专利新颖性评估数据集
收藏资源简介:
该数据集是专门为评估专利新颖性而设计的,包含了实际专利审查案例中专利声明与所引用先前艺术文档的对。数据集来源于美国专利商标局(USPTO)的审查案例,涵盖了2014至2015年间申请的、属于IPC G06F2技术领域的专利。数据集由专利声明和引用的先前艺术文档组成,分为未经修改的原始声明(标记为'非新颖')和经过修改后的声明(标记为'新颖'),旨在解决专利审查中的新颖性评估问题。
This dataset is specifically designed for evaluating patent novelty, and contains pairs of patent claims and their cited prior art documents from real patent examination cases. It is sourced from examination cases of the United States Patent and Trademark Office (USPTO), covering patents filed between 2014 and 2015 that belong to the IPC G06F2 technical domain. The dataset, which consists of patent claims and their cited prior art documents, is categorized into two groups: unmodified original claims marked as 'non-novel' and revised claims marked as 'novel', with the goal of addressing the novelty assessment issue in patent examination.

- 1Can AI Examine Novelty of Patents?: Novelty Evaluation Based on the Correspondence between Patent Claim and Prior Art卡内基梅隆大学语言技术研究所 · 2025年



