Knowledge graphs (KGs) play a crucial role in modern applications. However, automatically constructing a KG from natural language text is challenging due to the complexities of natural languages. This tutorial will focus on the state-of-the-art LLM-based methods, techniques, and tools for constructing knowledge graphs from text, discussing their capabilities, limitations, and current challenges. During the last year, emerging topics such as Retrieval Augmented Generation (RAG), GraphRAG, Chain-of-Thought, LLM Agents, and reasoning models have driven the development of numerous new entity and relation extraction methods that are helpful for KG generation from text.
This talk aims to summarize the research progress in KG construction from text, with a specific focus on the information acquisition branch that includes relation extraction, covering state-of-the-art transformer methods and tools. Our session will explore the theoretical foundations of innovative approaches while also providing practical, hands-on exercises for deeper understanding. This overview will be valuable for both practitioners engaged in building organizational knowledge graphs and academics interested in the cutting edge of research.