Fujian Province has set up a special science and technology project for the development of artificial intelligence industry for the first time
Recently, the Fujian Provincial Department of Science and Technology issued a notice to organize the application for the 2025 artificial intelligence industry development science and technology project, and launched the scientific research task of supporting the science and technology project. This is also the first time that Fujian Province has established a science and technology project for the development of artificial intelligence industry. According to the application guide for this batch of projects, two major directions are supported. The first is the research and development of key technologies for lightweight deployment of large models for multimodal application scenarios and demonstration applications. Relying on universities, research institutes and enterprises in the province to carry out research and development, two projects will be supported on a preferential basis, with a funding of 3 million yuan for each project and a project implementation period of 1 year and 6 months. The research and development content includes dynamic workflow management and automated orchestration technology for precise application of intelligent tasks, and the realization of five aspects such as model fine-tuning, data governance and efficient execution of multimodal human-machine collaboration. The second is the research and development and demonstration application of key technologies for network data security based on security-enhanced artificial intelligence models. Relying on universities, research institutes and enterprises in the province to carry out research and development, one project will be supported on a preferential basis with a funding of 2 million yuan and a project implementation period of 1 year and 6 months. The research and development content includes six aspects: vulnerability analysis technology for intelligent algorithm models, building a model security assessment framework in multimodal scenarios, and supporting automated vulnerability warnings for intelligent algorithm models.