Bridging Text and Crystal Structures: Literature-driven Contrastive Learning for Materials Science

Abstract

Understanding structure-property relationships is an essential yet challenging aspect of materials discovery and development. To facilitate this process, recent studies in materials informatics have sought latent embedding spaces of crystal structures to capture their similarities based on properties and functionalities. However, abstract feature-based embedding spaces are human-unfriendly and prevent intuitive and efficient exploration of the vast materials space. Here we introduce Contrastive Language--Structure Pre-training (CLaSP), a learning paradigm for constructing crossmodal embedding spaces between crystal structures and texts. CLaSP aims to achieve material embeddings that 1) capture property- and functionality-related similarities between crystal structures and 2) allow intuitive retrieval of materials via user-provided description texts as queries. To compensate for the lack of sufficient datasets linking crystal structures with textual descriptions, CLaSP leverages a dataset of over 400,000 published crystal structures and corresponding publication records, including paper titles and abstracts, for training. We demonstrate the effectiveness of CLaSP through text-based crystal structure screening and embedding space visualization.

19 pages, 8 figures. Accepted to Machine Learning: Science and Technology (2025). Preliminary versions appeared at NeurIPS 2024 AI4Mat and CVPR 2025 MM4Mat workshops

References in corpus (30)

• Adam: A Method for Stochastic Optimization
• Efficient Estimation of Word Representations in Vector Space
• Sequence to Sequence Learning with Neural Networks
• Decoupled Weight Decay Regularization
• LLaMA: Open and Efficient Foundation Language Models
• Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
• Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable Prediction of Material Properties
• Llama 2: Open Foundation and Fine-Tuned Chat Models
• SchNet - a deep learning architecture for molecules and materials
• Hierarchical Text-Conditional Image Generation with CLIP Latents
• On the Opportunities and Risks of Foundation Models
• Graph Networks as a Universal Machine Learning Framework for Molecules and Crystals
• Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision
• A Universal Graph Deep Learning Interatomic Potential for the Periodic Table
• BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models
• The Open Catalyst 2020 (OC20) Dataset and Community Challenges
• Atomistic Line Graph Neural Network for Improved Materials Property Predictions
• The Joint Automated Repository for Various Integrated Simulations (JARVIS) for data-driven materials design
• The Open Catalyst 2022 (OC22) Dataset and Challenges for Oxide Electrocatalysts
• Global Attention based Graph Convolutional Neural Networks for Improved Materials Property Prediction
• A foundation model for atomistic materials chemistry
• Atom2Vec: learning atoms for materials discovery
• MatterSim: A Deep Learning Atomistic Model Across Elements, Temperatures and Pressures
• Hierarchical Visualization of Materials Space with Graph Convolutional Neural Networks
• Periodic Graph Transformers for Crystal Material Property Prediction
• Multimodal Foundation Models for Material Property Prediction and Discovery
• Enhancing Activity Prediction Models in Drug Discovery with the Ability to Understand Human Language
• Sheared LLaMA: Accelerating Language Model Pre-training via Structured Pruning
• Efficient Approximations of Complete Interatomic Potentials for Crystal Property Prediction
• Crystalformer: Infinitely Connected Attention for Periodic Structure Encoding