TaDA-Workshop 2024 Best Short Paper Award
2024/09/02
We are proud to share that our recent paper “LLMs for Data Engineering on Enterprise Data” has won the Best Short Paper award of the Tabular Data Analysis Workshop at VLDB'24. Whereas recent studies are optimistic about the use of LLMs for data engineering tasks, we show that LLMs still display severe limitations when faced with real-world usage scenarios like enterprise data.
In this paper we investigate the performance of Large Language Models (LLMs) on data engineering tasks using a real-world enterprise dataset, specifically focusing on the task of column type annotation. The study reveals that LLMs face significant challenges when applied to enterprise data, highlighting the need for further adaptation to improve their accuracy in this context.
The paper can be found here (opens in new tab)
