I briefly scanned your GitHub repo, and it seeemst o me that you adopted the KG but you still use SQL query as the output. Unlike the original paper, where it uses KG and produces SPARQL as output. I suppose the key is the KG. Not so much whether SQL or SPARQL is used right?
I briefly scanned your GitHub repo, and it seeemst o me that you adopted the KG but you still use SQL query as the output. Unlike the original paper, where it uses KG and produces SPARQL as output. I suppose the key is the KG. Not so much whether SQL or SPARQL is used right?
That's exactly right - we produced SQL which is how the dbt Semantic Layer interacts with data platforms. The great things about this benchmark is its agnostic to the querying language - it's validating correctness of results. Our findings confirmed the original paper that layering semantic information into the natural language query yields superior results.
I briefly scanned your GitHub repo, and it seeemst o me that you adopted the KG but you still use SQL query as the output. Unlike the original paper, where it uses KG and produces SPARQL as output. I suppose the key is the KG. Not so much whether SQL or SPARQL is used right?
That's exactly right - we produced SQL which is how the dbt Semantic Layer interacts with data platforms. The great things about this benchmark is its agnostic to the querying language - it's validating correctness of results. Our findings confirmed the original paper that layering semantic information into the natural language query yields superior results.