Wang Haifeng, Baidu‘s Chief Technology Officer and Director of the National Engineering Laboratory of Deep Learning Technology and Applications, announced the latest progress of PaddlePaddle, China’s first deep learning platform, at the WAVE SUMMIT+2021 held on Sunday. To date, the platform has gathered 4.06 million developers, created 476,000 models, and served 157,000 enterprises and institutions, and has become the largest player in the field in the country.
The summit, hosted by the National Engineering Laboratory of Deep Learning Technology and Applications, was held in Shanghai on Sunday. Wang Haifeng said at the summit that PaddlePaddle is a typical AI platform, which empowers developers to support the mass production of AI technology on an industrial scale.
At the first summit in April 2019, Wang Haifeng proposed that artificial intelligence that is supported by deep learning had entered the stage of industrial mass production. In that stage, an AI mass production platform is required to deploy the technology to all walks of life in a standardized, automated and modular way. In addition, the platform would need to unite all parties in the development process in order to minimize conflicts within the code and workflow.
In the past three years, as technology and the industry developed, PaddlePaddle has continued to provide a convenient development platform, efficient training and multi-terminal deployment of AI models. As an open source platform, PaddlePaddle has brought developers and all sectors of society together to strengthen its ecology as well as trained professionals to foster the mutual promotion of technology, industry, specialization and ecology.
As Wang Haifeng believes, based on his observation of AI technology and application development as well as his experience at PaddlePaddle, AI technology and the industry as a whole are much more integrated and innovative today than ever before. However, despite the complexity, he also believes that the barrier to access for developing and applying AI technology is being constantly lowered.
At the summit, PaddlePaddle also released ten of its latest technological advances, including the industry’s first industrial practice example base. Other releases included PaddlePaddle’s own “Great Navigation” 2.0 Co-creation Plan; a program to accelerate text-based tasks; a multi-level low-cost hardware adaptation scheme; an industrial-level open source model library; and, finally, a desktop version of PaddlePaddle EasyDL, which is able to complete modeling on a local machine.
Wu Tian, Vice President of Baidu Group and Deputy Director of the National Engineering Laboratory of Deep Learning Technology and Applications, believes that “there are three key points that need to be grasped in the construction of PaddlePaddle: innovative technology, the needs of developers, and sharing.”
At present, the AI ecology is becoming much more prosperous. In the past year, the scale and breadth of AI active developers in cities across the country has increased significantly, covering a variety of fields such as industrial application, transportation, energy, urban development, and many others.
PaddlePaddle’s own Artificial Intelligence Industry Empowerment Centers have already been set up in Beijing, Shanghai, Jiaxing, Fuzhou, and others. More centers in 15 other cities, including Macao Special Administrative Region and Guangzhou, are being planned and are aimed at boosting industrial empowerment, personnel training and ecological co-construction.