Machine Learning of Implicit Combinatorial Rules in Mechanical Metamaterials

Abstract

Combinatorial problems arising in puzzles, origami, and (meta)material design have rare sets of solutions, which define complex and sharply delineated boundaries in configuration space. These boundaries are difficult to capture with conventional statistical and numerical methods. Here we show that convolutional neural networks can learn to recognize these boundaries for combinatorial mechanical metamaterials, down to finest detail, despite using heavily undersampled training sets, and can successfully generalize. This suggests that the network infers the underlying combinatorial rules from the sparse training set, opening up new possibilities for complex design of (meta)materials.

References in corpus (7)

• Learning phase transitions by confusion
• Combinatorial Design of Textured Mechanical Metamaterials
• Identifying structural flow defects in disordered solids using machine learning methods
• Deep Generative Modeling for Mechanistic-based Learning and Design of Metamaterial Systems
• Spiderweb nanomechanical resonators via Bayesian optimization: inspired by nature and guided by machine learning
• Learning physical properties of liquid crystals with deep convolutional neural networks
• Learning to Control Active Matter

Cited by in corpus (7)

• Perspective: How to overcome dynamical density functional theory
• Emergent Nonlocal Combinatorial Design Rules for Multimodal Metamaterials
• Symmetry-driven Phononic Metamaterials
• Combinatorial Design of Floppy Modes and Frustrated Loops in Metamaterials
• Prospecting for Pluripotency in Metamaterial Design
• Breaking Mechanical Holography in Combinatorial Metamaterials
• Defect Positioning in Combinatorial Metamaterials