Improved NL2SQL based on Multi-layer Expert Network
This addresses performance degradation in NL2SQL for database query systems, but it appears incremental as it builds on existing slot-filling approaches.
The paper tackled the problem of inaccurate SQL generation in NL2SQL due to negative migration in slot-filling methods by introducing MLEG-SQL, a multi-layer expert network that separates feature extraction and task-specific classification, and it was found effective on the WiKSQL dataset.
The Natural Language to SQL (NL2SQL) technique is used to convert natural language queries into executable SQL statements. Typically, slot-filling is employed as a classification method for multi-task cases to achieve this goal. However, slot-filling can result in inaccurate SQL statement generation due to negative migration issues arising from different classification tasks. To overcome this limitation, this study introduces a new approach called Multi-Layer Expert Generate SQL (MLEG-SQL), which utilizes a dedicated multi-task hierarchical network. The lower layer of the network extracts semantic features of natural language statements, while the upper layer builds a specialized expert system for handling specific classification tasks. This hierarchical approach mitigates performance degradation resulting from different task conflicts. The proposed method was evaluated on the WiKSQL dataset and was found to be effective in generating accurate SQL statements.