Get Started
Home
Topics
Search
Library
Research questionCan tabular foundation models learn transferable physical laws with units and noiseless mechanisms, not just interpolate data?High predictive accuracy on equation-generated tables may reflect interpolation rather than representation of governing physics. Physical modeling also requires handling units and noiseless mechanisms, which table completion may not capture.
AI
Evaluation & Benchmarks
Machine Learning
Research Paper
Statistical Machine Learning
Technology
Latest papersRecent research connected to this question, newest first.Do Tabular Foundation Models Know Physics? Contamination, Units, and the Deterministic LimitEvidence covers four tabular foundation models and six baselines evaluated on datasets sampled from 316 physical equations, both out of the box and after tuning. The reported limitations concern representing physical units and noiseless mechanisms; the evidence does not establish that these models can act as physical models.research paper · Sep 2, 2026
Related questions
How can reliable foundation-model scaling laws be constructed without training every configuration?How should model architecture and size scale with data for noisy tabular ratemaking?How should wireless foundation models be evaluated for transferable representations under task and distribution shifts?How can general-purpose tabular predictors balance adaptation-free accuracy with computational cost?