Research focuses on ML‑augmented data management: integrating heterogeneous data lakes, resolving entities with minimal labels, and enabling semantic table search and virtual knowledge graph mappings. Methods include active learning, graph neural networks, and ordinal‑aware loss design.