Document Haystacks: Vision-Language Reasoning Over Piles of 1000+ Documents

By Jun Chen, Dan-Nong Xu, Junjie Fei, Chun-Mei Feng, Mohamed Elhoseiny. 2024

Subfields: Information Retrieval & Search, Prompting & Reasoning, Multimodal Learning

Citations in the corpus

• Invoice Haystack: Benchmarking Document Retrieval and Visual Question Answering Under Strong Visual Homogeneity (2026, unknown): “Two invoices from the same template may exhibit cosine similarity exceeding 0.95 in vision-only embeddings [12] yet contain completely different semantic content distinguishable only through text analysis.”