On February 29, 2024, the National Natural Science Foundation of China released the “Top Ten Progress in Chinese Science” for 2023, mainly in the fields of life sciences and medicine, artificial intelligence, quantum, astronomy, chemical energy and other scientific fields. Among them, two scientific and technological research results from Shenzhen were selected, namely “Artificial Intelligence Large Model Brings New Breakthroughs for Accurate Weather Forecasting” from Huawei Cloud Computing Technology Co., Ltd. and one jointly completed by Southern University of Science and Technology, Shenzhen International Quantum Institute and other units. “Bose coding error correction extends qubit lifetime”.
Shenzhen Dream believes that the two scientific and technological research results in Shenzhen are of great significance. The scientific progress of artificial intelligence and quantum computing is pushing Shenzhen into the fourth industrial revolution. In 2023, Ren Zhengfei held an on-site dialogue with the ICPC competition team and said that we are about to enter the fourth industrial revolution, and the foundation is large computing power. “
At the same time, on February 26, the 2024 World Mobile Communications Conference (MWC24) opened in Barcelona. As one of the world’s leading companies in the communications industry, Huawei held multiple press conferences during the MWC24 Barcelona exhibition. Huawei announced that 2024 is the first year of 5G-A commercialization, and the company will actively explore the evolution to the 5G-A era with global operators.
It is worth noting that Huawei released the first large-scale model in the communications industry. In response to the industry’s high-level intelligent goals of agile service provisioning, accurate user experience guarantee, and efficient cross-domain operation and maintenance, the large model provides role-based and scenario-based intelligent applications to help operators empower employees and improve user satisfaction. Comprehensively enable network productivity.
At present, quantum computing technology is one of the strategic commanding heights that world powers are vying to seize. On March 22, 2023, the result of Academician Yu Dapeng’s team was published on the website of the top international academic journal “Nature” under the title “Logical qubits encoded with discrete variables break the break-even point”. In the view of Academician Yu Dapeng, quantum computers have the potential to provide a disruptive computing power and provide powerful computing power support for the fourth industrial revolution . As a “city of science and technology”, Shenzhen has great potential in the field of quantum technology.
On February 29, the National Natural Science Foundation of China released the “Top Ten Scientific Progress in China” for 2023 , mainly in the fields of life sciences and medicine, artificial intelligence, quantum, astronomy, chemical energy and other scientific fields.
Revealing the mechanism by which dark matter drives aging in the human genome
-Discover the existence of the brain’s “tangible” biological clock and its rhythm regulation mechanism
-Analysis and application of salt-alkali tolerance mechanism of crops
-New method enables precise manipulation of DNA from single bases to very large fragments
-Revealing a new mechanism of DNA replication initiation in human cells
-“Raso” discovered the extremely narrow jet and 10 trillion electron volt photons of the brightest gamma burst in history
-Bose coding error correction extends qubit life
-Revealing the mechanism of photoreception regulating blood sugar metabolism
-Discover a new mechanism of charge storage and aggregation reaction at the interface of lithium-sulfur batteries
The reporter learned on the spot that two research results from Shenzhen technology companies, universities and scientific research institutions were on the list. They are: the research results of Tian Qi and others from Huawei Cloud Computing Technology Co., Ltd. ” Artificial Intelligence Large Model Brings New Breakthroughs for Accurate Weather Forecasting” ; research from Yu Dapeng, Xu Yuan and others of Southern University of Science and Technology, Shenzhen International Quantum Research Institute The result is “Bose coding error correction extends qubit lifetime” .
It is understood that the selection activity for the “Top Ten Progresses in Chinese Science” was launched in 2005 and has been successfully held for 19 sessions. This event was hosted by the National Natural Science Foundation of China and voted by more than 2,100 high-level experts in the field of basic research, including academicians of the Chinese Academy of Sciences and the Chinese Academy of Engineering. After four steps of recommendation, preliminary selection, final selection, and deliberation, 600 people were finally selected. Ten major scientific research results were selected from a number of scientific research results . After review by the Advisory Committee of the National Natural Science Foundation of China, the list of results selected for the “Top Ten Progress in Chinese Science” in 2023 was finally determined.
Weather forecasting is an international scientific frontier issue and has great social value. The existing numerical weather prediction paradigm originated in the 1950s, which uses large-scale calculations on supercomputing platforms to solve a system of partial differential equations of atmospheric motion to predict future weather.
In recent years, as the frequency of extreme weather has increased, improving forecasts using this traditional method has become increasingly challenging. Based on artificial intelligence technology, Tian Qi, Bi Kaifeng, Xie Lingxi and others from Huawei Cloud Computing Technology Co., Ltd. proposed a three-dimensional neural network adapted to the earth coordinate system, which can effectively handle complex processes in weather data and use hierarchical time domain aggregation strategies. to effectively reduce iteration errors and successfully achieve accurate mid-term weather forecast.
After training on global weather reanalysis data from 1979 to 2017, a large Pangu meteorological model was constructed. The model is able to predict meteorological elements such as temperature, air pressure, humidity, and wind speed at the surface layer and 13 upper-altitude layers within 7 days, and improves the forecast timeliness of the world’s most advanced European Center for Medium and Long-term Weather Forecasts (ECMWF) integrated forecast system by 0.6 In about days, the tropical cyclone track forecast error is reduced by 25% compared with the ECMWF forecast system. In addition, the model can complete the global 7-day forecast of important meteorological elements in just 10 seconds, and the calculation speed is more than 10,000 times faster than numerical methods .
