GNN: Graph Neural Network and Large Language Model for Data Discovery

TL;DR

This paper introduces GNN, a novel algorithm combining graph neural networks and large language models to improve data discovery by understanding both numerical and text data, reducing human input and predefinition requirements.

Contribution

GNN extends previous data discovery methods by integrating GNNs and LLMs to better interpret text values and user preferences, enhancing prediction reliability.

Findings

Improved understanding of text type values in data discovery.

Reduced need for human input and predefined utility functions.

Enhanced prediction accuracy for diverse data types.

Abstract

Our algorithm GNN: Graph Neural Network and Large Language Model for Data Discovery inherit the benefits of \cite{hoang2024plod} (PLOD: Predictive Learning Optimal Data Discovery), \cite{Hoang2024BODBO} (BOD: Blindly Optimal Data Discovery) in terms of overcoming the challenges of having to predefine utility function and the human input for attribute ranking, which helps prevent the time-consuming loop process. In addition to these previous works, our algorithm GNN leverages the advantages of graph neural networks and large language models to understand text type values that cannot be understood by PLOD and MOD, thus making the task of predicting outcomes more reliable. GNN could be seen as an extension of PLOD in terms of understanding the text type value and the user's preferences, not only numerical values but also text values, making the promise of data science and analytics purposes.