CHORUS: Foundation Models for Unified Data Discovery and Exploration

TL;DR

This paper demonstrates that foundation models, especially large language models, excel in data discovery and exploration tasks, outperforming traditional models and sometimes even human experts, indicating a unified future approach for data management.

Contribution

The paper introduces a foundation-model-based approach for data discovery tasks, showing its superiority over task-specific models and establishing a new direction for unified data management.

Findings

Foundation models outperform task-specific models on key data discovery tasks.

The approach often surpasses human-expert performance.

The method generalizes across multiple foundation models and handles non-determinism.

Abstract

We apply foundation models to data discovery and exploration tasks. Foundation models include large language models (LLMs) that show promising performance on a range of diverse tasks unrelated to their training. We show that these models are highly applicable to the data discovery and data exploration domain. When carefully used, they have superior capability on three representative tasks: table-class detection, column-type annotation and join-column prediction. On all three tasks, we show that a foundation-model-based approach outperforms the task-specific models and so the state of the art. Further, our approach often surpasses human-expert task performance. We investigate the fundamental characteristics of this approach including generalizability to several foundation models and the impact of non-determinism on the outputs. All in all, this suggests a future direction in which disparate data management tasks can be unified under foundation models.