DeepInnovationAI
收藏资源简介:
DeepInnovationAI是一个全球性的综合数据集,旨在连接学术研究和工业专利之间的空白。该数据集由上海人工智能实验室创建,包含3511930篇学术论文和2356204项专利记录,覆盖了60多年的数据,跨越了全球主要的人工智能创新国家和地区。数据集通过深度学习技术,如关键字匹配、大型语言模型和双层BERT分类器,精确识别AI相关内容,并使用超图分析方法来评估创新的新颖性。DeepInnovationAI支持对技术创新模式和国际化竞争动态的详细分析,为建模AI创新动态和技术转移过程提供了坚实的基础。
DeepInnovationAI is a global comprehensive dataset designed to bridge the gap between academic research and industrial patents. Developed by the Shanghai AI Laboratory, this dataset contains 3,511,930 academic papers and 2,356,204 patent records, spanning over 60 years of data and covering major global countries and regions focused on AI innovation. The dataset accurately identifies AI-related content using deep learning techniques such as keyword matching, large language models (LLMs), and two-layer BERT classifiers, and employs hypergraph analysis methods to assess the novelty of innovations. DeepInnovationAI supports detailed analyses of technological innovation patterns and international competitive dynamics, providing a solid foundation for modeling AI innovation dynamics and technology transfer processes.

- 1DeepInnovation AI: A Global Dataset Mapping the AI innovation and technology Transfer from Academic Research to Industrial Patents上海人工智能实验室 · 2025年



