让AI生成专利时自动理清各条款的依赖关系，避免逻辑混乱。
Pointer-Augmented Autoregressive Generation of Patent Claims with Joint Topology and Content Decoding
- 用指针头在生成时同步预测条款依赖结构
- 恢复79%真实父节点链接，一致性提升至0.478
- 适合需要严谨逻辑的专利、法律文本生成场景
自回归解码器生成的是扁平的词元序列，无法强制输出段落间的层级约束，在专利权利要求生成中尤为明显——权利要求集构成一个作用域逐层收缩的依赖森林。拓扑与内容相互依赖：子项措辞需反映父项范围，但父项必须在子项存在前选定，因此事后解析或语法约束均不足。我们提出SPG（结构感知专利生成），在自回归过程中内嵌拓扑预测。指针头选择每个从属权利要求的父项，其梯度与深度自适应作用域正则化共同重塑共享解码器表示。第二阶段对自生成的错误候选施加违规加权偏好目标，提供授予专利语料库所缺乏的负信号。在HUPD-DCG数据集上，基于Llama-3-8B-Instruct的SPG恢复了79.0%的真实父链接（训练奖励未监督此量），将先行一致性从0.292提升至0.478，专家评估验证了提升效果。
原文摘要 · Abstract (English)
Autoregressive decoders emit flat token sequences and cannot enforce hierarchical constraints across output segments, a limitation that becomes acute in patent claim generation, where a claim set forms a dependency forest whose scope must narrow monotonically with depth. Topology and content are mutually dependent: a dependent claim's wording must reflect its parent's scope, yet the parent must be chosen before that wording exists, so neither post-hoc parsing nor grammar-constrained decoding suffices. We propose SPG (Structure-aware Patent Generation), which predicts topology inside the autoregressive pass. A pointer head selects each dependent claim's parent, and its gradients, together with a depth-adaptive scope regularizer, reshape the shared decoder's representations during training. A second stage then applies a violation-weighted preference objective over self-generated deficient candidates, supplying the negative signal that granted-patent corpora lack. On HUPD-DCG, SPG on Llama-3-8B-Instruct recovers 79.0\% of gold parent links, a quantity its training reward never supervises, and raises antecedent consistency from 0.292 to 0.478 over a supervised baseline of equal scale, with expert evaluation corroborating these gains.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。