Ad Hoc Table Retrieval using Semantic Similarity

Abstract

We introduce and address the problem of ad hoc table retrieval: answering a keyword query with a ranked list of tables. This task is not only interesting on its own account, but is also being used as a core component in many other table-based information access scenarios, such as table completion or table mining. The main novel contribution of this work is a method for performing semantic matching between queries and tables. Specifically, we (i) represent queries and tables in multiple semantic spaces (both discrete sparse and continuous dense vector representations) and (ii) introduce various similarity measures for matching those semantic representations. We consider all possible combinations of semantic representations and similarity measures and use these as features in a supervised learning model. Using a purpose-built test collection based on Wikipedia tables, we demonstrate significant and substantial improvements over a state-of-the-art baseline.

Web Table Extraction, Retrieval and Augmentation

“…Ahmadov et al (2015) Data completion Das Sarma et al (2012) Schema complement Entity complement Relation join Limaye et al (2010) To decide if two rows are duplicates of each other, they employ the method in (Gupta and Sarawagi, 2009). Zhang and Balog (2018a) perform semantic matching between queries and tables for keyword table search. Specifically, they (1) represent queries and tables in multiple semantic spaces (both discrete sparse and continuous dense vector representations) and (2) introduce various similarity measures for matching those semantic representations.…”

“…One limitation of existing work is that it often makes assumptions about underlying query intent and the preferred answer table types. For example, Zhang and Balog (2018a) assume that queries follow a class-property pattern, which can be successfully answered by relational tables. As a result, relational tables with this pattern are preferred, which might therefore result in lower coverage.…”

“…We organize relevant literature based on the task that is being addressed into six main categories. These are: ; Chen and Cafarella (2013); Cafarella et al (2008b); Balakrishnan et al (2015); Cafarella et al (2009); ; Bhagavatula et al (2015) Wang and Hu (2002b,a); ; Chen and Cafarella (2013); Cafarella et al (2008b); Crestan and Pantel (2011); Lautert et al (2013); Nishida et al (2017); Venetis et al (2011); Mulwad et al (2010); Fan et al (2014); Bhagavatula et al (2015); ; Efthymiou et al (2017); ; Hassanzadeh et al (2015); Mulwad et al (2013); Sekhavat et al (2014); Ibrahim et al (2016); Limaye et al (2010); Muñoz et al (2014); ; Ritze and Bizer (2017) Table search Query Ranked list of tables Cafarella et al (2009); Pimplikar and Sarawagi (2012); Cafarella et al (2008a); Bhagavatula et al (2013); ; ; Das Sarma et al (2012); Yakout et al (2012); Nguyen et al (2015); Zhang and Balog (2018a); Limaye et al (2010); Nargesian et al (2018); Zhang and Balog (2019b) Question Answering Natural language query Structured data Pasupat and Liang (2015); Sun et al (2016);…”

Table2Vec

“…Request the header row with additional headings. Table retrieval is the task of returning a ranked list of tables for a keyword query [15]. Prior table-related work has considered embeddings, both pretrained ones and task-specific ones.…”

Auto-completion for Data Cells in Relational Tables

“…TC) The WikiTables corpus[4] is extracted from Wikipedia and contains 1.6M high-quality tables. Following[24], we select the core column by taking the one among the left-most 2 columns with the highest entity rate. Based on a sample of 100 tables, this method has over 98% accuracy.…”