DSTI at LLMs4OL 2024 Task A: Intrinsic versus extrinsic knowledge for type classification

TL;DR

This paper compares intrinsic and extrinsic knowledge representations in large language models for ontology learning, highlighting a trade-off between performance and semantic grounding.

Contribution

It introduces semantic towers as an extrinsic knowledge method and evaluates their effectiveness against intrinsic knowledge in LLMs for ontology tasks.

Findings

Extrinsic knowledge offers better semantic grounding.

Intrinsic knowledge yields higher performance.

Trade-off exists between semantic grounding and performance.

Abstract

We introduce semantic towers, an extrinsic knowledge representation method, and compare it to intrinsic knowledge in large language models for ontology learning. Our experiments show a trade-off between performance and semantic grounding for extrinsic knowledge compared to a fine-tuned model intrinsic knowledge. We report our findings on the Large Language Models for Ontology Learning (LLMs4OL) 2024 challenge.