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.

11 pages, 7 figures

References in corpus (21)

• Novel type of phase transition in a system of self-driven particles
• Learning phase transitions by confusion
• Machine learning vortices at the Kosterlitz-Thouless transition
• AI Poincaré: Machine Learning Conservation Laws from Trajectories
• Machine-learning hidden symmetries
• Correlator Convolutional Neural Networks: An Interpretable Architecture for Image-like Quantum Matter Data
• Mean-field theory for the inverse Ising problem at low temperatures
• Attraction tames two-dimensional melting: from continuous to discontinuous transitions
• Machine-learning detection of the Berezinskii-Kosterlitz-Thouless transitions in the q-state clock models
• Few-shot machine learning in the three-dimensional Ising model
• Interpretable and unsupervised phase classification
• Machine-Learning Non-Conservative Dynamics for New-Physics Detection
• Metal-insulator transition and intermediate phases in the kagome lattice Hubbard model
• Inverse Ising problem in continuous time: A latent variable approach
• Statistical physics through the lens of real-space mutual information
• Symmetries and phase diagrams with real-space mutual information neural estimation
• 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
• 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 (2)

• Mapping out phase diagrams with generative classifiers
• Phases and Phase Transitions of the Disordered Quantum Clock Model