Skip to content
Open access

Language-guided inverse design of multiclass metamaterials using generative artificial intelligence

Nov 2026 · npj Metamaterials · Vol 2 · 0 citations · 44 references

TL;DR

MetaGen-LLM is introduced as a language-guided framework that maps free-form design requests to high-fidelity geometries across eight metamaterial classes, providing a unified route from natural language to continuous mesh representations of multiclass metamaterials.

Abstract

Inverse design of metamaterials is commonly conditioned on numerical targets and restricted to a single structural family, limiting accessibility and cross-family exploration. Here, MetaGen-LLM is introduced as a language-guided framework that maps free-form design requests to high-fidelity geometries across eight metamaterial classes. A fine-tuned large language model converts user prompts into structured morphological and property conditions. A condition encoder, generator classifier, and diffusion model then predict a geometry-generator class and modeling parameters, while a surrogate model refines candidates toward target relative density, directional Young’s moduli, and thermal conductivities. Procedural geometry generators render smooth mesh representations that can be directly exported for additive manufacturing without an additional voxel-to-mesh reconstruction step. Across in-distribution and controlled out-of-distribution tests, the generated designs closely matched the requested properties. The trained models are further integrated into an agentic interface that supports iterative language-based design and three-dimensional visualization. MetaGen-LLM provides a unified route from natural language to continuous mesh representations of multiclass metamaterials.

Read PDF

Similar papers

Preprint Nov 2026

An Agent-Driven, Cross-Validated Inverse-Design Pipeline for Dielectric Metasurfaces

Inverse design of dielectric metasurfaces is powerful but labour-intensive: each new target requires choosing a parameterization and a figure of merit, managing compute resources, and critically assessing results that may be numerically unconverged, sensitive to fabrication errors or trapped in local optima. Here we pr...

G. Lio, Roberto Caputo, Alain J. Corso et al. · 0 citations

ReGDiff: Guided Diffusion in Regulated Latent Space for Exploring Metamaterial Voxel Geometry

REGDIFF is proposed, a generative framework that couples voxel representation with latent space regulation and guided diffusion that outperforms voxel-based generative baselines in geometric plausibility, novelty, and diversity and is suggested to be a strong geometry candidate generator for downstream evaluation.

Wang-Zhi Zhan, Jian-Peng Chen, Dong-Qi Fu et al. · 0 citations
Open access Oct 2026

Autonomous Discovery of Multi-Material Metamaterials With Programmable Multimodal Thermal Responses via Reinforcement Learning.

Inverse design deep learning models have significantly accelerated materials discovery through mapping complex structure-property relationships. However, they remain constrained by prohibitive data demands and are limited to in-distribution interpolation designs. Efficient out-of-distribution inverse design under spars...

Chen-Xi Lu, Ke-Wen Pan, Xing-Ran Li et al. · 0 citations
Review Open access Oct 2026

Numerical Simulation Engines and Deep Learning Architectures for Inverse Design of Multilayer Metamaterials

The inverse design of multilayer metamaterials —artificial structures with engineered electromagnetic, acoustic, thermal, or mechanical properties— represents a high-dimensional optimisation challenge that has driven the integration of deep learning into computational photonics and materials science. This paper present...

Federico Rodas, Laura Malache, Osmani Mujica Pérez · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.