(The headline of this article is strictly prohibited from being moved, and the whole network has been opened for rights protection)
If you like it, you can click on the word "follow" above. Pay attention to the development of Chinese science and technology together!
When I woke up, I saw four good news in the field of domestic science and technology artificial intelligence, and I was refreshed to see it, and I quickly wanted to share my joy with everyone!
First, Nature magazine published Huawei's meteorological model research, and the accuracy of AI forecasting exceeds that of traditional numerical methods
According to global network reports, Nature, a well-known international academic journal, recently published the research results of the research and development team of HUAWEI CLOUD Pangu Meteorological Model. In the paper "3D Neural Network for Accurate Medium-term Global Weather Forecasting", the team proposed a weather prediction model based on 3D neural network, which surpasses the traditional numerical prediction method in accuracy and speed.
HUAWEI CLOUD's Pangea Meteorological Model is the world's first model that uses AI technology to achieve accurate medium-term weather forecasting, which is more than 10,000 times faster than traditional numerical forecasting. The model can provide second-level accuracy for global weather forecasting, covering multiple meteorological research sub-scenarios such as potential, humidity, wind speed, temperature, and sea level pressure. In addition, institutions such as the European Center for Medium-Range Forecasting and the Central Meteorological Observatory have verified the superiority of the Pangea Meteorological Large Model in practical applications.
Aiming at the insufficient accuracy of the existing AI weather prediction model, the Pangea Meteorological Model R&D team proposed a 3D Earth-Specific Transformer adapted to the Earth coordinate system to process complex uneven 3D meteorological data. At the same time, the team adopted a hierarchical time-domain aggregation strategy to reduce the number of forecast iterations and reduce iteration errors. After in-depth training on 43 years of global weather data, the Pangea Meteorological Model surpasses traditional numerical prediction methods in accuracy and speed.
In the future, HUAWEI CLOUD will cooperate with global meteorological agencies to continue to explore the application potential of AI in the meteorological field, and provide support for various industries, such as agriculture, forestry, animal husbandry and fishery, and aviation.
My opinion: The publication of the research results of HUAWEI CLOUD Pangu Meteorological Model in Nature marks an important breakthrough in Continental's AI technology in the meteorological field. The application of this model can not only improve the accuracy and speed of weather forecasting, but also provide more accurate meteorological data support for various industries such as agriculture, forestry, animal husbandry and fishery, aviation and navigation, etc., which will promote the development of the meteorological field in the future and make people re-examine the possibility of weather forecasting models.
Second: the scale of the mainland's AI industry exceeded 500 billion yuan: the second largest computing power in the world
According to Xinhua News Agency, the 2023 World Artificial Intelligence Conference was held in Shanghai from July 6 to 8. Xu Xiaolan, vice minister of the Ministry of Industry and Information Technology, said at the opening ceremony of the conference that the mainland's artificial intelligence industry is booming, the scale of core industries has reached 500 billion yuan, and the number of enterprises has exceeded 4,300. The conference attracted more than 1,400 domestic and foreign scholars, entrepreneurs, representatives of international organizations and other guests to participate, and displayed innovative achievements in the field of artificial intelligence at home and abroad.
At the conference, Xu Xiaolan also revealed that the mainland artificial intelligence infrastructure has accelerated its layout, and the scale of computing power ranks second in the world. Major projects such as east-west computing have been accelerated, and the number of 5G base stations has exceeded 2.8 million. At the same time, more than 2,500 digital workshops and smart factories have been built, shortening the R&D cycle by about 20.7%, increasing production efficiency by about 34.8%, reducing the defective product rate by about 27.4%, and reducing carbon emissions by about 21.2%.
There were more than 400 exhibitors in this conference, and the exhibition area reached the largest ever. The 50,000-square-meter World Expo exhibition covers four major sections: core technology, intelligent terminals, application empowerment and cutting-edge technology, displaying the latest technologies and products in the fields of large models, chips, robots, and intelligent driving.
My opinion: The rapid development of the artificial intelligence industry in the mainland has injected new vitality into the real economy and promoted industrial upgrading. The efforts of government departments in laying out computing power, integrating data, digging deep into scenarios, cultivating enterprises, and ensuring security have created a good development environment for the artificial intelligence industry. However, we should also see the challenges and remain vigilant, continuously strengthen technological innovation capabilities, and improve industrial competitiveness to cope with future changes and challenges.
Third: JD.com reshapes the business model with a 100-billion-level Yanxi big model: 3 billion words of marketing copywriting achievements
According to global network news, at the 2023 World Artificial Intelligence Conference, He Xiaodong, vice president of Jingdong Group and president of the Exploration Research Institute, said that multimodal intelligence is the only way to the end of artificial general intelligence (AGI), and the Jingdong Yanxi model is an important exploration on this road. Up to now, JD Yanxi has successfully created 3 billion words of marketing copy and tens of thousands of live broadcast scripts, demonstrating its strong strength in the business field.
He Xiaodong believes that to achieve AGI as smart as people, machines need to learn to acquire and fuse knowledge in a multimodal way like people. In the future human-computer interaction, multimodal communication methods such as language, vision, and voice will become the mainstream, so machines must have the ability to understand this information in order to better serve humans.
Jingdong Yanxi model is a new generation of pre-trained language model with hundreds of billions of parameters, its pre-training parameters cover 3000+ categories, and the manual review pass rate is as high as 95%+. As JD.com's version of ChatGPT, Yanxida's model will go deep into retail, logistics, industrial and other industrial scenarios, give full play to the advantages of multimodal intelligence, and promote business model innovation and change.
My opinion: The successful application of JD Yanxi model not only highlights the great potential of multimodal intelligence in the business field, but also provides enterprises with a new and efficient way of business operation. With the help of the Yanxi model, enterprises can create marketing copy, live scripts, and other content more quickly, reduce labor costs, and improve operational efficiency. However, we should also be wary of the risks that may be brought by over-reliance on machine intelligence, and always maintain a cautious attitude towards technology to ensure the healthy development of human-robot collaboration.
Fourth: Rice intelligent breeding has entered the 4.0 era: digital wings help agriculture
According to Science and Technology Daily, mainland rice breeding is entering the stage of intelligent breeding 4.0. At the Global Digital Economy Conference and the Special Forum on the Revitalization of the Seed Industry Enabled by the Digital Economy, Qian Qian, an academician of the Chinese Academy of Sciences, said that at this stage, biotechnology and digital technology will be used for gene editing to greatly improve the utilization rate of germplasm resources, thereby releasing the value of resources and helping to fight the battle of turning over the seed industry.
Digitalization is becoming a trend in technological innovation in the seed industry. Li Liang, deputy director of the Genetic Breeding Center of the Institute of Crop Science, believes that digital technology can achieve "prediction before verification", expand the scale of breeding, and accelerate the breeding process. The government and enterprises of Pinggu District of Beijing Municipality are jointly building a digital seed industry big database to promote the automatic trait collection in the livestock and poultry seed industry, which is expected to achieve an exponential leap in breeding work.
Beijing Tongzhou International Seed Technology Co., Ltd. has built a common technical service platform for smart breeding, cooperated with scientific research institutes to develop molecular markers on a variety of cash crops, and provided resistance screening for vegetable seed enterprises across the country. The National Crop Germplasm Bank 2.0 project is also underway, cooperating with Tencent to carry out digital germplasm storage and computing, and provide digital support for breeding institutions and researchers.
The application of digital technology and artificial intelligence in the field of breeding needs to build a bridge between technology enterprises and the field of agricultural breeding. For example, artificial intelligence ChatGPT can reduce the workload for breeders in statistics, mathematics and other fields, and extract more scientific breeding laws.
My opinion: Combining digital technology and artificial intelligence, the entry of mainland rice intelligent breeding into the 4.0 era is undoubtedly an important breakthrough. Digital technology can not only improve the utilization rate of germplasm resources, but also greatly accelerate the breeding process and help the development of mainland agriculture. However, this is not the end, in the future development, we need to continue to deepen the cooperation between science and technology enterprises and agricultural breeding, innovate the application of digital technology, and bring more intelligent and efficient solutions to agricultural production.
So, what do you think of this good news?