K-ON: Stacking Knowledge On the Head Layer of Large Language Model

TL;DR

K-ON introduces a novel method to integrate knowledge graph information into large language models by using multiple head layers for entity-level prediction, improving performance in KG-related tasks.

Contribution

The paper presents K-ON, a new approach that enhances LLMs with KG knowledge through multiple head layers and contrastive learning, addressing granularity mismatch issues.

Findings

K-ON outperforms existing methods in KG tasks.

It enables entity-level predictions in one step.

Contrastive loss improves KG representation learning.

Abstract

Recent advancements in large language models (LLMs) have significantly improved various natural language processing (NLP) tasks. Typically, LLMs are trained to predict the next token, aligning well with many NLP tasks. However, in knowledge graph (KG) scenarios, entities are the fundamental units and identifying an entity requires at least several tokens. This leads to a granularity mismatch between KGs and natural languages. To address this issue, we propose K-ON, which integrates KG knowledge into the LLM by employing multiple head layers for next k-step prediction. K-ON can not only generate entity-level results in one step, but also enables contrastive loss against entities, which is the most powerful tool in KG representation learning. Experimental results show that K-ON outperforms state-of-the-art methods that incorporate text and even the other modalities.