Large Language Model augmented Knowledge Graph embedding

Knowledge graphs are well-known for their ability to organize knowledge in a structured way. They are used in various tasks like question answering, recommendation systems, and web search. However, traditional knowledge graphs often suffer from incompleteness, and existing methods often overlook the incorporation of textual information. To tackle these challenges, recent studies have focused on enhancing knowledge graphs by integrating Language Model-based Models (LLMs). This integration aims to leverage textual information and enhance performance in downstream tasks. In this comparison, there are 8 different models are explored for LLMs as Text Encoders and LLMs for Joint Text and KG Embedding.