Unsupervised phase discovery with deep anomaly detection

Abstract

We demonstrate how to explore phase diagrams with automated and unsupervised machine learning to find regions of interest for possible new phases. In contrast to supervised learning, where data is classified using predetermined labels, we here perform anomaly detection, where the task is to differentiate a normal data set, composed of one or several classes, from anomalous data. Asa paradigmatic example, we explore the phase diagram of the extended Bose Hubbard model in one dimension at exact integer filling and employ deep neural networks to determine the entire phase diagram in a completely unsupervised and automated fashion. As input data for learning, we first use the entanglement spectra and central tensors derived from tensor-networks algorithms for ground-state computation and later we extend our method and use experimentally accessible data such as low-order correlation functions as inputs. Our method allows us to reveal a phase-separated region between supersolid and superfluid parts with unexpected properties, which appears in the system in addition to the standard superfluid, Mott insulator, Haldane-insulating, and density wave phases.

5 pages, 6 figures, suppl. material at http://journals.aps.org/prl/supplemental/10.1103/PhysRevLett.125.170603/PRL_suppl.pdf, readable code at https://github.com/Qottmann/phase-discovery-anomaly-detection

References in corpus (49)

• The density-matrix renormalization group in the age of matrix product states
• Machine learning and the physical sciences
• Efficient classical simulation of slightly entangled quantum computations
• A Practical Introduction to Tensor Networks: Matrix Product States and Projected Entangled Pair States
• Machine learning phases of matter
• Deep Learning for Anomaly Detection: A Survey
• One dimensional Bosons: From Condensed Matter Systems to Ultracold Gases
• Classical simulation of infinite-size quantum lattice systems in one spatial dimension
• Learning phase transitions by confusion
• Classical simulation of infinite-size quantum lattice systems in two spatial dimensions
• Discovering Phase Transitions with Unsupervised Learning
• Unsupervised learning of phase transitions: from principal component analysis to variational autoencoders
• The iTEBD algorithm beyond unitary evolution
• Discovering physical concepts with neural networks
• From density-matrix renormalization group to matrix product states
• Efficient numerical simulations with Tensor Networks: Tensor Network Python (TeNPy)
• The one-dimensional Bose-Hubbard Model with nearest-neighbor interaction
• Machine learning quantum phases of matter beyond the fermion sign problem
• Phases of the one-dimensional Bose-Hubbard model
• Machine Learning Topological States
• Machine Learning Topological Invariants with Neural Networks
• Rise and fall of hidden string order of lattice bosons
• Anomaly Detection using Autoencoders in High Performance Computing Systems
• Generalization in Deep Learning
• Mutual Information, Neural Networks and the Renormalization Group
• Unsupervised machine learning account of magnetic transitions in the Hubbard model
• Probing many-body localization with neural networks
• The Tensor Networks Anthology: Simulation techniques for many-body quantum lattice systems
• Phase diagram of the extended Bose Hubbard model
• Identifying Quantum Phase Transitions with Adversarial Neural Networks
• Supersolid phases in the one dimensional extended soft core Bosonic Hubbard model
• Deep learning generalizes because the parameter-function map is biased towards simple functions
• Discriminative Cooperative Networks for Detecting Phase Transitions
• Machine learning of quantum phase transitions
• Ground state fidelity and quantum phase transitions in free Fermi systems
• Entanglement spectrum of one-dimensional extended Bose-Hubbard models
• Automated discovery of characteristic features of phase transitions in many-body localization
• Spectral and Entanglement Properties of the Bosonic Haldane Insulator
• Supersolid and solitonic phases in one-dimensional Extended Bose-Hubbard model
• Neural network setups for a precise detection of the many-body localization transition: finite-size scaling and limitations
• Identifying and Categorizing Anomalies in Retinal Imaging Data
• Deep learning of topological phase transitions from entanglement aspects
• Phase diagrams of the extended Bose-Hubbard model in one dimension by Monte-Carlo simulation with the help of a stochastic-series expansion
• One-dimensional extended Bose-Hubbard model with a confining potential: a DMRG analysis
• Optical lattices as a tool to study defect-induced superfluidity
• Machine Learning Phase Diagram in the Half-filled One-dimensional Extended Hubbard Model
• Two supersolid phases in hard-core extended Bose-Hubbard model
• Credit Card Fraud Detection in e-Commerce: An Outlier Detection Approach
• Learning Deep Representations Using Convolutional Auto-encoders with Symmetric Skip Connections

Cited by in corpus (54)

• Artificial Intelligence and Machine Learning for Quantum Technologies
• Machine Learning for Quantum Matter
• Exploring QCD matter in extreme conditions with Machine Learning
• Unsupervised machine learning of topological phase transitions from experimental data
• Quantum phase detection generalisation from marginal quantum neural network models
• Topological quantum critical points in the extended Bose-Hubbard model
• Unsupervised Learning of Non-Hermitian Topological Phases
• Machine learning discovery of new phases in programmable quantum simulator snapshots
• Variational Quantum Anomaly Detection: Unsupervised mapping of phase diagrams on a physical quantum computer
• Phase diagram study of a two-dimensional frustrated antiferromagnet via unsupervised machine learning
• Proofs of network quantum nonlocality in continuous families of distributions
• Intrinsic dimension of path integrals: data mining quantum criticality and emergent simplicity
• Phase Detection with Neural Networks: Interpreting the Black Box
• Unsupervised learning of anomalous diffusion data
• Supersolid-Superfluid phase separation in the extended Bose-Hubbard model
• Novel Approach to Unveil Quantum Phase Transitions Using Fidelity Map
• Detecting ergodic bubbles at the crossover to many-body localization using neural networks
• Machine learning of Kondo physics using variational autoencoders and symbolic regression
• From observations to complexity of quantum states via unsupervised learning
• Hessian-based toolbox for reliable and interpretable machine learning in physics
• Quantum droplets with particle imbalance in one-dimensional optical lattices
• Replacing neural networks by optimal analytical predictors for the detection of phase transitions
• Neural networks in quantum many-body physics: a hands-on tutorial
• Unsupervised mapping of phase diagrams of 2D systems from infinite projected entangled-pair states via deep anomaly detection
• Explainable Representation Learning of Small Quantum States
• Unsupervised learning of phase transitions via modified anomaly detection with autoencoders
• Certificates of quantum many-body properties assisted by machine learning
• Detection of Berezinskii-Kosterlitz-Thouless transition via Generative Adversarial Networks
• Unsupervised Interpretable Learning of Phases From Many-Qubit Systems
• Noise-Robust End-to-End Quantum Control using Deep Autoregressive Policy Networks
• Quantum Reservoir Computing for Speckle-Disorder Potentials
• The Coming Decades of Quantum Simulation
• Phase Diagram Detection via Gaussian Fitting of Number Probability Distribution
• Machine learning of phases and structures for model systems in physics
• Unsupervised Learning of Rydberg Atom Array Phase Diagram with Siamese Neural Networks
• Phase determination with and without deep learning
• Deep Learning of Phase Transitions for Quantum Spin Chains from Correlation Aspects
• Rapid detection of phase transitions from Monte Carlo samples before equilibrium
• Characterizing out-of-distribution generalization of neural networks: application to the disordered Su-Schrieffer-Heeger model
• Machine learning the Ising transition: A comparison between discriminative and generative approaches
• Fluctuation based interpretable analysis scheme for quantum many-body snapshots
• Learning phases with Quantum Monte Carlo simulation cell
• Quantum Anomaly Detection with a Spin Processor in Diamond
• Scanning-probe and information-concealing machine learning intermediate hexatic phase and critical scaling of solid-hexatic phase transition in deformable particles
• Deep Neural Network-assisted improvement of quantum compressed sensing tomography
• Deep learning of topological phase transitions from entanglement aspects: An unsupervised way
• Mapping Phase Diagrams of Quantum Spin Systems through Semidefinite-Programming Relaxations
• Global exploration of phase behavior in frustrated Ising models using unsupervised learning techniques
• Topological Learning in Multi-Class Data Sets
• Connecting Topological Anderson and Mott Insulators in Disordered Interacting Fermionic Systems
• Emergent supercounterfluid and quantum phase diagram of two-component interacting bosons in one-dimensional optical lattice
• Machine learning analysis of dimensional reduction conjecture for nonequilibrium Berezinskii-Kosterlitz-Thouless transition in three dimensions
• Topological phases of bosons with local parity coupling on a dimerized lattice
• Visual configuration segmentation of quantum states for phase identification in many-body systems