Table representation learning for data set discovery and data integration in datalakes
The aim of this thesis proposal is to define and develop new solutions for structured tabular data discovery by learning table representations using Large Language Models (LLMs) and Graph Neural Networks (GNNs). The proposed approach suggests that the underlying transfer learning capabilities and the ability to handle graph-based data provide a robust framework for the challenges of modern data integration, enabling deeper analysis and accurate models for discovering and integrating heterogeneous datasets in a data lake. The scientific approach requires theoretical and practical experience in structured data processing and deep learning.
A. Boutaleb, A. Almutawa, B. Amann, R. Angarita, H. Naacke : “HEARTS: HypErgrAph-based Related Table Search”, ELLIS workshop on Representation Learning and Generative Models for Structured Data, Amsterdam, Netherlands (2025)