发表机构
Dalian Ocean University（大连海洋大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对长篇文学叙事的信息处理挑战，提出NS-ST-GraphRAG神经符号时空GraphRAG框架，构建首个中国古典文学多跳问答基准Red-Chamber-QA，实验显示其机械答案复现率与语义判断准确率优于基线模型。
AI 中文摘要
长篇文学叙事对检索增强生成（RAG）构成独特的信息处理挑战：相关证据分布在不同章节，关系随叙事时间演变，正确答案可能同时依赖时间、空间和关系约束。我们提出NS-ST-GraphRAG，这是一个神经符号时空GraphRAG框架，整合了本体引导提取、确定性约束检查、双时间坐标、空间场景属性及动态子图检索。该框架不从单一语料库级图中检索，而是选择查询时间和空间范围内有效的图状态，并将生成的答案基于可追溯证据。我们还引入Red-Chamber-QA，据我们所知，这是首个针对中国古典文学的开放多跳问答基准，包含时间、空间和通用问题类别、每部分证据跨度及确定性捷径控制。在120个问题的保留测试集上，NS-ST-GraphRAG的机械答案复现率为0.733，而冻结窗口基线为0.675，闭卷模型为0.083（McNemar精确检验p=0.092，方向有利但不显著）；语义判断准确率为0.866，而冻结窗口基线为0.850。H2的预设约束类别条件未得到所做比较的支持。这些结果表明，时间图表示、约束提取和可审计评估如何整合为一个统一框架，用于长篇叙事的可验证知识处理。
英文摘要
Long-form literary narratives pose a distinctive information-processing challenge for retrieval-augmented generation: relevant evidence is distributed across chapters, relations evolve over narrative time, and correct answers may depend jointly on temporal, spatial, and relational constraints. We propose NS-ST-GraphRAG, a neuro-symbolic spatio-temporal GraphRAG framework that integrates ontology-guided extraction, deterministic constraint checking, dual temporal coordinates, spatial scene attributes, and dynamic sub-graph retrieval. Instead of retrieving from a single corpus-level graph, the framework selects the graph state valid for the temporal and spatial scope of a query and grounds generated answers in traceable evidence. We further introduce Red-Chamber-QA, to our knowledge the first open multi-hop question-answering benchmark for classical Chinese literature, with time-, space-, and general-question categories, per-part evidence spans, and deterministic shortcut controls. On a 120-question held-out split, NS-ST-GraphRAG achieves mechanical answer reproduction of 0.733 versus 0.675 for the frozen window baseline and 0.083 for a closed-book model (McNemar exact p = 0.092, directionally favorable but not significant); semantic-judge accuracy is 0.866 versus 0.850. The pre-specified constrained-category condition of H2 is not supported by the delivered comparison. These results show how temporal graph representation, constrained extraction, and auditable evaluation integrate into a unified framework for verifiable knowledge processing over long-form narrative.
CommentsSubmitted to Information Processing and Management
