EFVLR first retrieves candidate pages through hybrid multimodal and lexical retrieval followed by cross-modal reranking, then asks a vision-language model to verify direct evidence independently on each labeled candidate page, while original PDF page numbers are recovered deterministically from metadata.
Yan-Jiang Zhou, Shi-Cheng Wang, Yang Liu et al.· Proceedings of the 34th ACM...· 0 citations
Document visual question answering requires models to understand visually rich documents while identifying the pages that support each answer. We present EFVLR, an evidence-first vision-language reasoning framework for multilingual PDF question answering. EFVLR first retrieves candidate pages through hybrid multimodal...
Yan-Jiang Zhou, Shi-Cheng Wang, Yang Liu et al.· Proceedings of the 34th ACM...· 0 citations
GrowPage is introduced, an on-demand KV budgeting framework that treats KV capacity as a runtime resource and achieves a superior performance--throughput trade-off over existing approaches.
Qian-Kun Ma, Yan-Jiang Zhou, Zi-Nan Xiong et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.