MicroEvoEval: A Systematic Evaluation Framework for Image-Based Microstructure Evolution Prediction

2025-11-12

AI Summary

Current microstructure evolution prediction lacks standardized evaluation benchmarks: existing studies fail to systematically compare domain-specific models against state-of-the-art spatiotemporal architectures (e.g., VMamba), overemphasize numerical accuracy while neglecting physical fidelity, and omit error propagation analysis over time. Method: We introduce the first systematic evaluation framework for image-based microstructure evolution, featuring a benchmark dataset with four representative tasks and co-designed short- and long-term sequences; we propose multi-scale structural fidelity metrics and error propagation analysis to jointly assess numerical accuracy, physical consistency, and computational efficiency. Results: Extensive experiments across 14 spatiotemporal models demonstrate that modern architectures significantly outperform traditional methods in long-term stability and physical fidelity, while accelerating inference by an order of magnitude—establishing a reproducible, scalable paradigm for surrogate model evaluation in data-driven materials science.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Computer Vision: Diffusion Models for Vision

Application Category

• Web Mining and Content Analysis: Models for Web evolution
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization

Abstract

Simulating microstructure evolution (MicroEvo) is vital for materials design but demands high numerical accuracy, efficiency, and physical fidelity. Although recent studies on deep learning (DL) offer a promising alternative to traditional solvers, the field lacks standardized benchmarks. Existing studies are flawed due to a lack of comparing specialized MicroEvo DL models with state-of-the-art spatio-temporal architectures, an overemphasis on numerical accuracy over physical fidelity, and a failure to analyze error propagation over time. To address these gaps, we introduce MicroEvoEval, the first comprehensive benchmark for image-based microstructure evolution prediction. We evaluate 14 models, encompassing both domain-specific and general-purpose architectures, across four representative MicroEvo tasks with datasets specifically structured for both short- and long-term assessment. Our multi-faceted evaluation framework goes beyond numerical accuracy and computational cost, incorporating a curated set of structure-preserving metrics to assess physical fidelity. Our extensive evaluations yield several key insights. Notably, we find that modern architectures (e.g., VMamba), not only achieve superior long-term stability and physical fidelity but also operate with an order-of-magnitude greater computational efficiency. The results highlight the necessity of holistic evaluation and identify these modern architectures as a highly promising direction for developing efficient and reliable surrogate models in data-driven materials science.

Problem

• Establishing standardized benchmarks for microstructure evolution prediction models
• Evaluating physical fidelity and error propagation in deep learning approaches
• Comparing domain-specific and general-purpose architectures across multiple tasks

Innovation

• Introduces MicroEvoEval benchmark for microstructure evolution prediction
• Evaluates 14 models across short- and long-term MicroEvo tasks
• Incorporates structure-preserving metrics to assess physical fidelity