review Research2643-15642021-07-013303305210.1103/PhysRevResearch.3.033052Interpretable and unsupervised phase classificationJulian ArnoldFrank SchäferMartin ŽondaAxel U. J. LodeFully automated classification methods that provide direct physical insights into phase diagrams are of current interest. Interpretable, i.e., fully explainable, methods are desired for which we understand why they yield a given phase classification. Ideally, phase classification methods should also be unsupervised. That is, they should not require prior labeling or knowledge of the phases of matter to be characterized. Here, we demonstrate an unsupervised machine-learning method for phase classification, which is rendered interpretable via an analytical derivation of the functional relationship between its optimal predictions and the input data. Based on these findings, we propose and apply an alternative, physically-motivated, data-driven scheme, which relies on the difference between mean input features. This mean-based method does not rely on any predictive model and is thus computationally cheap and directly explainable. As an example, we consider the physically rich ground-state phase diagram of the spinless Falicov-Kimball model.http://doi.org/10.1103/PhysRevResearch.3.033052 |
spellingShingle | Julian Arnold Frank Schäfer Martin Žonda Axel U. J. Lode Interpretable and unsupervised phase classification Physical Review Research |
title | Interpretable and unsupervised phase classification |
title_full | Interpretable and unsupervised phase classification |
title_fullStr | Interpretable and unsupervised phase classification |
title_full_unstemmed | Interpretable and unsupervised phase classification |
title_short | Interpretable and unsupervised phase classification |
title_sort | interpretable and unsupervised phase classification |
url | http://doi.org/10.1103/PhysRevResearch.3.033052 |
work_keys_str_mv | AT julianarnold interpretableandunsupervisedphaseclassification AT frankschafer interpretableandunsupervisedphaseclassification AT martinzonda interpretableandunsupervisedphaseclassification AT axelujlode interpretableandunsupervisedphaseclassification |