专访北京移动刘南：“5G+工业互联网”还需关注个性化需求

Core Viewpoint - The integration of 5G and industrial internet is crucial for expanding the large-scale application of 5G, facing challenges such as high costs and fragmented demand across various industries [3][4]. Group 1: 5G and Industrial Internet Integration - The deep integration of 5G and industrial internet is essential for expanding 5G applications, currently facing challenges in cost and demand matching [3][4]. - High costs associated with 5G network construction and equipment upgrades pose a significant burden, especially for small and medium-sized enterprises [3]. - The current "5G + industrial internet" applications primarily address common needs, lacking sufficient alignment with the unique demands of different industries [3]. Group 2: 5G-A Development and Its Implications - 5G-A is a key transitional phase towards 6G, enhancing network bandwidth and providing valuable insights for future 6G development [5]. - The development of 5G-A has led to the emergence of new application scenarios, indicating that 6G should focus on deeper integration with vertical industries [5]. - The emphasis on industry collaboration in 5G-A development highlights the need for a robust ecosystem to support 6G advancements [5]. Group 3: AI for Industry Trends - AI for Industry is expected to experience rapid growth, integrating with 5G-A and industrial internet to create a comprehensive intelligent system [9][10]. - The expansion of AI applications is moving from a few leading sectors to broader industries, enhancing production efficiency and reducing costs [9][10]. - Key issues for the information and communication industry include data quality and security, computational power support, standardization, and talent cultivation [10][11]. Group 4: Challenges in AI Model Development - The transition from general AI models to specialized industry applications faces challenges such as data barriers, algorithm precision, and scene adaptation [10][11]. - Data governance is prioritized to address the scarcity of high-quality datasets, utilizing proprietary tools for data cleaning and transformation [10][11]. - Collaboration with industry partners is essential for developing benchmark applications and optimizing models through practical scenarios [12].