Leveraging Schema Labels to Enhance Dataset Search

Abstract

A search engine's ability to retrieve desirable datasets is important for data sharing and reuse. Existing dataset search engines typically rely on matching queries to dataset descriptions. However, a user may not have enough prior knowledge to write a query using terms that match with description text. We propose a novel schema label generation model which generates possible schema labels based on dataset table content. We incorporate the generated schema labels into a mixed ranking model which not only considers the relevance between the query and dataset metadata but also the similarity between the query and generated schema labels. To evaluate our method on real-world datasets, we create a new benchmark specifically for the dataset retrieval task. Experiments show that our approach can effectively improve the precision and NDCG scores of the dataset retrieval task compared with baseline methods. We also test on a collection of Wikipedia tables to show that the features generated from schema labels can improve the unsupervised and supervised web table retrieval task as well.

Table Search Using a Deep Contextualized Language Model

“…Trabelsi et al [37] propose custom embeddings for column headers based on multiple contexts for table retrieval, and find representing numerical cell values to be useful. Chen et al [8] utilize matrix factorization to generate additional table headers and then show that those generated headers can improve the performance of unsupervised table search.…”

Modular framework for similarity-based dataset discovery using external knowledge

“…The metadata is then mapped to the Google's knowledge graph, which is then used for dataset duplicates detection and for dataset discovery. Chen et al (2020) enrich metadata records with labels based on the dataset content. Chapman et al (2020) describe the whole dataset discovery process comprising querying for datasets, query processing resulting in a list of datasets, result handling and its presentation.…”

Neural ranking models for document retrieval

“…Zhang & Balog, (2018) propose a semantic matching method for table retrieval where various embedding features are used. Chen et al (2020a) first learn the embedding representations of table headers and generate new headers with embedding features and curated features (Chen et al, 2018) for data tables. They show that the generated headers can be combined with the original fields of the table in order to accurately predict the relevance score of a query-table pair, and improve ranking performance.…”