Salvato in:
Dettagli Bibliografici
Autori principali: Yu, Rosen Ting-Ying, Sung, Nicholas, Ahmed, Faez
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:https://arxiv.org/abs/2601.22371
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908799026069504
author Yu, Rosen Ting-Ying
Sung, Nicholas
Ahmed, Faez
author_facet Yu, Rosen Ting-Ying
Sung, Nicholas
Ahmed, Faez
contents Multi-fidelity (MF) regression often operates in regimes of extreme data imbalance, where the commonly-used Gaussian-process (GP) surrogates struggle with cubic scaling costs and overfit to sparse high-fidelity observations, limiting efficiency and generalization in real-world applications. We introduce FIRE, a training-free MF framework that couples tabular foundation models (TFMs) to perform zero-shot in-context Bayesian inference via a high-fidelity correction model conditioned on the low-fidelity model's posterior predictive distributions. This cross-fidelity information transfer via distributional summaries captures heteroscedastic errors, enabling robust residual learning without model retraining. Across 31 benchmark problems spanning synthetic and real-world tasks (e.g., DrivAerNet, LCBench), FIRE delivers a stronger performance-time trade-off than seven state-of-the-art GP-based or deep learning MF regression methods, ranking highest in accuracy and uncertainty quantification with runtime advantages. Limitations include context window constraints and dependence on the quality of the pre-trained TFM's.
format Preprint
id arxiv_https___arxiv_org_abs_2601_22371
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle FIRE: Multi-fidelity Regression with Distribution-conditioned In-context Learning using Tabular Foundation Models
Yu, Rosen Ting-Ying
Sung, Nicholas
Ahmed, Faez
Machine Learning
Multi-fidelity (MF) regression often operates in regimes of extreme data imbalance, where the commonly-used Gaussian-process (GP) surrogates struggle with cubic scaling costs and overfit to sparse high-fidelity observations, limiting efficiency and generalization in real-world applications. We introduce FIRE, a training-free MF framework that couples tabular foundation models (TFMs) to perform zero-shot in-context Bayesian inference via a high-fidelity correction model conditioned on the low-fidelity model's posterior predictive distributions. This cross-fidelity information transfer via distributional summaries captures heteroscedastic errors, enabling robust residual learning without model retraining. Across 31 benchmark problems spanning synthetic and real-world tasks (e.g., DrivAerNet, LCBench), FIRE delivers a stronger performance-time trade-off than seven state-of-the-art GP-based or deep learning MF regression methods, ranking highest in accuracy and uncertainty quantification with runtime advantages. Limitations include context window constraints and dependence on the quality of the pre-trained TFM's.
title FIRE: Multi-fidelity Regression with Distribution-conditioned In-context Learning using Tabular Foundation Models
topic Machine Learning
url https://arxiv.org/abs/2601.22371