PolyGraphCL: A Multi-View Graph Contrastive Learning Framework for Grain-Level Fatigue Damage Prediction in Polycrystalline Materials
Bibliographic record
Abstract
Accurately predicting fatigue damage at the grain scale in polycrystalline materials is challenging, primarily due to the complex microstructural topology, anisotropic deformation, and severe class imbalance caused by the rarity of slip-band-marked damage events relative to the vast population of intact grains. Conventional machine learning methods and single-view graph neural networks often lack the capacity to model such heterogeneity across scales. To bridge this gap, we introduce PolyGraphCL, a novel multi-view graph contrastive learning framework integrating heterogeneous inductive biases from three backbones—Graph Convolutional Network for localized neighborhood aggregation, Graph Attention Network for globally attentive interactions, and Graph Sample and Aggregate for multi-scale sampling. These diverse structural views, derived from applying different graph neural network architectures to the same input graph, are fused through a learnable attention mechanism, enabling dynamic weighting of view-specific representations per node to capture both fine-grained and holistic structural characteristics. To further address extreme label imbalance, we incorporate cross-view contrastive learning that aligns intra-node representations across views while repelling inter-node embeddings, facilitating the formation of class-discriminative manifolds. Evaluated on a ferritic steel microstructure dataset comprising 7,633 grains (311 damaged) with 100 descriptors per node, PolyGraphCL achieves an average F1 score of 0.8816±0.0505 and balanced accuracy of 0.7788±0.1606 under stratified five-fold cross-validation-surpassing both conventional machine learning baselines and single-view graph neural networks. Furthermore, GNNExplainer-based attribution reveals that PolyGraphCL’s predictions are predominantly governed by local stress concentration, with moderate influence from topological substructures, offering interpretable insights grounded in underlying physical mechanisms. Altogether, PolyGraphCL offers a robust, interpretable, and domain-adaptive framework for advancing data-driven fatigue prediction in computational materials science.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.003 | 0.001 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".