光明科学城论坛·大湾区智能算力与大模型智能体论坛在深举办

Core Insights - The forum focused on the integration of intelligent computing power and large model intelligent agents, emphasizing their role in driving industrial empowerment and innovation in various sectors such as meteorology, finance, and healthcare [1][10]. Group 1: Forum Overview - The "2025 Guangming Science City Forum" was held in Shenzhen, co-hosted by various governmental and educational institutions, highlighting the importance of intelligent computing and large models [1]. - Experts and scholars from fields like artificial intelligence and high-performance computing gathered to discuss cutting-edge topics related to intelligent computing infrastructure and large model technology innovation [1][3]. Group 2: Key Statements - Gao Wen, director of Pengcheng Laboratory, emphasized the strategic importance of artificial intelligence and computing power for national sovereignty in the digital age, highlighting the ongoing construction of the "Pengcheng Cloud Brain III" [3]. - Wang Fangcheng, deputy district mayor of Guangming, stated the district's commitment to becoming a leading hub for artificial intelligence, with over 300 high-level talents and nearly 100 quality AI enterprises, contributing to an industry scale exceeding 30 billion [4]. Group 3: Technological Innovations - The forum showcased significant technological advancements, including the release of the "Pengcheng Brain Sea 2.1" open-source model and the AI forecaster assistant "Afu," which integrates advanced computing with meteorological applications [6][8]. - Other innovations included the FenixCOS inference engine and financial intelligent agents developed in collaboration with major financial institutions, demonstrating the practical applications of domestic computing power [6]. Group 4: Collaborative Agreements - Several strategic cooperation agreements were signed, including partnerships between Pengcheng Laboratory and the Shenzhen Meteorological Bureau to enhance AI applications in meteorology [8]. - Agreements were also made with other laboratories to advance research and development in domestic computing power and large model tools, aiming to foster innovation in intelligent agent technologies [8]. Group 5: Academic Contributions - Notable presentations included discussions on advancements in data selection and training methods for large models, as well as the application of multi-modal models in precision medicine [9]. - The forum served as a platform for academic and industry leaders to explore innovative paths for integrating artificial intelligence into the real economy [10].