Leveraging small language models for Text2SPARQL tasks to improve the resilience of AI assistance

TL;DR

This paper demonstrates that small language models under one billion parameters can effectively translate natural language into SPARQL queries after fine-tuning, enabling accessible AI assistance for semantic web users.

Contribution

It shows that small, affordable language models can be fine-tuned for Text2SPARQL tasks, broadening accessibility and resilience of AI tools in semantic web applications.

Findings

Small models achieve effective translation after fine-tuning

Training success depends on specific data prerequisites

Enables AI assistance on commodity hardware

Abstract

In this work we will show that language models with less than one billion parameters can be used to translate natural language to SPARQL queries after fine-tuning. Using three different datasets ranging from academic to real world, we identify prerequisites that the training data must fulfill in order for the training to be successful. The goal is to empower users of semantic web technology to use AI assistance with affordable commodity hardware, making them more resilient against external factors.