Improving Access to Historical Archives with Real-time RAG-based Systems
2026-07-03 • Information Retrieval
Information Retrieval
AI summaryⓘ
The authors worked on improving how people search through old historical archives that have lots of text errors due to poor OCR (optical character recognition). They used large language models (LLMs) to clean up the messy text and to make searching smarter by understanding natural language questions better. Their experiments on Swiss newspaper archives showed that the text got much clearer and search results became more accurate compared to traditional keyword methods. This approach helps turn simple digital archives into systems that understand and respond to user questions more like a conversation.
Optical Character Recognition (OCR)Large Language Models (LLMs)Retrieval-Augmented Generation (RAG)Cross-Encoder RerankingNatural Language Question AnsweringInformation RetrievalCharacter Error Rate (CER)Word Error Rate (WER)NDCG@10Digital Archives
Authors
Stergios Konstantinidis, Hayman Lotfy, Alexis Erne, Faruk Zahiragic, Min-Yen Kan, Michalis Vlachos
Abstract
Digitized historical archives are large, heterogeneous cultural heritage repositories, but access methods for such archives face challenges such as noisy optical character recognition (OCR) output and rigid keyword-based retrieval, which limit retrieval quality. In this work, we present an end-to-end archival processing and retrieval framework that integrates large language models (LLMs) into the archival pipeline. Our system introduces two core components: (i) an LLM-based OCR refinement module that improves text quality, and (ii) a semantic retrieval and cross-encoder reranking pipeline supporting natural-language question answering via retrieval-augmented generation (RAG). Our evaluations are done on a historical archival dataset of 500,000 Swiss newspaper segments spanning over three centuries (1762 to 2001). Experiments are conducted across 384 natural-language test queries. Our results highlight that LLM refinements reduce OCR errors by up to 44.52% (CER) and 60.95% (WER). More importantly, this is accompanied by downstream information retrieval improvements. Compared to traditional keyword baselines, our reranking pipeline increases NDCG@10 by 31.9% (from 65.99% to 87.05%) and achieves statistically significant gains in both answer correctness and context relevance. These results demonstrate that integrating LLMs with established document processing and retrieval pipelines can elevate digital libraries from static repositories to interactive, semantically searchable archival systems.