Automated discovery of characteristic features of phase transitions in many-body localization

Abstract

We identify a new "order parameter" for the disorder driven many-body localization (MBL) transition by leveraging artificial intelligence. This allows us to pin down the transition, as the point at which the physics changes qualitatively, from vastly fewer disorder realizations and in an objective and cleaner way than is possible with the existing zoo of quantities. Contrary to previous studies, our method is almost entirely unsupervised. A game theoretic process between neural networks defines an adversarial setup with conflicting objectives to identify what characteristic features to base efficient predictions on. This reduces the numerical effort for mapping out the phase diagram by a factor of ~100x. This approach of automated discovery is applicable specifically to poorly understood phase transitions and exemplifies the potential of machine learning assisted research in physics.

arXiv comment3 pages + 3 pages appendix, 4 figures

References in corpus (24)

• Many body localization and thermalization in quantum statistical mechanics
• Metal-insulator transition in a weakly interacting many-electron system with localized single-particle states
• The many-body localization phase transition
• Machine learning phases of matter
• Many-body localization edge in the random-field Heisenberg chain
• Phenomenology of fully many-body-localized systems
• A phenomenology of certain many-body-localized systems
• Local conservation laws and the structure of the many-body localized states
• Learning phase transitions by confusion
• Discovering Phase Transitions with Unsupervised Learning
• Many-body localization in a disordered quantum Ising chain
• Unsupervised learning of phase transitions: from principal component analysis to variational autoencoders
• Discovering Phases, Phase Transitions and Crossovers through Unsupervised Machine Learning: A critical examination
• Machine Learning Topological Invariants with Neural Networks
• Machine Learning Phases of Strongly Correlated Fermions
• Many-body delocalization as a quantum avalanche
• Probing many-body localization with neural networks
• Identifying Quantum Phase Transitions with Adversarial Neural Networks
• Shift-invert diagonalization of large many-body localizing spin chains
• Many-body localization in infinite chains
• Many-body localization, symmetry, and topology
• Total correlations of the diagonal ensemble herald the many-body localization transition
• Quantum phase recognition via unsupervised machine learning
• Finite-size scaling with respect to interaction and disorder strength at the many-body localization transition

Cited by in corpus (30)

• Machine learning and the physical sciences
• Discovering physical concepts with neural networks
• Many-Body Localization in the Age of Classical Computing
• Unsupervised phase discovery with deep anomaly detection
• Topological quantum phase transitions retrieved through unsupervised machine learning
• Unsupervised machine learning of topological phase transitions from experimental data
• Stability of many-body localization in Floquet systems
• Transition to a many-body localized regime in a two-dimensional disordered quantum dimer model
• Universality in Anderson localization on random graphs with varying connectivity
• Machine learning discovery of new phases in programmable quantum simulator snapshots
• Machine learning topological invariants of non-Hermitian systems
• Neural network setups for a precise detection of the many-body localization transition: finite-size scaling and limitations
• Variational Quantum Anomaly Detection: Unsupervised mapping of phase diagrams on a physical quantum computer
• Unsupervised learning using topological data augmentation
• Phase Detection with Neural Networks: Interpreting the Black Box
• Detecting ergodic bubbles at the crossover to many-body localization using neural networks
• Hessian-based toolbox for reliable and interpretable machine learning in physics
• Distinguishing an Anderson Insulator from a Many-Body Localized phase through space-time snapshots with Neural Networks
• Replacing neural networks by optimal analytical predictors for the detection of phase transitions
• Unsupervised mapping of phase diagrams of 2D systems from infinite projected entangled-pair states via deep anomaly detection
• Detection of Berezinskii-Kosterlitz-Thouless transition via Generative Adversarial Networks
• Exponentially improved efficient machine learning for quantum many-body states with provable guarantees
• Machine learning of phases and structures for model systems in physics
• Neural network enhanced hybrid quantum many-body dynamical distributions
• Unsupervised Learning of Rydberg Atom Array Phase Diagram with Siamese Neural Networks
• Unsupervised interpretable learning of topological indices invariant under permutations of atomic bands
• Machine learning of mirror skin effects in the presence of disorder
• Characterizing out-of-distribution generalization of neural networks: application to the disordered Su-Schrieffer-Heeger model
• Many-body mobility edges in 1D and 2D revealed by convolutional neural networks
• Machine learning analysis of dimensional reduction conjecture for nonequilibrium Berezinskii-Kosterlitz-Thouless transition in three dimensions