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

Abstract

For performing regression tasks involved in various physics problems, enhancing the precision or equivalently reducing the uncertainty of regression results is undoubtedly one of the central goals. Here, somewhat surprisingly, we find that the unfavorable regression uncertainty in performing the regression tasks of inverse statistical problems actually contains hidden information concerning the phase transitions of the system under consideration. By utilizing this hidden information, we develop a new unsupervised machine learning approach for automated detection of phases of matter, dubbed learning from regression uncertainty. This is achieved by revealing an intrinsic connection between regression uncertainty and response properties of the system, thus making the outputs of this machine learning approach directly interpretable via conventional notions of physics. We demonstrate the approach by identifying the critical points of the ferromagnetic Ising model and the three-state clock model, and revealing the existence of the intermediate phase in the six-state and seven-state clock models. Comparing to the widely-used classification-based approaches developed so far, although successful, their recognized classes of patterns are essentially abstract, which hinders their straightforward relation to conventional notions of physics. These challenges persist even when one employs the state-of-the-art deep neural networks that excel at classification tasks. In contrast, with the core working horse being a neural network performing regression tasks, our new approach is not only practically more efficient, but also paves the way towards intriguing possibilities for unveiling new physics via machine learning in a physically interpretable manner.

arXiv comment11 pages, 7 figures

References in corpus (70)

• Adam: A Method for Stochastic Optimization
• Novel type of phase transition in a system of self-driven particles
• Active Particles in Complex and Crowded Environments
• Collective motion
• Machine learning and the physical sciences
• Machine learning phases of matter
• Learning phase transitions by confusion
• Colloquium: Criticality and dynamical scaling in living systems
• Unsupervised learning of phase transitions: from principal component analysis to variational autoencoders
• Machine learning quantum phases of matter beyond the fermion sign problem
• Full phase diagram of active Brownian disks: from melting to motility-induced phase separation
• Identifying topological order through unsupervised machine learning
• Inverse statistical problems: from the inverse Ising problem to data science
• Machine Learning Phases of Strongly Correlated Fermions
• Identifying Quantum Phase Transitions using Artificial Neural Networks on Experimental Data
• Solving Statistical Mechanics Using Variational Autoregressive Networks
• Machine learning vortices at the Kosterlitz-Thouless transition
• Inverse Ising inference using all the data
• Unsupervised machine learning and band topology
• Neural Network Renormalization Group
• Machine Learning for Quantum Matter
• Probing many-body localization with neural networks
• Machine learning out-of-equilibrium phases of matter
• Dualities and the phase diagram of the $p$-clock model
• AI Poincaré: Machine Learning Conservation Laws from Trajectories
• Classifying Snapshots of the Doped Hubbard Model with Machine Learning
• Machine learning of phase transitions in the percolation and XY models
• Toward an AI Physicist for Unsupervised Learning
• Machine learning of quantum phase transitions
• Artificial Neural Network Approach to the Analytic Continuation Problem
• Unveiling phase transitions with machine learning
• Machine Learning Many-Body Localization: Search for the Elusive Nonergodic Metal
• Parameter diagnostics of phases and phase transition learning by neural networks
• Critical properties of the two-dimensional $q$-state clock model
• Unsupervised identification of topological order using predictive models
• Machine-learning hidden symmetries
• Correlator Convolutional Neural Networks: An Interpretable Architecture for Image-like Quantum Matter Data
• Symbolic Pregression: Discovering Physical Laws from Distorted Video
• Mean-field theory for the inverse Ising problem at low temperatures
• Response to a twist on systems with Zp symmetry
• Attraction tames two-dimensional melting: from continuous to discontinuous transitions
• Machine Learning Holographic Mapping by Neural Network Renormalization Group
• Divergence of predictive model output as indication of phase transitions
• Interpretable Machine Learning Study of Many-Body Localization Transition in Disordered Quantum Ising Spin Chains
• Learning the Ising Model with Generative Neural Networks
• Machine-learning detection of the Berezinskii-Kosterlitz-Thouless transitions in the q-state clock models
• Absorbing phase transition with a continuously varying exponent in a quantum contact process: a neural network approach
• Generative models for sampling and phase transition indication in spin systems
• Few-shot machine learning in the three-dimensional Ising model
• Interpretable and unsupervised phase classification
• Residual discrete symmetry of the five-state clock model
• Machine-Learning Non-Conservative Dynamics for New-Physics Detection
• Logarithmic finite-size scaling correction to the leading Fisher zeros in the p-state clock model: A higher-order tensor renormalization group study
• Inverse Ising problem in continuous time: A latent variable approach
• Metal-insulator transition and intermediate phases in the kagome lattice Hubbard model
• Higher-order interactions in statistical physics and machine learning: A model-independent solution to the inverse problem at equilibrium
• Statistical physics through the lens of real-space mutual information
• Deep-learning-assisted detection and termination of spiral- and broken-spiral waves in mathematical models for cardiac tissue
• Solving the inverse Ising problem by mean-field methods in a clustered phase space with many states
• Symmetries and phase diagrams with real-space mutual information neural estimation
• Continuous-mixture Autoregressive Networks for efficient variational calculation of many-body systems
• Clock model interpolation and symmetry breaking in O(2) models
• Replacing neural networks by optimal analytical predictors for the detection of phase transitions
• Deep neural networks for inverse problems in mesoscopic physics: Characterization of the disorder configuration from quantum transport properties
• Machine learning phases and criticalities without using real data for training
• Topological defects and confinement with machine learning: the case of monopoles in compact electrodynamics
• Intermediate Phase in Interacting Dirac Fermions with Staggered Potential
• Inverse problem for the mean-field monomer-dimer model with attractive interaction
• Inverse Ising inference from high-temperature re-weighting of observations
• Scanning-probe and information-concealing machine learning intermediate hexatic phase and critical scaling of solid-hexatic phase transition in deformable particles

Cited by in corpus (3)

• Mapping out phase diagrams with generative classifiers
• Phases and Phase Transitions of the Disordered Quantum Clock Model
• Reveal flocking phase transition of self-propelled active particles by machine learning regression uncertainty