How can the Shanghai artificial intelligence industry, which has a total scale of more than 300 billion yuan in Zhangjiang Artificial Intelligence Island, more efficiently empower all walks of life? On Zhangjiang Artificial Intelligence Island in Pudong New District, Baidu Paddle

Zhangjiang Artificial Intelligence Island Overview

Shanghai artificial intelligence industry with a scale of over 300 billion yuan, how to empower all walks of life more efficiently? At the Zhangjiang Artificial Intelligence Island in Pudong New Area, Baidu Fei Paddle (hereinafter referred to as " Fei Paddle ") presented a "construction map": source creation, technology open source, application transformation, industry empowerment, technology venture capital . In the exhibition hall, the crystal gray columns are arranged like a building frame, giving a visual annotation to the "drawings" - "Feipiao open source deep learning platform is like a building framework, enabling thousands of industries to realize application transformation on this basis, and then create an industrial ecology." said the person in charge of Feipiao Artificial Intelligence Industry Empowerment Center.

A few steps away, in the exhibition hall of the Shanghai Autonomous Intelligent Unmanned Systems Science Center of Tongji University (hereinafter referred to as the "Unmanned Systems Science Center"), this "construction drawing" has an interlocking path: disciplines, major scientific issues, research directions, science center goals, key core technologies, and industrial transformation. The person in charge "sets the focus": obtain original innovative research results, develop key core technologies, work hard on the formulation of standards and specifications, and promote industrial change.

ecology and standards - in the recent intensive inspection of the "artificial intelligence industry development" by Municipal People's Congress deputies, these two words were mentioned repeatedly, reflecting Shanghai's persistent approach to refining the "construction drawings" of the artificial intelligence industry: thinking outside the project and finding the coupling between projects and industry elements.

technology is open source to expand the ecosystem.

Fei Paddle is looking for a coupling path by creating an open source framework platform to link small and medium-sized enterprises to create an industrial ecology.

In the industry, there is an image metaphor for companies empowered by open source frameworks: it is like a pre-made dish. The main ingredients and condiments are already prepared, and you only need to take it back and put it in the pot. For example, the person in charge of Feipiao said that an old software company in Zhangjiang with electronic customs as its main business wanted to create artificial intelligence application scenarios for complex document processing processes. "The role of Feipiao is to analyze the customs document layout and form more accurate rule results to achieve the purpose of automated document processing. The software company conducts secondary development based on Feipiao's open source algorithm model and optimizes it into App or system software."

But empowerment does not mean targeted cultivation, and there is no "restriction clause that only allows one company to use it alone in a specific scenario." The person in charge said that in more cases, Feipiao uses open source technology as industry-wide artificial intelligence (AI), allowing companies to make innovative applications based on their own businesses. On the other hand, Feipiao provides learning resources and guidance, allowing enterprises to go through a complete scenario development chain and truly in-depth learning. "The current algorithm model developed based on general AI belongs to the enterprise, and the assets can be applied for software copyright , etc."

These practices, which may seem to have little benefit in the short term, are actually paving the way for the creation of an industrial ecosystem. Looking from a distance, the picture is broader: partners empowered by Flying Paddle have the opportunity to leverage their comparative advantages and undertake Baidu business opportunities to jointly promote industrial empowerment and iteration to achieve a win-win situation.

From another perspective, even if the ultimate goal is to create an industrial-level general-purpose AI, it cannot be achieved overnight. The person in charge said, "Small and medium-sized enterprises are the capillaries of various industries. Only through co-creation with industrial innovative enterprises can we find the coupling with the elements of the industry."

has created an open source framework platform. In the global situation where a hundred schools of thought are contending, it has become Feipiao's top priority to make the ecosystem bigger and stronger. As of the end of November, Feipiao has gathered 5.35 million developers, assisted 670,000 creative models, and served 200,000 enterprises and institutions.

Dai Liu, chairman of the Municipal People's Congress Financial and Economic Committee, said, "Fei Paddle has developed rapidly in the short term, but user power levels can be further improved. The point behind this is whether the open source architecture can meet higher-level needs? Currently, the world's top open source framework has good data on the number and level of its user power levels. Fei Paddle should continue to strengthen technology upgrades. level iteration, application services, and making a fuss about improving customer capabilities."

Baidu Flying Paddle's outdoor scene

sets standards and is the source of

Lin Paddle's unmanned system science center. Breaking through the key and formulating standards are its starting points to open up the elements.

In October this year, the first phase of the unmanned system multi-body collaborative major scientific and technological infrastructure construction project led by Tongji University was launched on Zhangjiang Artificial Intelligence Island. After completion, the project will become an important experimental device for breakthroughs in original artificial intelligence theory and verification of key technologies.

"This device is directly related to the formulation of standards and specifications for artificial intelligence and unmanned systems, and is an important manifestation of participating in international competition in the future." The head of the Unmanned Systems Science Center said that for example, in the future, data security, route safety calculation and other issues of multiple unmanned systems such as drone in the low-altitude field, scientific experiments are required.

has a strong orientation in artificial intelligence application, and enterprises directly connect to the market. Although it is convenient to reverse research and development from the demand side, scientific research institutions have a longer-term vision from the perspective of public research. "For example, standards are the right to speak in some industries." The person in charge said that one of the focus points of scientific research institutions is to achieve high-level scientific and technological self-reliance in basic original theories and key areas, and promote the innovative development of artificial intelligence-related industries.

looks at the city. At present, high-end platforms such as Shanghai Artificial Intelligence Laboratory, Unmanned Systems Science Center, Qizhi Research Institute, and Zhejiang University Shanghai High-tech Research Institute have started operations to explore the "unmanned zone" in the cutting-edge artificial intelligence technology; 5 new generation artificial intelligence open innovation platforms, 17 municipal artificial intelligence innovation centers, etc. are laid out in Shanghai to enhance the supply capacity of industry innovation . The

standard is also a circulation factor between scientific research, industry and enterprises.

Dai Liu pointed out that some leading enterprises are guiding industry standards in the process of moving forward. "In some fields, local standards are still gradually being established. Unmanned Systems Science Center can consider making good efforts in industry standards and group standards. For the development of artificial intelligence enterprises, I think more important is how the industry viewes it and how the group recognizes it. There is still a lot of room for industry and group standards."

Manuscript Source: Wenhui APP