A fine-tuned LLM can generate valid execution plans that beat traditional and learned optimizers on three workloads, by distilling the best plans from multiple existing optimizers.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DB 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Can Large Language Models Be Query Optimizer for Relational Databases?
A fine-tuned LLM can generate valid execution plans that beat traditional and learned optimizers on three workloads, by distilling the best plans from multiple existing optimizers.