Tag
This paper introduces ARI, a framework that uses retrieval-augmented large language models to restore illegible portions of historical documents, significantly improving named entity restoration by combining implicit LLM knowledge with explicitly retrieved external historical context.