LEIA: Learned Environment for Interactive Architected Materials

2026-05-27

AI Summary

Traditional physics-based simulations struggle to enable real-time interactive modeling of engineering materials that exhibit nonlinear constitutive behavior, history-dependent internal states, inertial effects, and multiscale structures. This work proposes an interactive autoregressive world model that allows users to incrementally apply boundary conditions and instantly observe the evolution of deformation and stress fields on three-dimensional unstructured meshes. The method uniquely unifies the modeling of two distinct material classes—explicit microstructural lattices and implicit internal-variable-based homogenized plates—and integrates load sequence modeling with internal state evolution mechanisms. Evaluated on the MicroPlate benchmark, the approach significantly outperforms four baseline methods, generating designs that finite element validation confirms achieve high stress prediction accuracy, thereby substantially accelerating inverse material design workflows.

Abstract

World models have enabled interactive exploration of game environments and robotic manipulation, but physical engineering remains beyond their reach: real materials exhibit nonlinear constitutive laws, carry history-dependent internal state, undergo inertial dynamics, and may possess hierarchical structures spanning multiple length scales. We present LEIA (Learned Environment for Interactive Architected materials), a world model that lets engineers apply boundary conditions step by step and observe the resulting deformation and stress fields in real time. LEIA handles large three-dimensional unstructured meshes and generates autoregressive responses to user-specified loading. We introduce MicroPlate, a benchmark of architected plates spanning two regimes of microstructure modeling: architected lattices that resolve microstructure explicitly through three-dimensional geometry, and a homogeneous plate where microstructural change is modeled implicitly through internal degrees of freedom. MicroPlate is used to assess LEIA alongside four baseline methods across both regimes. Finally, we demonstrate that LEIA enables efficient candidate generation and ranking for fast surrogate-guided search for de novo designs of architected materials, with stress-accurate candidate ranking validated by finite element ground truth.

Problem

• world models
• architected materials
• nonlinear constitutive laws
• history-dependent internal state
• multi-scale structures

Innovation

• world model
• architected materials
• real-time simulation
• microstructure modeling
• surrogate-guided design