SenseTime, a leading global artificial intelligence (AI) software company, today hosted an industry forum themed “AI+: Meta-Innovation” at the 2022 World Artificial Intelligence Conference (WAIC), where academics and industry leaders discuss how AI technology can enable and bring promising innovations to various industrial sectors.
The doctor. Xu Li, Executive Chairman and CEO of SenseTime, delivered a speech titled AI’s Yingzao Fashi: Innovative Industrial Handbook and announced four new AI platforms at the forum. The four innovative AI platforms spanning multiple fields including AI cloud computing, vehicle collaboration, AI in games, and smart hospitals aim to drive scalable industrial adoption of AI through standardized platform services, more efficiently, easily to use, and at low cost.
Several senior government officials attended the forum, along with key guest speakers who shared their views on how AI is changing the game in various industries, including Gong Ke, Executive Director of the China Institute for Artificial Intelligence Development Strategies for Next Generation, Ning Guang, Chinese Academy of Engineering Sciences, President of Ruijin Hospital associated with Shanghai Jiaotong University School of Medicine, Zhang Yongwei, Vice President and General Secretary, China EV100, Li Qian, VP of Vehicle Intelligence at HiPhi and Qiu Zhijie, Vice President and Dean of the School of Experimental Art, Central Academy of Fine Arts.
Inspired by Yingzao Fashi (Treatise on Architectural Methods or State Building Standards), the oldest surviving Chinese engineering manual on architecture originally published during the Song Dynasty, Xu Li shared his insights into creating highly scalable AI innovations for a wide range of applications applications based on a standardized industrial manual. He pointed out that many famous Chinese architectures, including the Old Drunken Pavilion, Foguang Temple, and Prince Teng Pavilion, were built in imitation of Yingzao Fashi, who specified the ideal of universal standardization.
“These norms and consensus form gradually with the continuous accumulation of people’s experience through abstraction, model deduction, and simulation.” Xu Li said in his speech. As AI moves towards commercialization, AI models are becoming standard features for the intelligent age, supporting various applications such as Advanced Driver Assistance System (ADAS), city management, surgery planning, and agents. in games, where problems can be solved by the standardized AI models themselves. To date, SenseTime has produced over 49,000 AI business models, which can be merged into various AI solutions for different industries.
A standardized manufacturing process is also needed to facilitate industry standardization. Under this concept, SenseTime achieved common AI capabilities by standardizing the production of AI models, including AI infrastructure, data labels, pre-trained master models, and industrial applications. SenseTime’s automated data production line increases the speed of automated data labeling by 600 times.
“Industries cannot be formed without standards,” said Xu Li, standardization will lead to the flourishing industrialization of AI. Through a powerful infrastructure, we can produce high-quality standardized models with replication in different scenarios and efficiently manage the entire model manufacturing process. It is worth remembering that standardization is not a limitation and will eventually create infinite possibilities in AI applications.