Query Assistant for Conversational Database Access Using LLM

DOI

10.2991/978-94-6463-866-0_25How to use a DOI?

Keywords

Natural Language to SQL (NL2SQL); Large Language Models (LLMs); T5 model; Llama 3.1; Gemma 3; Phi4; Qwen; SQL query translation; language model fine-tuning; model performance evaluation; complex queries; database schemas; accuracy metrics; modern AI models; comparative analysis; implementation considerations; consumer-grade hardware; AI democratization; machine learning models; advanced NLP models; SQL query generation; natural language processing (NLP); research in AI; AI model deployment; performance benchmarks

Abstract

This comprehensive report presents an in-depth analysis of our groundbreaking research on improving Natural Language to SQL (NL2SQL) conversion through the application of modern Large Language Models (LLMs). We meticulously fine-tuned and rigorously evaluated five distinct language models—T5 (serving as our baseline), Llama 3.1, Gemma 3, Phi4, and Qwen—measuring their capabilities in accurately translating natural language queries into precise, executable SQL queries. Our extensive findings demonstrate that modern LLMs substantially outperform traditional approaches across all metrics, with particularly dramatic improvements observed for complex queries and large database schemas. Llama 3.1 emerged as the definitive top performer with exceptional accuracy rates, followed closely by Qwen. This report provides exhaustive performance metrics, multilayered comparative analyses, and detailed implementation considerations that thoroughly document how these sophisticated models can be efficiently fine-tuned and deployed even using consumer-grade hardware, democratizing access to this powerful technology.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - K. Akila
AU  - P. Revanth Balaji
AU  - P. Sri Sai Rishith
AU  - V. Sai Sankar Raju
PY  - 2025
DA  - 2025/10/31
TI  - Query Assistant for Conversational Database Access Using LLM
BT  - Proceedings of the International Conference on Intelligent Systems and Digital Transformation (ICISD 2025)
PB  - Atlantis Press
SP  - 286
EP  - 307
SN  - 2589-4919
UR  - https://doi.org/10.2991/978-94-6463-866-0_25
DO  - 10.2991/978-94-6463-866-0_25
ID  - Akila2025
ER  -