New AI model RheOPT uses Newtonian pretraining for efficient non-Newtonian flow prediction

Researchers have developed RheOPT, a novel transformer-based neural operator designed for efficient prediction of non-Newtonian fluid flow. This model leverages Newtonian simulations for pretraining, significantly reducing the data requirements for learning complex non-Newtonian dynamics. RheOPT demonstrates substantial improvements in few-shot learning scenarios, outperforming models trained from scratch with more data, and shows promise in accurately capturing phenomena like vortex shedding. AI

IMPACT Establishes a new method for data-efficient learning in complex fluid dynamics simulations, potentially accelerating research in fields relying on non-Newtonian flow.

RANK_REASON The cluster contains a research paper detailing a new AI model and its application. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• arXiv
• Bingham
• Hugging Face
• Revin
• RheOFormer
• Ulrich Giesekus

• paper
• model release

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