TAHOE: New System Improves Text-to-SQL Performance Through Automated Hint Optimization
Researchers have introduced TAHOE, a system that improves Large Language Model performance on Text-to-SQL tasks by automatically learning and organizing reusable hints from past errors. The system addresses key production challenges—strict SQL dialects, large schemas, and evolving user needs—without requiring costly model fine-tuning. On a supervised benchmark using GPT-5.5, TAHOE raised pass rates from 61.95% to 79.42% and achieved 100% Snowflake syntax compliance.
TAHOE (Text-to-SQL with Automated Hint Optimization from Experience) is a prompt optimization framework that treats hint management as a dynamic data management problem rather than a model training challenge. It operates through an error-driven pipeline that consolidates debugging traces into a structured Hint Bank, separating dialect-specific Syntax Hints from schema- and user-specific Semantic Hints. A Strategy Layer handles conflicting user intents by modeling them as competing strategies under shared natural-language triggers, using recency signals and attribution statistics to track each strategy's empirical success, harm, inertness, and support. At inference time, the system retrieves relevant hints and guides the LLM through a two-stage process of Logic Planning followed by SQL Synthesis. Evaluated on the Spider 2.0-Snow benchmark with 113 supervised examples, TAHOE improved pass-at-4 from 72.57% to 87.61% and dramatically reduced compiler-feedback critic rounds from 2.79 to 0.12 per candidate. Notably, the learned Hint Bank transferred to weaker backbone models, yielding a 19.7 percentage-point pass-rate gain on Doubao-2.0-lite, suggesting broad applicability across model tiers. The deployment-phase component incorporating live human feedback is identified as future work.
What's missing
The study evaluates only the development-phase workflow; deployment-time human-feedback updates are explicitly left for future work, leaving real-world production performance unvalidated. The benchmark used (Spider 2.0-Snow) is Snowflake-specific, so generalizability to other SQL dialects or database platforms is not established. The paper does not report computational overhead or latency costs of hint retrieval at inference time, which are relevant for production feasibility assessments.
What different sources said
- arXiv cs.AICenter
TAHOE: Text-to-SQL with Automated Hint Optimization from Experience
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.