Scientific Table Search Using Keyword Queries

Tables are common and important in scientific documents, yet most text-based document search systems do not capture structures and semantics specific to tables. How to bridge different types of mismatch between keywords queries and scientific tables and what influences ranking quality needs to be carefully investigated. This paper considers the structure of tables and gives different emphasis to table components. On the query side, thanks to external knowledge such as knowledge bases and ontologies, key concepts are extracted and used to build structured queries, and target quantity types are identified and used to expand original queries. A probabilistic framework is proposed to incorporate structural and semantic information from both query and table sides. We also construct and release TableArXiv, a high quality dataset with 105 queries and corresponding relevance judgements for scientific table search. Experiments demonstrate significantly higher accuracy overall and at the top of the rankings than several baseline methods.