EGMOF: Efficient Generation of Metal-Organic Frameworks Using a Hybrid Diffusion-Transformer Architecture

2025-11-05

AI Summary

Addressing the challenges of vast chemical space, scarce labeled data, and poor cross-task generalization in MOF inverse design, this work proposes a modular two-stage generative framework. In the first stage, a 1D diffusion model (Prop2Desc) maps target properties to interpretable chemical descriptors; in the second stage, a Transformer-based model (Desc2MOF) generates topologically explicit MOF structures from these descriptors. The framework supports multi-attribute conditional generation, drastically reduces data dependency (requiring only ~1,000 samples), and exhibits strong generalizability and scalability. On hydrogen adsorption prediction, it achieves >95% structural validity and an 84% hit rate—outperforming state-of-the-art methods. Furthermore, it successfully transfers to 29 diverse property prediction datasets, demonstrating broad applicability across MOF design tasks.

Technology Category

• Machine Learning: Mixture of Experts (MoE)
• Natural Language Processing: Generation
• Computer Vision: Diffusion Models for Vision

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

Designing materials with targeted properties remains challenging due to the vastness of chemical space and the scarcity of property-labeled data. While recent advances in generative models offer a promising way for inverse design, most approaches require large datasets and must be retrained for every new target property. Here, we introduce the EGMOF (Efficient Generation of MOFs), a hybrid diffusion-transformer framework that overcomes these limitations through a modular, descriptor-mediated workflow. EGMOF decomposes inverse design into two steps: (1) a one-dimensional diffusion model (Prop2Desc) that maps desired properties to chemically meaningful descriptors followed by (2) a transformer model (Desc2MOF) that generates structures from these descriptors. This modular hybrid design enables minimal retraining and maintains high accuracy even under small-data conditions. On a hydrogen uptake dataset, EGMOF achieved over 95% validity and 84% hit rate, representing significant improvements of up to 57% in validity and 14% in hit rate compared to existing methods, while remaining effective with only 1,000 training samples. Moreover, our model successfully performed conditional generation across 29 diverse property datasets, including CoREMOF, QMOF, and text-mined experimental datasets, whereas previous models have not. This work presents a data-efficient, generalizable approach to the inverse design of diverse MOFs and highlights the potential of modular inverse design workflows for broader materials discovery.

Problem

• Addresses inverse design challenges in materials with limited property-labeled data
• Overcomes retraining limitations for new target properties through modular workflow
• Enables efficient MOF generation across diverse chemical spaces and properties

Innovation

• Hybrid diffusion-transformer architecture for MOF generation
• Descriptor-mediated workflow enabling minimal retraining
• Modular design maintaining accuracy with small datasets