Search papers, labs, and topics across Lattice.
This paper introduces SAGE, a multi-agent framework that enhances Chinese ancient document understanding by shifting from direct answer generation to evidence-grounded inference. By coordinating specialized agents for planning, evidence acquisition, and verification, SAGE effectively addresses the limitations of current Large Vision-Language Models (LVLMs) that often produce overconfident and poorly grounded responses. Experimental results on the AncientDoc benchmark demonstrate that SAGE consistently outperforms traditional direct-answering approaches, even surpassing larger models, underscoring the value of structured inference in this domain.
SAGE reveals that a multi-agent, evidence-grounded approach can outperform larger models in understanding Chinese ancient documents, challenging the notion that size alone drives performance.
Chinese ancient document understanding demands complex visual, linguistic, and historical reasoning. Current Large Vision-Language Models (LVLMs) typically rely on an opaque, single-pass generation paradigm, often producing overconfident and weakly grounded responses. To address this, we propose SAGE, an evidence-grounded multi-agent framework that reformulates Chinese ancient document understanding as evidence-grounded inference rather than direct answer generation. SAGE coordinates specialized agents for task-aware planning, tool-mediated evidence acquisition, claim-level verification, and bounded replanning under a constrained shared-state runtime. This design supports bounded evidence seeking, answer revision, and abstention when grounding is insufficient. Experiments on the AncientDoc benchmark show that SAGE consistently outperforms matched direct-answering baselines across three LVLM backbones. Remarkably, SAGE with Qwen3.5-9B surpasses much larger monolithic LVLMs on most evaluated metrics, highlighting the importance of structured, evidence-grounded inference beyond model scaling.
