Learning to Transpile AMR into SPARQL
This addresses the challenge of improving KBQA by leveraging pre-trained parsers with minimal paired data, though it is incremental as it builds on existing relations between AMR and SPARQL.
The paper tackles the problem of transpiling Abstract Meaning Representation (AMR) into SPARQL for Knowledge Base Question Answering (KBQA), resulting in a model that outperforms recent approaches on datasets like DBPedia and Wikidata.
We propose a transition-based system to transpile Abstract Meaning Representation (AMR) into SPARQL for Knowledge Base Question Answering (KBQA). This allows us to delegate part of the semantic representation to a strongly pre-trained semantic parser, while learning transpiling with small amount of paired data. We depart from recent work relating AMR and SPARQL constructs, but rather than applying a set of rules, we teach a BART model to selectively use these relations. Further, we avoid explicitly encoding AMR but rather encode the parser state in the attention mechanism of BART, following recent semantic parsing works. The resulting model is simple, provides supporting text for its decisions, and outperforms recent approaches in KBQA across two knowledge bases: DBPedia (LC-QuAD 1.0, QALD-9) and Wikidata (WebQSP, SWQ-WD).