The percolating cluster is invisible to image recognition with deep learning

Abstract

We study the two-dimensional site-percolation model on a square lattice. In this paradigmatic model, sites are randomly occupied with probability $p$; a second-order phase transition from a non-percolating to a fully percolating phase appears at occupation density $p_c$, called percolation threshold. Through supervised deep learning approaches like classification and regression, we show that standard convolutional neural networks (CNNs), known to work well in similar image recognition tasks, can identify $p_c$ and indeed classify the states of a percolation lattice according to their $p$ content or predict their $p$ value via regression. When using instead of $p$ the spatial cluster correlation length $\xi$ as labels, the recognition is beginning to falter. Finally, we show that the same network struggles to detect the presence of a spanning cluster. Rather, predictive power seems lost and the absence or presence of a global spanning cluster is not noticed by a CNN with local convolutional kernel. Since the existence of such a spanning cluster is at the heart of the percolation problem, our results suggest that CNNs require careful application when used in physics, particularly when encountering less-explored situations.

arXiv comment10 pages, 12 figures

References in corpus (24)

• ADADELTA: An Adaptive Learning Rate Method
• Solving the Quantum Many-Body Problem with Artificial Neural Networks
• Machine learning phases of matter
• A high-bias, low-variance introduction to Machine Learning for physicists
• Learning phase transitions by confusion
• Restricted-Boltzmann-Machine Learning for Solving Strongly Correlated Quantum Systems
• Quantum Loop Topography for Machine Learning
• Machine Learning Phases of Strongly Correlated Fermions
• Self-Learning Monte Carlo Method
• Accelerate Monte Carlo Simulations with Restricted Boltzmann Machines
• Approximating quantum many-body wave-functions using artificial neural networks
• Typical medium theory of Anderson localization: A local order parameter approach to strong disorder effects
• NetKet 3: Machine Learning Toolbox for Many-Body Quantum Systems
• Machine learning out-of-equilibrium phases of matter
• Solving the Bose-Hubbard model with machine learning
• Machine learning of phase transitions in the percolation and XY models
• Machine Learning for Condensed Matter Physics
• Unsupervised machine learning of topological phase transitions from experimental data
• Gated-Dilated Networks for Lung Nodule Classification in CT scans
• Drawing Phase Diagrams of Random Quantum Systems by Deep Learning the Wave Functions
• Method to solve quantum few-body problems with artificial neural networks
• Supervised and unsupervised learning of directed percolation
• Machine Learning Percolation Model
• Machine Learning as an Accurate Predictor for Percolation Threshold of Diverse Networks

Cited by in corpus (5)

• Machine learning of phases and structures for model systems in physics
• Phase determination with and without deep learning
• Minimalist Neural Networks training for phase classification in diluted-Ising models
• Learning phase transitions by siamese neural network
• Large-Angle Convergent-Beam Electron Diffraction Patterns via Conditional Generative Adversarial Networks