Unsupervised machine learning of topological phase transitions from experimental data

2021/01/31 by Niklas Käming, Anna Dawid, Korbinian Kottmann +5 · 2 citations
Biochemistry, Genetics and Molecular Biology · Materials Science · Mathematics · Physics and Astronomy · #Algorithm #Artificial intelligence #Combinatorics #Computer science #Condensed matter physics #Machine Learning in Materials Science #Machine learning #Mathematics #Phase transition #Physics #Protein Structure and Dynamics #Quantum many-body systems #Statistical physics #Topological data analysis #Topology (electrical circuits) #Unsupervised learning #cond-mat.dis-nn #cond-mat.mes-hall #cond-mat.quant-gas #quant-ph

paper · pdf · doi:10.1088/2632-2153/abffe7

published as Mach. Learn.: Sci. Technol. 2 035037 (2021) · 19 pages, 12 figures

openalex created_date 2021/01/18 · arxiv created 2021/02/01 · openalex publication_date 2021/05/11 · arxiv updated 2021/07/14 · openalex updated_date 2026/08/06

Abstract

Abstract Identifying phase transitions is one of the key challenges in quantum many-body physics. Recently, machine learning methods have been shown to be an alternative way of localising phase boundaries from noisy and imperfect data without the knowledge of the order parameter. Here, we apply different unsupervised machine learning techniques, including anomaly detection and influence functions, to experimental data from ultracold atoms. In this way, we obtain the topological phase diagram of the Haldane model in a completely unbiased fashion. We show that these methods can successfully be applied to experimental data at finite temperatures and to the data of Floquet systems when post-processing the data to a single micromotion phase. Our work provides a benchmark for the unsupervised detection of new exotic phases in complex many-body systems.

Citations

• Adam: A Method for Stochastic Optimization
• Stochastic Backpropagation and Approximate Inference in Deep Generative Models
• Universal high-frequency behavior of periodically driven systems: from dynamical stabilization to Floquet engineering
• Auto-Encoding Variational Bayes
• An Aharonov-Bohm interferometer for determining Bloch band topology
• Auto-association by multilayer perceptrons and singular value decomposition
• Photovoltaic Hall effect in graphene
• Measuring the Chern number of Hofstadter bands with ultracold bosonic atoms
• The NumPy array: a structure for efficient numerical computation
• Solving the Quantum Many-Body Problem with Artificial Neural Networks
• Tutorial on Variational Autoencoders
• Understanding Black-box Predictions via Influence Functions
• Quantum phase recognition via unsupervised machine learning
• Deep Koalarization: Image Colorization using CNNs and Inception-ResNet-v2
• Machine learning and the physical sciences
• Integrating Neural Networks with a Quantum Simulator for State Reconstruction
• Detection of Influential Observation in Linear Regression

Cited by

• Machine learning classification of two-dimensional vortex configurations
• Connecting Topological Anderson and Mott Insulators in Disordered Interacting Fermionic Systems