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

2026-09-28

AI Summary

This study addresses the challenge of simultaneously achieving geometric validity, novelty, and diversity in metamaterial voxel generation by proposing the REGDIFF framework. This method integrates voxel discretization with a variational autoencoder for latent space regulation and employs a conditional diffusion model to enable high-quality geometric synthesis. Its core innovations include a repulsion-sinking mechanism designed to smooth the latent distribution, a short-range repulsion guidance strategy that effectively mitigates overfitting, and a dedicated benchmark dataset. Experimental results demonstrate that REGDIFF achieves superior performance across two datasets, improving geometric validity, novelty, and diversity by 8.9%, 46.4%, and 128.6%, respectively. These findings significantly advance the intelligent inverse design of metamaterials.

Abstract

Metamaterials are artificially engineered structures whose mechanical and physical behaviors are strongly shaped by geometry rather than composition. Voxel representation provides a unified format for metamaterial geometry generation, as it can express diverse classes such as truss, shell, and porous structures within a single cubic discretization. However, voxel-based generation faces a plausibility-novelty trade-off: staying close to known geometries helps preserve geometric regularities, while moving away from them is necessary for novelty but may produce degenerate geometries. To address this challenge, we propose REGDIFF, a generative framework that couples voxel representation with latent space regulation and guided diffusion. REGDIFF introduces a repel-and-sink (RAS) mechanism to smooth the latent distribution of plausible geometries, and short-range repulsion (SRR) guidance to discourage generation overly close to known samples while maintaining geometric plausibility. We further contribute a voxel-based benchmark covering truss- and shell-type metamaterial geometries, together with an evaluation module for geometric plausibility, novelty, and diversity. Experiments show that REGDIFF outperforms voxel-based generative baselines, achieving +8.9% in geometric plausibility, +46.4% in novelty, and +128.6% in diversity on average across two datasets. These results suggest that REGDIFF is a strong geometry candidate generator for downstream evaluation. Our code is provided at https://github.com/wzhan24/ReGDiff.

Problem

• metamaterials
• voxel generation
• plausibility-novelty trade-off
• geometry design

Innovation

• Guided Diffusion
• Latent Space Regulation
• Metamaterials
• Voxel Geometry
• Repel-and-Sink Mechanism