UniMate: A Unified Model for Mechanical Metamaterial Generation, Property Prediction, and Condition Confirmation

2025/06/05 by Zhan, Wangzhi, Chen, Jianpeng, Fu, Dongqi +1
Engineering · #Topology Optimization in Engineering #Cellular and Composite Structures #3D Shape Modeling and Analysis

paper · pdf · doi:10.48550/arxiv.2506.15722

Abstract

Metamaterials are artificial materials that are designed to meet unseen properties in nature, such as ultra-stiffness and negative materials indices. In mechanical metamaterial design, three key modalities are typically involved, i.e., 3D topology, density condition, and mechanical property. Real-world complex application scenarios place the demanding requirements on machine learning models to consider all three modalities together. However, a comprehensive literature review indicates that most existing works only consider two modalities, e.g., predicting mechanical properties given the 3D topology or generating 3D topology given the required properties. Therefore, there is still a significant gap for the state-of-the-art machine learning models capturing the whole. Hence, we propose a unified model named UNIMATE, which consists of a modality alignment module and a synergetic diffusion generation module. Experiments indicate that UNIMATE outperforms the other baseline models in topology generation task, property prediction task, and condition confirmation task by up to 80.2%, 5.1%, and 50.2%, respectively. We opensource our proposed UNIMATE model and corresponding results at https://github.com/wzhan24/UniMate.

Citations

• Equivariant Diffusion for Crystal Structure Prediction
• MetamatBench: Integrating Heterogeneous Data, Computational Tools, and Visual Interface for Metamaterial Discovery
• MetaScientist: A Human-AI Synergistic Framework for Automated Mechanical Metamaterial Design
• Complete and Efficient Graph Transformers for Crystal Material Property Prediction
• Fine-Tuned Language Models Generate Stable Inorganic Materials as Text
• Energy-conserving equivariant GNN for elasticity of lattice architected metamaterials
• Towards Symmetry-Aware Generation of Periodic Materials
• Symmetry-Informed Geometric Representation for Molecules, Proteins, and Crystalline Materials
• M²Hub: Unlocking the Potential of Machine Learning for Materials Discovery
• FairGen: Towards Fair Graph Generation
• Variational Graph Generator for Multi-View Graph Clustering
• Diffusion Models: A Comprehensive Survey of Methods and Applications
• Equiformer: Equivariant Graph Attention Transformer for 3D Atomistic Graphs
• MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fields
• Equivariant Diffusion for Molecule Generation in 3D
• Predicting deformation mechanisms in architected metamaterials using GNN
• Crystal Diffusion Variational Autoencoder for Periodic Material Generation
• Spherical Message Passing for 3D Graph Networks
• Equivariant message passing for the prediction of tensorial properties and molecular spectra
• Generative Modeling by Estimating Gradients of the Data Distribution
• Symmetry-adapted generation of 3d point sets for the targeted discovery of molecules
• Neural Discrete Representation Learning
• Variational Graph Auto-Encoders

Related

• MetamatBench: Integrating Heterogeneous Data, Computational Tools, and Visual Interface for Metamaterial Discovery
• MetaScientist: A Human-AI Synergistic Framework for Automated Mechanical Metamaterial Design
• Gyroid-like metamaterials: Topology optimization and Deep Learning
• Recent Advances and Applications of Machine Learning in Experimental Solid Mechanics: A Review
• MetaGen: A DSL, Database, and Benchmark for VLM-Assisted Metamaterial Generation