Manpowered intelligent efficiency unimaginable in the past

Progress in scientific research, but

Advancing a new scientific revolution remains to be seen

In recent days, the Shanghai Handicapped Intelligent Laboratory and several other scientific institutions have jointly published the Global Medium-Term Weather Forecasting Model, Wind Wau. Based on reanalysis data validation, “window” was reduced by 19.4 per cent over the 10-day forecast error in traditional physical models.

Global medium-term weather forecasts target weather conditions for the next 14 days. Previous studies have shown that the effectiveness of global medium-term weather forecasts has only increased by one day per 10 years, owing to the complexity of the physical processes in the atmosphere and the size of the resources required for atmospheric modelling.

In an interview in the Chinese News Week, the scientist of the Shanghai artificial intelligent laboratory described “window” as a multi-model, multi-mission learning issue for the global medium-term meteorological forecasting mission, and as a basis for designing manual, smart forecasting methods. Based on retrospective forecasts, the performance of the “window” exceeds the most recent model published by Deep Mind, a world-renowned artificial intelligence company, with a breakthrough of 10.75 days of available forecasts.

The introduction of artificial intelligent methods to model the global atmospheric system is only one of AI for Science (industrial research driven by manual intelligence). In many areas, AI is changing multidisciplinary research with more than ever and more than tens of thousands. Visiting scientists noted that modern science is becoming increasingly complicated and that the rapid breakthrough of relevant AI technologies in recent years has made AI for Science a front line for international scientific research.

At the end of March, the Ministry of Science and Technology and the Natural Science Fund Committee jointly launched the “AI for Science” deployment. The Officer-in-Charge of the Ministry of Science and Technology stated that China has a good basis for manual, intellectual, scientific and arithmetical resources, and that further strengthening of systems layouts and integrated guidance is required to promote the deep integration of manual intelligentness and scientific research, to facilitate the pooling of resources and to enhance related innovation capabilities.

“The long-term impact of manual intelligent intelligence on scientific exploration has only just begun, from the protein structure forecast to the modelling of the climate system, from induced wave detection to understanding the universe.” Data science-renowned agency Dataconomy wrote in an article in November 2022.

In the past, totally unimaginable efficiency

The processing of data would change the conduct of scientific research, and the research fellow of the Institute of Physics of the Chinese Academy of Sciences, Liu Daqun, was deeply felt. As a resource scientist, he said that, more than 10 years ago, the completion of the study of three or four materials was smooth. Today, on the basis of advances in artificial intelligent, supercaling, etc., materials consisting of different elements are screened and even projected in hundreds of thousands of possibilities, judged by their material attributes, do not need to be counted for another time, experimentation, but only for rats.

According to Liu Daqun, five years ago a number of scientists, including him, pre-judged that, with technological advances, the next step in material science should not be to use data to help scientific research by merely coltaning individual materials. His team developed a database of materials called “Atomly”, including data on more than 300,000 inorganic crystals.

He described the attributes of virtually all substances in the natural world as a certain electronic act. Early in the 1960s, the academic community had found that electronic behaviour could be calculated through a method of calculating the neutron stream, with a view to predicting the nature of the material. The calculation of the materials has accelerated considerably, thanks to the calculation of supercomputers, which have created the “Atomly” data base; moreover, the structure of many inorganic materials in the database is to be forecast using manual intelligent models, which will be followed up with initial judgement.

With a strong database and high flux calculations, Liu stated that scientists in any group of elements could quickly search possible new compounds and presume their physical nature. In the “Atomly” database, if two elements of oxygen are hit, there will be a potential compound of 280 elements. If one of these is hit, it will be possible to see further specific data, such as its atomic space placement structure, its media nature and its capability. If scholars want to find a material in such compounds, they can look at the indicators, their nature and then carry out the next study.

This increased efficiency, he described as if the desired material had been shelled, and today it was “divided the sense that the fish were caught”.

On 8 March this year, Ranga Diyas, an assistant professor from the University of Roscher, United States, declared that the team discovered a hybrid of hydrogen, nitrogen and a rare earth element known as arsenic, which could achieve superconductivity under pressure of 21°C and about 1GPa (approximately 10,000 standard atmospheric pressure). This result was then heavily shelled within and outside the circle.

In order to validate this result, the Liu Daqun team proceeded quickly on 9 March. Using the aforementioned database, they calculated more than 1,500 compounds less than a week and submitted papers on 21 March, which found that hydrogen-nitrogen-toxicity could not be stable in three yuan renminbi. In other words, the results of the hyperlinked paper are to be consulted. Mr. Liu Daqun said that this was a time when there was no imagination.

At the end of 2022, an article published in the Gradient magazine wrote that the mother-to-child science behind them was chemical when predicting a hyperlinkage of protein and seeking new hyperconductor materials, vaccines or any other material that met specific needs. The magazine was established in 2017 and was created as a group of students and researchers at the AAI laboratory at the University of Stanford, United States.

Traditionally, chemical research has often been completed in laboratories equipped with test pipes and bottles. This article states that, with current artificial intelligent, data-centred technological advances and growing data volume, we may be witnessing a change in the methodology used not only to assist experiments but also to guide experiments.

Not only then, AI can become a “chemical” in fact. For example, in July 2020, researchers from the University of Liverpool developed a manual smart machine chemical. The machinery is human in character and can work independently in standard laboratories, like human beings.