A multi-fidelity tabular prior-data fitted network model for accurate prediction and uncertainty quantification
Yan Shi, Cheng Liu, Aodi Yu, Zhenzhou Lü, Said Elias, Kai Cheng, Jiaqing Kou, Xin Chen, Yu Liu, H Huang, Michael Beer · Nature Communications · 2026
Accurate prediction of unknown labels from feature-label datasets using machine learning is critical for applications spanning drug discovery, disease diagnostics, and climate science. However, challenges persist with limited data, high-dimensional inputs, and multi-fidelity scenarios. We developed multi-fidelity tabular prior-data fitted network (MFTabPFN), a general-purpose multi-fidelity model integrating low- and high-fidelity data through a hierarchical transformer architecture to enhance prediction accuracy and uncertainty quantification (UQ). MFTabPFN captures cross-fidelity correlations while seamlessly adapting to single-fidelity data. An active learning framework further enhances scalability by prioritizing high-value data for model refinement, minimizing resource demands in resource-intensive tasks. Evaluated on various tasks such as forest fire burned area prediction, wine quality assessment, and computational fluid dynamics, MFTabPFN outperforms state-of-the-art methods, achieving varying degrees of prediction accuracy improvement. Its versatility and robust prediction and UQ capabilities across single- and multi-fidelity datasets position MFTabPFN as a promising tool for data-driven discovery in diverse applications. This paper presents MFTabPFN, a transformer-based multi-fidelity learning framework that integrates low- and high-fidelity data for improved prediction and uncertainty quantification. Across diverse benchmarks, it outperforms existing methods and supports efficient active learning for resource intensive applications.