Generating Table Vector Representations Article Swipe

PDF

Related Concepts

Table (database) Computer science Task (project management) Artificial intelligence Annotation Natural language processing Semantics (computer science) Encoding (memory) Information retrieval Machine learning Class (philosophy) Data mining Theoretical computer science Programming language Management Economics

Aneta Koleva , Martin Ringsquandl , Mitchell Joblin , Volker Tresp ·

YOU? · · 2021 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2110.15132 · OA: W3208020742

High-quality Web tables are rich sources of information that can be used to populate Knowledge Graphs (KG). The focus of this paper is an evaluation of methods for table-to-class annotation, which is a sub-task of Table Interpretation (TI). We provide a formal definition for table classification as a machine learning task. We propose an experimental setup and we evaluate 5 fundamentally different approaches to find the best method for generating vector table representations. Our findings indicate that although transfer learning methods achieve high F1 score on the table classification task, dedicated table encoding models are a promising direction as they appear to capture richer semantics.

Keywords: Table (database) · Computer science · Task (project management) · Artificial intelligence · Annotation · Natural language processing · Semantics (computer science) · Encoding (memory) · Information retrieval · Machine learning · Class (philosophy) · Data mining · Theoretical computer science · Programming language