[China Science Journal] Become a storm driver – 夜上海论坛-爱上海419论坛-上海后花园论坛-上海龙凤1314

　　Scientific research is experiencing a storm. The storm originated from sturm und drang’s artificial intelligence technology. "Embrace" or "escape" has become a multiple-choice question for every scientific researcher.

　　In China University of Science and Technology (hereinafter referred to as China University of Science and Technology), a group of "ambitious" young people chose to hug. They focused on the most traditional subject-chemistry field, started from humble ideas, broke the research convention, and their achievements were frequently published in Nature and Science. They should strive to be the masters of the storm in this transformation of scientific research paradigm.

　　Initial thought: let the computer help.　

　　If we didn’t choose to "join hands" with computers, the latest paper published in Nature by Yao Hongbin, a professor at China University of Science and Technology, may be just an experimental report.

　　In early April, this paper caused a sensation as soon as it was published. Yao Hongbin’s team’s research on the new family of solid-state electrolytes is considered to be expected to realize the "Holy Grail" all-solid lithium metal battery in the field of electrochemical energy storage.

　　Yao Hongbin and others tried to find a metal halide to improve the efficiency of solid-state lithium batteries. In the process of searching for halide electrolyte, they once considered replacing the "oxygen" in garnet oxide-lithium lanthanum zirconium oxide with "chlorine", but how to synthesize this compound is unknown. According to the traditional chemical research methods, researchers have to stay in the laboratory for a little trial and error. I don’t know when I can try it right.

　　Yao Hongbin thought, "Is it possible for the computer to help?"

　　Since there is no one in the team who can play with computers, Yao Hongbin began to collect talents. In 2021, during the postgraduate re-examination, he chose Luo Jinda, an undergraduate from the School of Materials Science and Engineering of Xiangtan University, because he had the foundation of computer programming and was very interested in it.

　　On the night when the admission list was confirmed, Yao Hongbin dialed Luo Jinda’s phone: "We have a material system which is very important and needs computational chemistry. Would you like to join us? I will contact the teacher to teach you. "

　　The reason why Yao Hongbin is so active is that this research, which has been done for four years, is already facing financial difficulties and cannot be delayed any longer.

　　After picking Luo Jinda, Yao Hongbin quickly found Li Zhenyu, a professor of computational chemistry at China University of Science and Technology, and gave him Luo Jinda.

　　Under the joint guidance of Yao Hongbin and Li Zhenyu, Luo Jinda wrote a program to meet the research needs. After that, Yao Hongbin’s team and Li Zhenyu’s team designed a lanthanide metal chloride that can exist stably at room temperature according to the computer simulation results. Following the guidance of computer, they successfully synthesized lanthanide metal chloride solid electrolyte with high performance in the laboratory.

　　The sample has finally been made, but it can only be regarded as a "not bad experimental report". How the internal structure of the sample is, how lithium ions are conducted in it, and why this solid electrolyte has high performance are unknown.

　　At this time, the use of computers can be demonstrated again. Based on the principle of molecular dynamics, the team members input their own experimental data and massive experimental data of related research in history into the Supercomputing Center. After a long period of calculation, simulation and analysis, the lithium ion conduction principle of lanthanide metal halide frame structure is finally proved.

　　"In this study, the component of analog calculation accounts for about one third. Assuming that there is no such one-third, the research will not be satisfactory, because we may not be able to synthesize the optimal electrolyte materials and explain the principles behind the experimental phenomena clearly. " Yao Hongbin said.

　　The reviewer of the paper thinks: "The author puts forward a brand-new solid electrolyte structural material, and the calculation and simulation are very perfect."

　　After tasting the sweetness of analog computing for the first time, Yao Hongbin was enlightened: "We suddenly found that we can do something else."

　　They now have a plan that they never dreamed of before: "Combine many elements in the periodic table in different proportions, or find materials that meet our expectations in the existing massive data to make the materials more diverse."

　　Breaking through tradition: training computers into "chemists"　

　　Yao Hongbin’s team made the computer a "helper", while Jiang Jun, a professor at China University of Science and Technology, tried to train the computer into a thinking "chemist".

　　Jiang Jun didn’t come from a chemistry class. He first studied physics, and then slowly turned to theory and computational chemistry. Cross-discipline has made Jiang Jun somewhat uncomfortable: "The physics field pursues common laws, but chemistry pursues individual characteristics. Every laboratory hopes to be different from others, and the more different it is, the better."

　　When he gives lectures to students, he often finds that he has just finished a rule of chemical experience today, and there is an exception the next day. "So-and-so experiment has broken the previous theoretical limit."

　　The disconnection between theory and practice once made Jiang Jun "very collapsed". He began to try to find the possible common laws in this ancient discipline. Until 2014, he saw the possibility of solving problems-using big data and artificial intelligence.

　　"The reason why chemistry is personalized and can’t see the law is because it is a complex system with high dimensions. Any small change in conditions will cause changes in the whole system." Jiang Jun said, "Maybe we can build a’ chemical brain’ through data intelligence and find the common laws in chemistry."

　　This can’t be done by yourself. Like Yao Hongbin’s original choice, he began to look around for like-minded people. In his team, there are students from artificial intelligence, electronic science and technology, mathematics and other directions. They take data as the entry point, write intelligent programs, build databases and build knowledge maps, so that computers can gradually read documents and do experiments.

　　At first, someone disdainfully said, "It’s just automation." However, after eight years of exploration, the machine chemist has been upgraded from version 1.0 to version 2.0, and what the "chemical brain" can do covers the whole chain of chemical research such as literature reading, experimental design, formula and condition optimization.

　　After the birth of machine chemist, the first attempt of Jiang Jun’s team was to let it calculate and predict protein spectrum. This paper was recommended by the editorial department of Science magazine. "Generally speaking, to calculate a protein spectrum, it takes more than 1,000 dynamic conformations, and each conformation takes more than a week. We used machine learning to let machine chemists do it, which was shortened from more than 20 years to one or two hours. " Jiang Jun said.

　　Later, Jiang Jun’s team asked the upgraded machine chemists to create chemicals and materials such as high-entropy catalysts. Machine chemists independently selected five non-precious metal elements from 16,000 papers, and found the global optimal solution from 550,000 possible metal ratios, shortening the 1400 years required by the traditional "trial and error method" to five weeks.

　　According to the international peer evaluation, the network operating system, workstation and intelligent chemical brain of machine chemists are all the most advanced, and it is predicted that it will have a great impact on chemical science.

　　All these make Jiang Jun feel that "the original direction was not wrong." By 2022, the name of "Machine Chemist" appeared in the list of "Annual Team of China Academy of Sciences" and "Top Ten Progress in Online Information Work".

　　In the future, their goal is to promote Shenyang Institute of Automation of Chinese Academy of Sciences, Shenzhen Institute of Advanced Technology of Chinese Academy of Sciences, Institute of Computing Technology of Chinese Academy of Sciences, and China University of Science and Technology to jointly set up smart laboratories. At the same time, they will join hands with the Literature and Information Center of Chinese Academy of Sciences, the Institute of Automation of Chinese Academy of Sciences and the Institute of Information Engineering of Chinese Academy of Sciences to build a global knowledge platform and form a full-scenario solution for accurate and intelligent chemistry.

　　In addition, there is another "ambition" hidden in Jiang Jun’s heart-he wants to understand and even crack the high-dimensional dependence of chemistry through machine chemists, and rely on powerful computing power to realize the innovation and reform of chemical theory; He also wants to find out the commonalities of chemistry, physics, biology and other disciplines, and derive machine physicists and machine biologists from machine chemists …

　　At present, in a laboratory of more than 100 square meters in the Physical and Chemical Building of the East Campus of China University of Science and Technology, two machine chemists, Xiao Lai and Xiao Fu, shuttle between different experimental platforms and instruments. Jiang Jun lamented: "In the future, the core competitiveness of scientists may no longer be the ability to do experiments, but a more solid professional foundation, a stronger understanding and an imaginative imagination."

　　Set off a storm: concentrate on exploring the fourth paradigm　

　　Like young people, Academician Yang Jinlong, head of the single molecule science team of China University of Science and Technology and vice president of China University of Science and Technology, saw the impact of artificial intelligence and big data on chemistry, and also saw the attempts of young people from different majors to break through the shackles of scientific research tradition.

　　Thus, in 2020, in China University of Science and Technology, the matter of "alliance" was put on the agenda. After 23 meetings, they combined the existing scientific research resources of China University of Science and Technology and concentrated on exploring the "fourth paradigm" of chemical research in the form of institutionalized scientific research institutions.

　　"Paradigm change" is a term often mentioned by the scientific community after realizing the impact of big data and artificial intelligence. It is like a warning, warning the whole scientific community that a storm sweeping the world outlook, values and methodology of the discipline is coming.

　　In the field of chemistry, the development history of more than 150 years is divided into four paradigm stages: first normal form discovered chemical laws through experimental observation and induction based on experiments, and lavoisier discovered the law of conservation of mass as the representative study; The second paradigm is based on theory and explains chemical phenomena through mathematical model and theoretical framework. The representative research is Mendeleev’s discovery of the periodic law of elements. The third paradigm is to use computers and artificial intelligence to simulate and predict chemical processes and results, and the representative research is quantum chemical calculation, molecular dynamics simulation, etc. The fourth paradigm is driven by data intelligence, such as material property prediction based on machine learning.

　　"In more than 150 years of exploration, chemical research mainly relies on the’ trial and error method’, and accuracy and efficiency are the dreams of all chemists." Li Zhenyu said that every paradigm change has promoted the involvement of computers in the field of chemistry, and precise and intelligent chemistry has become an opportunity to realize the leap-forward development of chemistry.

　　On January 18th this year, the Key Laboratory of Precision Intelligent Chemistry of China Academy of Sciences (hereinafter referred to as the Laboratory) was formally established in China University of Science and Technology, with Li Zhenyu as the laboratory director.

　　The laboratory takes the chemistry discipline of China University of Science and Technology as the main body, gives full play to the three advantages of interdisciplinary, "integration of science and education, collaboration between colleges and universities" and Hefei scientific apparatus gathering, focuses on the problem of "how to change the paradigm of chemical research", and explores the establishment of a dual-drive model of precision and intelligence in chemical research.

　　Talking about the significance of the fourth paradigm, Li Zhenyu made an analogy: "Riding a bicycle is too tired to ride on the moon. The fourth paradigm is like building a rocket and going to places that you couldn’t go before."

　　"We hope to build this laboratory into a top international research institution in the field of precision intelligent chemistry, form a new paradigm of precision natural chemistry, and establish our country-led precision chemical data system and intelligent chemical software and hardware standards." Li Zhenyu said.

　　At present, they have gathered Yao Hongbin, Jiang Jun and other research teams from theoretical calculation, accurate characterization, intelligent chemistry, precise synthesis, and physical property control. In addition to their achievements in Nature and Science, these teams have one thing in common-related to precise and intelligent chemistry.

　　The team led by Xu Tongwen, a professor at China University of Science and Technology, has been studying ionic membranes for nearly 30 years, and recently published a new achievement in Nature. "We all rely on the calculation and simulation methods related to precision chemistry in the selection of materials and the regulation and design of material structure channels." Xu Tongwen said that the laboratory has opened a new window for them to cross and cooperate.

　　For Li Zhenyu, director of the laboratory, the future of the laboratory is bright, but the road is tortuous. As a laboratory that relies on interdisciplinary development, an important problem that must be solved is "how to attract more high-level cooperative teams and top talents".

　　Kuhn, the American philosopher of science who first proposed the concept of "paradigm change", once said: "At first, the new paradigm candidates may have only a few supporters, and sometimes the motives of these supporters are suspicious. However, if they are really capable, they will improve it, explore all its possibilities, and show what prospects the community will have under its guidance. "

　　Li Zhenyu often persuades his collaborators to join and support them on various occasions. Ineffective, he will sum up the prospect of the fourth paradigm and their attitude with the simplest and simplest 12 words, "embracing the future of chemistry precision and intelligence".