Google Introduces TabFM, a Zero-Shot Foundation Model for Tabular Data

Google Research has unveiled TabFM, a foundation model designed to work with tabular data without requiring task-specific training. Unlike traditional machine learning models that need to be trained separately for each dataset, TabFM can generalize across unseen tabular tasks in a zero-shot manner. The model aims to bring the power of large foundation models — previously dominant in text and image domains — to structured, table-based data. Google detailed the development and capabilities of TabFM in a post on its official research blog. The release marks a notable step toward making AI more versatile for the vast amounts of structured data used across industries.