Innovation is revitalized and the integration of artificial intelligence and the real economy is accelerating

◎Reporter Wang Chun

A new generation of artificial intelligence (AI) is booming around the world, injecting new momentum into economic and social development. At the 2022 World Artificial Intelligence Conference held from September 1st to 3rd, reporters learned that artificial intelligence and the real economy focus on intelligent empowerment and accelerated integration, and are becoming a new growth point for economic development.

The proportion of Chinese enterprises adopting artificial intelligence ranks first in the world

The World Artificial Intelligence Conference is a concentrated show for high-level applications of AI technology. It showcases not only application innovations such as lighthouse factories and unmanned ships, but also infrastructure such as computing power and chips; there are both service-type integration such as AI chess playing and surgical robots, as well as cutting-edge technologies such as privacy computing and multi-modal large models. explore. Cai Yinghua, vice president of Alibaba Group and president of global sales of Alibaba Cloud, said at the plenary meeting on industrial development that the current adoption rate of artificial intelligence by Chinese enterprises is 58%, ranking first in the world. "It is expected that by 2025, China will become the world's largest data circle, and intelligent computing will also become the main driving force for the growth of computing power."

Zhang Wen, founder and chairman of Biren Technology, believes that in the future, Computing power will present five unprecedented new characteristics: "manufacturing intelligence", "feedback and superposition", "simulation and simulation", "virtual world" and "economic leverage", and will move towards "efficient computing power", "green computing power" and "diversified computing power". "Computing power" is moving forward in three new directions.

In recent years, large models have demonstrated unprecedented understanding and creative capabilities, bringing humans one step closer to the goal of general artificial intelligence. However, there are still many technical difficulties in realizing the versatility of large models in all modes and tasks. At the same time, due to limitations of computing resources, their training and implementation are quite challenging. At the 2022 World Artificial Intelligence Conference "Large-scale Pre-training Model" theme forum, Zhou Jingren, deputy director of DAMO Academy, released Alibaba's latest "Tongyi" large model series and announced that the relevant core models will be open source to developers around the world. Facing the problem of lack of versatility and ease of use of large models, the unified AI base of "Tongyi" builds a hierarchical artificial intelligence system for collaboration between large and small models, which will provide advanced information for AI to move from perceptual intelligence to knowledge-driven cognitive intelligence. infrastructure.

“Large models imitate the process of human beings building cognition, which is an important opportunity we are facing now. By integrating AI’s knowledge systems in different modalities and fields such as language, speech, and vision, we expect multi-modal The state-of-the-art model can become the cornerstone of the next generation of artificial intelligence algorithms, allowing AI to go from being able to use only a 'single sense' to having all five senses, and using a brain with a wealth of knowledge to understand the world and think, and ultimately achieve levels close to human levels. Cognitive intelligence." Zhou Jingren said. According to reports, in order to accelerate the large-scale application of large models, DAMO Academy has also developed the S4 framework, a key technology for the implementation of ultra-large models. The multi-task accuracy of large models with tens of billions of parameters can be close to lossless when the compression rate reaches 99%.

Many companies are building an industrial metaverse

The integration of the metaverse and AI is becoming a new path to empower the real economy. Cao Peng, senior vice president of JD.com, believes that only by integrating artificial intelligence into the entire process of the industrial chain in a large-scale and systematic way and creating industrial AI can we achieve more leading industrial competitiveness. In terms of retail scenarios, more than 85% of JD.com’s purchases are automatically completed by machines. It uses operations optimization, deep learning and other technologies to apply AI capabilities to various supply chain scenarios, helping JD.com’s inventory turnover reduce to 31.2 days.

A series of high-level applications are setting off a wave of intelligence, promoting high-quality development and exerting the "goose effect" that empowers all industries. Many companies are also looking for more new growth points by building an industrial metaverse. Ge Yue, vice president of Apple and managing director of Greater China, said at the industry development plenary meeting that machine learning plays a vital role in people's innovation and creation, and it can improve people's lives in all aspects, especially in the In the field of accessibility and health functions, “Our machine learning models not only rely on powerful chips, but also require high-quality input, including touch, motion, sound and vision information. These sensors will be integrated into our devices. , Provide fast and highly accurate signals for our machine learning models.”

Technology companies must be cautious about their own technical capabilities

Technology companies in the digital economy era also need to be cautious about their own technical capabilities and influence, and always be alert to the boundaries of technological innovation. "As we move towards the era of intelligence and digitalization, the challenges of scientific and technological ethics will follow and accompany each other." Zhou Hanmin, vice chairman of the Shanghai Municipal Committee of the Chinese People's Political Consultative Conference, said at the meeting that new needs call for new governance and the establishment of a global artificial intelligence governance system has become a common demand.

At the World Artificial Intelligence Conference, Alibaba Group announced the establishment of a Technology Ethics Governance Committee and introduced seven external advisory committee members to strengthen tripartite supervision. According to reports, the Alibaba Group Technology Ethics Governance Committee is led by Cheng Li, chief technology officer of Alibaba Group, and its core members come from Alibaba Research Institute, Damo Academy, Legal Compliance, Alibaba Artificial Intelligence Governance and Sustainable Development Research Center and related The business section has multiple specific working groups. "The Science and Technology Ethics Governance Committee should be the gatekeeper of technological innovation, be people-oriented, guided by social values, and use ethics and morality as the boundary to build social responsibility for science and technology enterprises," Cheng Li said.

The above is the detailed content of Innovation is revitalized and the integration of artificial intelligence and the real economy is accelerating. For more information, please follow other related articles on the PHP Chinese website!