# Accelerating Materials Discovery: Learning a Universal Representation of Chemical Processes for Cross-Domain Property Prediction

arXiv: https://arxiv.org/abs/2512.05979 · HTML: https://mlsanity.com/paper/2512.05979/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 12 of its 19 references that are in the map.
- First posted on arXiv 26 Nov 2025.

## Builds on (most cited first, 12 of 12)

- [Graph Attention Networks](https://mlsanity.com/paper/1710.10903/) (2017)
- [Neural Message Passing for Quantum Chemistry](https://mlsanity.com/paper/1704.01212/) (2017)
- [UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction](https://mlsanity.com/paper/1802.03426/) (2018)
- [Fast Graph Representation Learning with PyTorch Geometric](https://mlsanity.com/paper/1903.02428/) (2019)
- [Masked Label Prediction: Unified Message Passing Model for Semi-Supervised Classification](https://mlsanity.com/paper/2009.03509/) (2020)
- [The Road Less Scheduled](https://mlsanity.com/paper/2405.15682/) (2024)
- [Materials science in the era of large language models: a perspective](https://mlsanity.com/paper/2403.06949/) (2024)
- [ChemDFM-X: Towards Large Multimodal Model for Chemistry](https://mlsanity.com/paper/2409.13194/) (2024)
- [Agent-based Learning of Materials Datasets from Scientific Literature](https://mlsanity.com/paper/2312.11690/) (2023)
- [Mining experimental data from Materials Science literature with Large Language Models: an evaluation study](https://mlsanity.com/paper/2401.11052/) (2024)
- [Artificial Intelligence and Generative Models for Materials Discovery -- A Review](https://mlsanity.com/paper/2508.03278/) (2025)
- [Learning inducing points and uncertainty on molecular data by scalable variational Gaussian processes](https://mlsanity.com/paper/2207.07654/) (2022)

## Abstract (arXiv)

Experimental validation of chemical processes is slow and costly, limiting exploration in materials discovery. Machine learning can prioritize promising candidates, but existing data in patents and literature is heterogeneous and difficult to use. We introduce a universal directed-tree process-graph representation that unifies unstructured text, molecular structures, and numeric measurements into a single machine-readable format. To learn from this structured data, we developed a multi-modal graph neural network with a property-conditioned attention mechanism. Trained on approximately 700,000 process graphs from nearly 9,000 diverse documents, our model learns semantically rich embeddings that generalize across domains. When fine-tuned on compact, domain-specific datasets, the pretrained model achieves strong performance, demonstrating that universal process representations learned at scale transfer effectively to specialized prediction tasks with minimal additional data.