![[2602.19915] Fully Convolutional Spatiotemporal Learning for Microstructure Evolution Prediction](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.19915] Fully Convolutional Spatiotemporal Learning for Microstructure Evolution Prediction
Summary
This article presents a deep learning framework for predicting microstructure evolution in materials science, enhancing accuracy while reducing computational costs compared to traditional methods.
Why It Matters
Understanding microstructure evolution is crucial for materials science, impacting material properties and performance. This research offers a scalable, data-driven alternative to traditional simulation methods, potentially accelerating advancements in material development and applications.
Key Takeaways
- Introduces a deep learning framework for microstructure evolution prediction.
- Achieves state-of-the-art performance with reduced computational costs.
- Utilizes self-supervised learning on sequential images from simulations.
- Demonstrates generalization to unseen spatiotemporal domains.
- Establishes a baseline for future spatiotemporal learning in materials science.
Computer Science > Machine Learning arXiv:2602.19915 (cs) [Submitted on 23 Feb 2026] Title:Fully Convolutional Spatiotemporal Learning for Microstructure Evolution Prediction Authors:Michael Trimboli, Mohammed Alsubaie, Sirani M. Perera, Ke-Gang Wang, Xianqi Li View a PDF of the paper titled Fully Convolutional Spatiotemporal Learning for Microstructure Evolution Prediction, by Michael Trimboli and 4 other authors View PDF HTML (experimental) Abstract:Understanding and predicting microstructure evolution is fundamental to materials science, as it governs the resulting properties and performance of materials. Traditional simulation methods, such as phase-field models, offer high-fidelity results but are computationally expensive due to the need to solve complex partial differential equations at fine spatiotemporal resolutions. To address this challenge, we propose a deep learning-based framework that accelerates microstructure evolution predictions while maintaining high accuracy. Our approach utilizes a fully convolutional spatiotemporal model trained in a self-supervised manner using sequential images generated from simulations of microstructural processes, including grain growth and spinodal decomposition. The trained neural network effectively learns the underlying physical dynamics and can accurately capture both short-term local behaviors and long-term statistical properties of evolving microstructures, while also demonstrating generalization to unseen spatiotemporal ...
