Most available semantic parsing datasets, comprising of pairs of natural utterances and logical forms, were collected solely for the purpose of training and evaluation of natural language understanding systems. As a result, they do not contain any of the richness and variety of natural-occurring utterances, where humans ask about data they need or are curious about. In this work, we release SEDE, a dataset with 12,023 pairs of utterances and SQL queries collected from real usage on the Stack Exchange website. We show that these pairs contain a variety of real-world challenges which were rarely reflected so far in any other semantic parsing dataset, propose an evaluation metric based on comparison of partial query clauses that is more suitable for realworld queries, and conduct experiments with strong baselines, showing a large gap between the performance on SEDE compared to other common datasets.

No abstract

The manual creation and update of a Wigmore Diagram (WD) is laborious and tedious. The system described in this paper supports the human user in building the diagram. The system transforms a list of evidence items into a WD semi-automatically by interleaving human effort and machine strength. This is done using a new data structure, a generalisation of a suffix tree.