"How many mathematicians in China have been dedicated to the research of basic algorithms for artificial intelligence?" This question proposed by several academicians including Xu Kuangdi of the Chinese Academy of Engineering at an academician salon held in Shanghai two years ago gave the industry insiders some food for thought.

Data, computing power and algorithms are hailed as the three cornerstones of artificial intelligence. The iteration and innovation of algorithms, which drives the evolution of artificial intelligence technology, will be the key to the breakthrough and development of artificial intelligence in the future. Without core algorithms, one will be at the mercy of others in case of critical problems. Currently, China's artificial intelligence innovation ability has been enhanced continuously, but the industrial development still depends heavily on open source code and existing mathematical models.

The artificial intelligence industry is underpinned by algorithms. In fact, artificial intelligence without algorithms is like a body without a soul. Shanghai aspires to build a highland of artificial intelligence, and it will need to make major breakthroughs in the field of algorithms to lay a solid foundation for a world-class artificial intelligence industry cluster. 

Just recently, Shanghai has specially formulated and released the Action Plan for New-generation Artificial Intelligence Algorithm Innovation (2021-2023) to promote the development of algorithms, aiming to create an active innovation ecosystem for theoretical research and development of algorithms and industrial application by making full use of Shanghai's advantages of computing power and data resources.

According to the Action Plan, Shanghai aims to maintain a leading position in the field of artificial intelligence algorithms in China, and rank top in some areas by 2023. It will initially build itself into an international artificial intelligence algorithm highland with fruitful cutting-edge results, active application ecology and numerous innovation talent, provide substantial support for the quality development of Shanghai’s artificial intelligence industry, and empower the city's overall digital transformation.

Specifically, it will produce about 10 leading results in such fields as basic theories, core algorithms, general models and software frameworks; provide 100 algorithm products and 100 algorithm application cases that serve as models and can be replicated in key areas of urban digital transformation; bring in and cultivate about 50 international leading algorithm scholars and excellent young scientists, and select and train about 100 chief algorithmists from enterprises of relevant industries; and cultivate a group of algorithm application developers of 100,000 people.

Committed to building Shanghai into a highland of algorithms, the Action Plan proposes five major moves including making breakthroughs in basic algorithms, innovating algorithm application, cultivating algorithm ecology, attracting algorithm talent and opening up algorithm communities.

In order to make breakthroughs in basic algorithms, it's necessary to build an algorithm innovation platform system, support basic algorithm theory research, and promote breakthroughs in core algorithm technologies. Currently, UniDT Shanghai Artificial Intelligence Innovation Center, a subsidiary of UniDT (Shanghai) Co., Ltd., has become an important artificial intelligence innovation R&D platform in Shanghai. With an eye on the frontier of global science, it focuses on the two research fields of basic computing theory and cognitive intelligence (such as Conscious Turing Machine) to promote the development of basic theories. Xuan Xiaohua, founder and chairman of UniDT, said that UniDT will step up efforts to develop the cognitive intelligence technology, provide the underlying intelligence engine, lead independent innovation of algorithms, offer intelligent products and services for different industries, promote the transformation and upgrading of intelligent development of industries, and transform research results into actual productivity.

Another example is the Shanghai Institute of Artificial Intelligence Algorithms of Fudan University, which was established at the first World Artificial Intelligence Conference. The institute is making advance deployments in basic research that may lead to a revolution of the artificial intelligence paradigm, researching and developing world-leading artificial intelligence mathematical theories and algorithm foundations, and cultivating strategic science and technology power for artificial intelligence to provide a strong scientific reserve for the sustainable development and in-depth application of artificial intelligence.

In terms of innovating algorithm application, the Action Plan proposes to promote breakthroughs and upgrade of key algorithms such as autonomous driving, AI+industry, AI+health care, AI+finance and AI+governance in the three major digital transformation areas of economy, life and governance, and form a batch of algorithm products that serve as models and can be replicated. At the same time, it encourages various enterprises to open up application scenarios, promote the docking of algorithms and enterprise business, and launch a batch of demonstration projects for algorithm conversion.

In addition, it's necessary to establish a number of operation platforms for algorithm intellectual property through cultivating algorithm ecology, launch pilot programs for the operation and trading of artificial intelligence algorithms, encourage the development of new models such as "algorithm application stores," and gradually improve algorithm intellectual property trading rules. By implementing algorithm talent attraction programs, efforts will be made to cultivate and select a batch of chief algorithmists who are proficient in artificial intelligence algorithms and business operations and take the lead in promoting algorithm application, and to assist industry enterprises in building core algorithm empowerment teams with chief algorithmists at the core. It's crucial to build brand open source communities by opening up algorithm communities, focus on building community ecology such as the Baiyulan Open AI, connect with well-known open source communities at home and abroad, and form a key node of the international artificial intelligence development ecology network.