Table annotation

Table annotation is a task in the field of natural language processing that aims to annotate cells, columns, and rows in tables with terms or concepts from a knowledge graph or database schema. This task primarily consists of five subtasks: cell entity annotation, column type annotation, column property annotation, table type detection, and row annotation. The goal of table annotation is to enhance the semantic understanding and interpretability of tabular data, enabling it to better connect with external knowledge bases, thereby supporting more efficient data integration and querying. The SemTab challenge is held annually, focusing on the first three subtasks, to evaluate and compare different table annotation systems.