Abstract:The power of machine learning algorithms to automatically classify different phases of matter and detect quantum phase transitions without necessity to characterize phases by various quantities like local order parameters or topological invariants as in conventional approaches defined machine learning phases of matter as a new research frontier and basic research tool in condensed matter and statistical physics. We study quantum criticality in the spin-1/2 square-lattice J1-J2 model with additional plaquette structure by combination of reinforcement and supervised machine learning techniques. In our calculations the ground-state spin-spin correlation matrices for several system sizes are first found by restricted Boltzmann machine based variational Monte Carlo method, equivalent to reinforcement learning, and then used as a training data for convolutional neural network based supervised machine learning algorithm for phases classification. The model exhibits a quantum phase transition from paramagnetic plaquette resonating valence bond state to an antiferromagnetic state and has been a topic of great interest because of its close connection to cuprate superconductors and possibility of realization in cold atoms experiments. We consider both frustrated and unfrustrated regimes and compare our results with the results obtained previously with other methods. We find that our results are in good agreement with the results obtained by coupled cluster and real space renormalization group methods for both frustrated and unfrustrated regimes. The quantum Monte Carlo and finite-size scaling result for the unfrustrated case however slightly differs from our result for the critical value of the inter-plaquette coupling strength, although the results are still in reasonable agreement.

Unsupervised Machine Learning Phase Classification for Falicov-Kimball Model

Unsupervised machine learning of topological phase transitions from experimental data

Machine Learning Phases of Strongly Correlated Fermions

Phase determination with and without deep learning

Detecting the phase transition in a strongly-interacting Fermi gas by unsupervised machine learning

Machine learning of phases and structures for model systems in physics

Unsupervised and supervised learning of interacting topological phases from single-particle correlation functions

Learning What a Machine Learns in a Many-Body Localization Transition

Machine learning quantum criticality in the spin-1/2 quantum antiferromagnets on the square lattice with plaquette structure

Explainable Machine Learning Identification of Superconductivity from Single-Particle Spectral Functions

Few-shot Machine Learning in the Three-Dimensional Ising Model

Self-Supervised Ensemble Learning: A Universal Method for Phase Transition Classification of Many-Body Systems

A learning by confusion approach to characterize phase transitions

Interpretable machine learning for inferring the phase boundaries in a nonequilibrium system

Learning phase transitions from regression uncertainty: A new regression-based machine learning approach for automated detection of phases of matter

Unsupervised identification of Floquet topological phase boundaries

Unsupervised machine learning of quenched gauge symmetries: A proof-of-concept demonstration

Deep learning of topological phase transitions from entanglement aspects: An unsupervised way

Topological Data Analysis Assisted Machine Learning for Polar Topological Structures in Oxide Superlattices

Machine Learning of Frustrated Classical Spin Models. II. Kernel Principal Component Analysis

Machine learning vortices at the Kosterlitz-Thouless transition