ATHENA++
Bibliographic record
Abstract
Natural Language Interfaces to Databases (NLIDB) systems eliminate the requirement for an end user to use complex query languages like SQL, by translating the input natural language (NL) queries to SQL automatically. Although a significant volume of research has focused on this space, most state-of-the-art systems can at best handle simple select-project-join queries. There has been little to no research on extending the capabilities of NLIDB systems to handle complex business intelligence (BI) queries that often involve nesting as well as aggregation. In this paper, we present Athena++, an end-to-end system that can answer such complex queries in natural language by translating them into nested SQL queries. In particular, Athena++ combines linguistic patterns from NL queries with deep domain reasoning using ontologies to enable nested query detection and generation. We also introduce a new benchmark data set ( FIBEN ), which consists of 300 NL queries, corresponding to 237 distinct complex SQL queries on a database with 152 tables, conforming to an ontology derived from standard financial ontologies (FIBO and FRO). We conducted extensive experiments comparing Athena++ with two state-of-the-art NLIDB systems, using both FIBEN and the prominent Spider benchmark. Athena++ consistently outperforms both systems across all benchmark data sets with a wide variety of complex queries, achieving 88.33% accuracy on FIBEN benchmark, and 78.89% accuracy on Spider benchmark, beating the best reported accuracy results on the dev set by 8%.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.008 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.004 | 0.006 |
| Open science | 0.005 | 0.006 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.087 | 0.119 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".