重思结构化生成：基于图的推理能否解决歧义？
原标题：Rethinking Structured Generation: Can Graph-Based Reasoning Resolve Ambiguity?
AI 摘要
该论文指出结构化生成通常被建模为从输入到单一结构化输出的确定性映射，这与自然语言常存在多种结构不同但语义有效的解释相矛盾，单参考监督会将歧义压缩为单一预测，导致输出不稳定。作者将结构化生成重新表述为歧义感知推理问题，采用多有效学习框架在优化中保留歧义，构建候选结构化输出空间并将每个候选表示为图，通过基于图的语义对齐与关系一致性推理进行选择。在结构化生成基准上的实验显示，该方法在准确率、稳定性和鲁棒性上均有提升，作者据此主张从单输出预测转向集合级推理。
正文节选
Rethinking Structured Generation: Can Graph-Based Reasoning Resolve Ambiguity? Abstract Structured generation is typically formulated as a deterministic mapping from input to a single structured output. This assumption is fundamentally misaligned with natural language, which often admits multiple structurally distinct yet semantically valid interpretations. As a result, existing methods rely on single-reference supervision that collapses ambiguity into one prediction, leading to unstable and inc