电子信息制造高质量专利数据集

电子信息制造高质量专利数据集面向电子信息制造领域，深度融合跨域语义融合、动态知识管控等前沿技术，构建覆盖数据全生命周期的价值转化体系。数据集总规模不低于100GB，涵盖四大核心模块：（1）不低于100组"技术交底书—专利文案"对齐数据，实现研发语言与专利语言的精准语义映射；（2）覆盖电子类、机械类不少于10个细分技术方向的行业专利文本样本；（3）不少于50套专利写作规范、结构模板及错误案例库；（4）常见撰写错误与审核要点数据。数据集经过脱敏、结构化处理，100%合规可用，支持专利生成模型训练、关键词检索增强及企业知识产权管理体系建设。

The High-Quality Patent Dataset for Electronic Information Manufacturing is targeted at the electronic information manufacturing industry. It deeply integrates cutting-edge technologies such as cross-domain semantic fusion and dynamic knowledge management, and establishes a value realization system covering the entire data lifecycle. The total size of the dataset is no less than 100 GB, and it includes four core modules:
1. No less than 100 sets of aligned data between technical disclosure documents and patent documents, enabling accurate semantic mapping between R&D language and patent language;
2. Industry patent text samples covering no fewer than 10 subdivided technical directions in electronic and mechanical fields;
3. No less than 50 sets of patent writing specifications, structural templates and error case bases;
4. Data on common writing errors and patent examination key points.
The dataset has undergone de-identification and structured processing, and is 100% compliant and usable. It supports the training of patent generation models, keyword retrieval augmentation, and the construction of enterprise intellectual property management systems.

搜集汇总

背景与挑战

背景概述

该数据集面向电子信息制造领域，深度融合前沿技术，构建覆盖数据全生命周期的高质量专利价值转化体系。其规模不低于100GB，包含四大核心模块：技术交底书与专利文案对齐数据、行业专利文本样本、专利写作规范与结构模板及错误案例库、撰写错误与审核要点数据。数据集经过脱敏和结构化处理，支持专利生成模型训练、关键词检索增强及企业知识产权管理体系建设。