Tackling prediction tasks in relational databases with LLMs
This work addresses the challenge of using LLMs for relational database tasks, establishing them as a new baseline for machine learning in this domain.
The paper tackled the problem of applying large language models (LLMs) to predictive tasks in relational databases, showing that a straightforward application achieves competitive performance on the RelBench benchmark.
Though large language models (LLMs) have demonstrated exceptional performance across numerous problems, their application to predictive tasks in relational databases remains largely unexplored. In this work, we address the notion that LLMs cannot yield satisfactory results on relational databases due to their interconnected tables, complex relationships, and heterogeneous data types. Using the recently introduced RelBench benchmark, we demonstrate that even a straightforward application of LLMs achieves competitive performance on these tasks. These findings establish LLMs as a promising new baseline for ML on relational databases and encourage further research in this direction.