Determining liquid crystal properties with ordinal networks and machine learning

Abstract

Machine learning methods are becoming increasingly important for the development of materials science. In spite of this, the use of image analysis in the development of these systems is still recent and underexplored, especially in materials often studied via optical imaging techniques such as liquid crystals. Here we apply the recently proposed method of ordinal networks to map optical textures obtained from experimental samples of liquid crystals into complex networks and use this representation jointly with a simple statistical learning algorithm to investigate different physical properties of these materials. Our research demonstrates that ordinal networks formed by only 24 nodes encode crucial information about liquid crystal properties, thus allowing us to train simple machine learning models capable of identifying and classifying mesophase transitions, distinguishing among different doping concentrations used to induce chiral mesophases, and predicting sample temperatures with outstanding accuracy. The precision and scalability of our approach indicate it can be used to probe properties of different materials in situations involving large-scale datasets or real-time monitoring systems.

arXiv comment9 two-column pages, 4 figures; accepted for publication in Chaos, Solitons & Fractals

References in corpus (14)

• Scikit-learn: Machine Learning in Python
• Machine learning and the physical sciences
• Searching for Exotic Particles in High-Energy Physics with Deep Learning
• A Statistical Measure of Complexity
• Insightful classification of crystal structures using deep learning
• Complexity-Entropy Causality Plane as a Complexity Measure for Two-dimensional Patterns
• ordpy: A Python package for data analysis with permutation entropy and ordinal network methods
• Deep-Learning continuous gravitational waves
• Transient Classification in LIGO data using Difference Boosting Neural Network
• Discriminating image textures with the multiscale two-dimensional complexity-entropy causality plane
• Learning physical properties of liquid crystals with deep convolutional neural networks
• Estimating physical properties from liquid crystal textures via machine learning and complexity-entropy methods
• Characterizing stochastic time series with ordinal networks
• Mapping images into ordinal networks

Cited by in corpus (5)

• Characterizing unstructured data with the nearest neighbor permutation entropy
• Machine Eye for Defects: Machine Learning-Based Solution to Identify and Characterize Topological Defects in Textured Images of Nematic Materials
• Nearest neighbor permutation entropy detects phase transitions in complex high-pressure systems
• Gradient dynamics in reinforcement learning
• Interpretable liquid crystal phase classification via two-by-two ordinal patterns