Generative AI and Transfer Learning Enhance Surrogate Modeling for Engineering

Researchers have developed a novel framework for probabilistic multi-fidelity surrogate modeling that leverages generative AI and transfer learning to address data scarcity in complex engineering systems. The approach uses a normalizing flow generative model, first pre-trained on abundant low-fidelity data and then fine-tuned on limited high-fidelity data. This method allows for accurate probabilistic predictions with quantified uncertainty, outperforming traditional low-fidelity baselines and reducing the need for expensive high-fidelity simulations. The framework has been validated on benchmark systems, demonstrating its potential for data-efficient AI-driven surrogates in engineering. AI

IMPACT This research offers a path toward more data-efficient AI-driven surrogates for complex engineering simulations.

RANK_REASON The cluster contains a research paper detailing a new methodology for surrogate modeling. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv stat.ML →

• arXiv
• DagsHub
• generative artificial intelligence
• Hugging Face
• Jice Zeng
• Normalizing Flow
• Probabilistic Multi-Fidelity Surrogate Modeling
• transfer learning

• paper
• model release

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