AI framework predicts wear fields with high accuracy using uncertainty guidance

Researchers have developed a novel uncertainty-guided active learning framework to predict wear fields in stream finishing processes. This method utilizes a deep ensemble to estimate epistemic uncertainty, allowing for selective DEM simulations on the most uncertain orientations. The framework accurately predicts key fields governing erosion, achieving high Spearman rank correlations for normal impact velocity, tangential impact velocity, and particle impact flux. The predicted uncertainty is well-calibrated, reliably anticipating prediction errors and the fidelity of the reconstructed wear field. AI

IMPACT This research demonstrates a method for improving the efficiency and accuracy of complex simulations through AI-driven uncertainty quantification.

RANK_REASON Academic paper detailing a novel machine learning approach for a specific engineering problem. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• alphaXiv
• arXiv
• CatalyzeX Code Finder for Papers
• CORE Recommender
• DagsHub
• digital elevation model
• Finnie wear model
• Gotit.pub
• Hugging Face
• IArxiv Recommender
• Influence Flower
• Koushik Viswanathan
• ScienceCast

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →