Table2answer: Read the database and answer without SQL
This work addresses the challenge of hard label work in semantic parsing for database question answering, though it appears incremental as it matches rather than surpasses baseline performance.
The paper tackles the problem of semantic parsing for question answering by proposing to bypass explicit logic form generation, using only schema and answer information, and demonstrates that their BERT-based model achieves baseline results on the WikiSQL dataset.
Semantic parsing is the task of mapping natural language to logic form. In question answering, semantic parsing can be used to map the question to logic form and execute the logic form to get the answer. One key problem for semantic parsing is the hard label work. We study this problem in another way: we do not use the logic form any more. Instead we only use the schema and answer info. We think that the logic form step can be injected into the deep model. The reason why we think removing the logic form step is possible is that human can do the task without explicit logic form. We use BERT-based model and do the experiment in the WikiSQL dataset, which is a large natural language to SQL dataset. Our experimental evaluations that show that our model can achieves the baseline results in WikiSQL dataset.