作者位次 #2/7ACL 2026Poster· ACL Industry Track
When Good OCR Is Not Enough: Benchmarking OCR Robustness for Retrieval-Augmented Generation
Lin Sun · Wangdexian · Jingang Huang · Linglin Zhang · Change Jia · Zhengwei Cheng · Xiangzheng Zhang
摘要
Industrial Retrieval-Augmented Generation (RAG) systems depend on optical character recognition (OCR) to transform visual documents into text. Existing OCR benchmarks rely on character-level metrics, which inadequately measure downstream RAG effectiveness under real-world conditions. We introduce...