AI赋能科学研究，北京发布全国首个科学智能专项地方政策

Group 1 - The core viewpoint of the news is the official release of the "Beijing Action Plan for Accelerating AI Empowerment in Scientific Research for High-Quality Development (2025-2027)", marking a strategic roadmap for AI for Science in Beijing and the first local policy in China focused on this area [1] - AI for Science is recognized globally as a new paradigm to accelerate scientific research, with Beijing positioning itself as a leader in this field by establishing the Beijing Academy of Scientific Intelligence in 2021 [2][6] - The Action Plan aims to integrate AI with scientific research, focusing on breakthroughs in fundamental theories and interdisciplinary collaboration, with goals to build at least 10 high-quality scientific databases and serve over 10 million users by 2027 [3] Group 2 - The Beijing Municipal Development and Reform Commission plans to enhance the overall planning and systematic layout of scientific intelligence, targeting key disciplines such as materials and life sciences to develop specialized models [4] - The Beijing Economic and Information Technology Bureau emphasizes the importance of AI in accelerating scientific innovation across various sectors, including biomedicine and new materials, while acknowledging existing challenges in the industrial application of scientific intelligence [5] - The Haidian District aims to strengthen the foundation for scientific intelligence development by supporting advanced research elements and creating a comprehensive support system for scientific innovation [6] Group 3 - Significant achievements in the field of AI for Science include the "Bohler Research Space Station," which integrates literature review, computation, and experimentation, currently utilized by over 900,000 users across more than 40 universities and companies [7] - The "DPA Large Atom Model," developed by over 30 organizations, aims to enhance the understanding of atomic interactions, achieving world-leading stability and predictive performance while reducing data computation costs by 90% [8] - The "Uni-Lab-OS" intelligent laboratory operating system transforms traditional labs into autonomous "AI scientists," improving efficiency and data sharing among instruments [9] - The "DeepFlame" software enables comprehensive numerical simulations of rocket engines, significantly reducing development costs and time by allowing for virtual testing of extreme conditions [10]