Toward a Robust and Generalizable Metamaterial Foundation Model

TL;DR

MetaFO is a Bayesian transformer-based foundation model that enables zero-shot, out-of-distribution predictions and inverse design for metamaterials, significantly advancing AI-driven discovery and understanding of complex structure-property relationships.

Contribution

Introduces MetaFO, a novel probabilistic transformer model that learns metamaterial mechanics, enabling generalizable, zero-shot predictions and inverse design under OOD conditions.

Findings

MetaFO achieves accurate zero-shot predictions for unseen metamaterial configurations.

MetaFO outperforms task-specific models in nonlinear inverse design tasks.

The model uncovers complex structure-property relationships, expanding design possibilities.

Abstract

Advances in material functionalities drive innovations across various fields, where metamaterials-defined by structure rather than composition-are leading the way. Despite the rise of artificial intelligence (AI)-driven design strategies, their impact is limited by task-specific retraining, poor out-of-distribution(OOD) generalization, and the need for separate models for forward and inverse design. To address these limitations, we introduce the Metamaterial Foundation Model (MetaFO), a Bayesian transformer-based foundation model inspired by large language models. MetaFO learns the underlying mechanics of metamaterials, enabling probabilistic, zero-shot predictions across diverse, unseen combinations of material properties and structural responses. It also excels in nonlinear inverse design, even under OOD conditions. By treating metamaterials as an operator that maps material properties to structural responses, MetaFO uncovers intricate structure-property relationships and significantly expands the design space. This scalable and generalizable framework marks a paradigm shift in AI-driven metamaterial discovery, paving the way for next-generation innovations.