$R^3$-NL2GQL: A Model Coordination and Knowledge Graph Alignment Approach for NL2GQL
This addresses the challenge of NL2GQL for users of graph databases, offering a novel approach in an emerging field with limited datasets, though it appears incremental as it adapts existing NL2SQL methods to GQL.
The paper tackles the problem of converting natural language to Graph Query Language (NL2GQL) by introducing $R^3$-NL2GQL, a method that integrates small and large foundation models for ranking, rewriting, and refining tasks, and demonstrates promising efficacy and robustness on a newly developed bilingual dataset.
While current tasks of converting natural language to SQL (NL2SQL) using Foundation Models have shown impressive achievements, adapting these approaches for converting natural language to Graph Query Language (NL2GQL) encounters hurdles due to the distinct nature of GQL compared to SQL, alongside the diverse forms of GQL. Moving away from traditional rule-based and slot-filling methodologies, we introduce a novel approach, $R^3$-NL2GQL, integrating both small and large Foundation Models for ranking, rewriting, and refining tasks. This method leverages the interpretative strengths of smaller models for initial ranking and rewriting stages, while capitalizing on the superior generalization and query generation prowess of larger models for the final transformation of natural language queries into GQL formats. Addressing the scarcity of datasets in this emerging field, we have developed a bilingual dataset, sourced from graph database manuals and selected open-source Knowledge Graphs (KGs). Our evaluation of this methodology on this dataset demonstrates its promising efficacy and robustness.