Automated Construction of a Knowledge Graph of Nuclear Fusion Energy for Effective Elicitation and Retrieval of Information

Domain Knowledge Structuring: The paper addresses the need to structure and represent domain-specific knowledge from large document corpora.

Specialized Domain Heterogeneity: Nuclear fusion energy is presented as a highly specialized field with broad scope and heterogeneous information.

Entity Extraction And Resolution: Automatic named entity recognition and entity resolution are identified as key challenges for the pipeline.

Knowledge Graph Question Answering: The paper addresses answering natural-language queries, including multihop questions requiring reasoning over interconnected entities.

Domain Knowledge Structuring: The paper addresses the need to structure and represent domain-specific knowledge from large document corpora.

Automated KG Construction Pipeline: The authors propose a multistep approach for automated construction of a domain-specific knowledge graph from large document corpora.

LLM-Based Entity Processing: The authors use pretrained large language models to address named entity recognition and entity resolution challenges.

Knowledge-Graph RAG System: The authors develop a knowledge-graph retrieval-augmented generation system using multiple prompts with LLMs for natural-language query answering.

Automated KG Construction Pipeline: The authors propose a multistep approach for automated construction of a domain-specific knowledge graph from large document corpora.

LLM-Based Entity Processing: The authors use pretrained large language models to address named entity recognition and entity resolution challenges.

First Nuclear Fusion Energy KG:

LLM Performance Evaluated Against Zipf:

Contextually Relevant Query Answers: