Global Reasoning over Database Structures for Text-to-SQL Parsing

By Ben Bogin, Matt Gardner, Jonathan Berant. 2019

115 citations recorded by the source corpus.

State-of-the-art semantic parsers rely on auto-regressive decoding, emitting one symbol at a time. When tested against complex databases that are unobserved at training time (zero-shot), the parser often struggles to select the correct set of database constants in the new database, due to the local nature of decoding. %since their decisions are based on weak, local information only. In this work, we propose a semantic parser that globally reasons about the structure of the output query to make a more contextually-informed selection of database constants. We use message-passing through a graph neural network to softly select a subset of database constants for the output query, conditioned on the question. Moreover, we train a model to rank queries based on the global alignment of database constants to question words. We apply our techniques to the current state-of-the-art model for Spider, a zero-shot semantic parsing dataset with complex databases, increasing accuracy from 39.4% to 47.4%.

Subfields: Artificial Intelligence, Natural Language Processing, Information Extraction

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

In Information Extraction

Citations in the corpus

UnifiedSKG: Unifying and Multi-Tasking Structured Knowledge Grounding with Text-to-Text Language Models (2022, methodology): “Structured knowledge encoding has been widely explored (Bogin et al., 2019; Lin et al., 2019; Agar-wal et al., 2020; Saxena et al., 2020; Yasunaga and Liang, 2020; Yasunaga et al., 2022; and others detailed in Section 2).”
RAT-SQL: Relation-Aware Schema Encoding and Linking for Text-to-SQL Parsers (2019, methodology): “Then we use a cross-entropy loss, referred as alignment loss, to strengthen the model’s belief: align_loss =− 1 |Rel(C)| ∑ j∈Rel(C) logmax i Lcoli,j − 1 |Rel(T )| ∑ j∈Rel(T ) logmax i Ltabi,j where Rel(C) and Rel(T ) denote the set of relevant columns and tables that appear in the SQL.”
Benchmarking Meaning Representations in Neural Semantic Parsing (2020, methodology): “Since it is domain-agnostic and has wellestablished execution engines, the subtask of semantic parsing, Text-to-SQL, has received a lot of interests.”
RESDSQL: Decoupling Schema Linking and Skeleton Parsing for Text-to-SQL (2023, methodology): “Bogin, Gardner, and Berant (2019) calculate a relevance score for each schema item, which is then used as the soft coefﬁcient of the schema items in the subsequent graph encoder.”
TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data (2020, methodology): “Global–GNN (Bogin et al., 2019a) 52.7 EditSQL + B ERT (Zhang et al., 2019a) 57.6 RatSQL (Wang et al., 2019a) 60.9 IRNet + B ERT (Guo et al., 2019) 60.3 + Memory + Coarse-to-Fine 61.9 IRNet V2 + B ERT 63.9 RyanSQL + B ERT (Choi et al., 2020) 66.6 Our System based on TranX (Yin and Neubig, 2018…”
Photon: A Robust Cross-Domain Text-to-SQL System (2020, background): “GNN (Bogin et al., 2019a) 40.7 Global-GNN (Bogin et al., 2019b) 52.7 EditSQL + BERT (Zhang et al., 2019) 57.6 GNN+Bertrand-DR † (Kelkar et al., 2020) 57.9 EditSQL+Bertrand-DR † (Kelkar et al., 2020) 58.5 IRNet + BERT (Guo et al., 2019) 61.9 RYANSQL + BERT † (Choi et al., 2020) 66.6 P HOTON 63.2 † de”