Anchoring the creation of artificial intelligence "Shanghai Highland", Shanghai is accelerating the construction of a globally competitive large-scale model industrial ecology. On December 16, the Digital Intelligence Shanghai 2023 Summit "Digital Intelligence Talk" Generative Artificial Intelligence Innovation Application Empowering Thousands of Industries Forum was successfully held. Experts and scholars in the field of large models, as well as representatives of leading enterprises, gathered together to share the latest technological progress of generative AI and cutting-edge empowerment practices in the fields of finance, industry, healthcare, and education, and discuss the latest development trends of large models.
This forum is guided by the Office of the Shanghai Municipal Leading Group for Urban Digital Transformation, the Shanghai Municipal Commission of Economy and Information Technology, and the People's Government of Yangpu District, Shanghai, hosted by the Science and Technology Commission of Yangpu District, and the Shanghai Urban Digital Transformation Application Promotion Center.
The person in charge of the Artificial Intelligence Development Department of the Municipal Economic Information Commission attended the forum and delivered a speech.
Zhang Qi, professor and doctoral supervisor of the School of Computer Science and Technology of Fudan University and director of the China Chinese Information Society, brought the report "AI Empowers Industries, Large Models Help the Digital and Intelligent Upgrading of Industries", on the basis of combing the existing language model technology route, he summarized the new research directions and methods, and introduced how large language models can help the digital and intelligent upgrading of the industry.
Zhang Shaoting, Vice President of SenseTime, shared the theme of "Building a New Paradigm for Future Medical Scenarios Based on AI Large Models". He pointed out that the medical multimodal basic model based on cross-modal, cross-center, and cross-disease large-scale data can break through the bottlenecks of existing medical artificial intelligence, such as high annotation cost, weak knowledge transferability, and difficulty in covering long-tail diseases. In specific applications, the medical multimodal basic model has empowered a series of medical scenarios such as medical image analysis, clinical assisted decision-making, and patient medical services, significantly improving the clinical diagnosis and treatment efficiency of doctors, optimizing the medical experience of patients, and creating a new blueprint for the future of healthcare.
Sheng Zhichao, vice president of the iFLYTEK AI Research Institute, brought the report "Practical Exploration of Xinghuo Cognitive Model Empowering Education", focusing on the innovative application of iFLYTEK Xinghuo cognitive model in multiple scenarios such as essay correction, oral dialogue, teacher preparation, interactive supplementary learning, mental health companionship, etc., the basic requirements to be followed in the application of large model technology in the education industry, and the broad prospects for the digital intelligence upgrade of the education industry empowered by large models.
Productization is the only way for large models to land. Guo Jiangliang, Vice President of Innovation Qizhi Technology, brought the report "Large Model Helps New Industrialization: Reconstructing Industrial Productivity and Production Relations", sharing the AIGC application products built for industrial robots, enterprise private domain data analysis, and enterprise private domain knowledge base, and interpreting its solutions for factory logistics, intelligent BI, intelligent manufacturing training and other fields to empower the industry.
Large model technology is changing the financial industry faster and deeper than expected. Chen Hong, head of Ant Financial Large Model Algorithm of Ant Financial Insurance Technology Department, brought the report "What kind of large model does the financial industry need", introducing that Ant Group is building reliable financial AI and launching a 1+2 product matrix (Ant Financial Large Model + Zhi Xiaobao 2.).0 and Zhi Xiaosuke series), sharing an understanding of the boundaries of large models and crossing boundaries.
The roundtable session was presided over by Liu Jiangxian, Chief Strategy Officer of Daguan Data, Zhang Qi, professor and doctoral supervisor of the School of Computer Science and Technology of Fudan University, director of the China Chinese Information Society, Du Xiawei, general manager of Sugon Intelligent Computing Product Department, Sun Linjun, founder and CEO of Real Intelligence, Zhou Jian, founder and CEO of Lanma Technology, and Liu Jingguo, partner of Shanghai Artificial Intelligence Industry ** Investment Research, respectively from their own perspectives, focusing on "AI agent." Dealing with the potential and technological breakthroughs in complex language understanding and generation tasks", "Opportunities and strategies for enterprises on the scenario side and technology side to promote the implementation of large models and AI agents", "Generative AI industry investment strategies" and other topics, shared the key experience of the integration of various AI technologies such as agents to achieve human-machine collaboration, and the technological breakthroughs, innovative applications and strategic layout of AI agents.
With the theme of "Intelligent Innovation Empowering Industrial Digital Transformation", the Digital Intelligence Shanghai 2023 Summit will bring 1 main forum, 5 parallel forums of "Digital Intelligence" and a series of blockbuster releases, where more than 100 government, industry, academia, research and application experts will gather together to jointly develop the deep integration and innovation of digital intelligence technology and industry. The summit was broadcast live through multiple platforms such as Heart of the Machine, Cailianshe APP, Bilibili, Migu, Himalaya, Douyin, etc., and more than 2 million people live-streamed online. More than 40 partners, including Xinhua News Agency, People's Daily, China Business Network, and Jiefang, attended the conference.