Scientists and professors shared their views on artificial intelligence at the first Shanghai Master Forum on Science, held on Nov 15 in Fudan University's Xianghui Hall.
Professor Michael Levitt, laureate of the 2013 Nobel Prize in chemistry, was the forum's keynote speaker. He joined Fudan University in 2018, where he founded the Multiscale Research Institute of Complex Systems.
Levitt's speech, titled "AI for Science: Frontiers in Computational Biology" gave answers to defining science.
The speech focused on initial thoughts related to science, types of science, education in science, science practice and future science.
"Nature can do amazing things. Basic molecules such as DNA, proteins and ribosomes are really complicated. I would image one cell is as complicated as a big city, so that nature is a very smart constructor," Levitt said, adding that perceptrons and deep-learning were what people took away from biology.
He later elaborated on a new kind of science — crisis science, which "is when you have to make scientific observations when time is very short", he said.
Levitt said he wanted to use artificial intelligence for something different. As AI can be used to draw pictures, he gave text to engines and created scenes of crisis, including the Hoover Tower at Stanford University on fire and Cambridge University in a flood. "I think it is useful to make us scared. We need to realize there will be these sudden events and we need every crisis to learn," he said.
For him, the most important future use of AI will not be protein folding or translation.
"One thing that human beings are very bad at doing is assessing risk. We don't know how risky it is to walk on the street or to inhale bad air or to eat bad food or to smoke," he said.
Levitt said in the next decade, people may be able to ask their phones whether they should panic because of a certain circumstance and they will be confident that the answer is impartial and accurately supported by all available data. "Human beings will be much more rational," he added.
During his speech, he noted a recipe for winning Nobel Prizes. That is ample research support, freely available supplies, advanced equipment and computing, small groups about five in all, intense peer pressure, and an atmosphere in which students feel as good as Nobel laureates. "Young people ought to be smart, have crazy ideas and not have to fill in forms. If you want a chemical, you just went and took it. There is no need to justify anything," he said. Levitt said his mentors should take credit for his success. When he was 20, he was told to go out and do his own work. With that, he wrote his first paper.
"A good mentor is not somebody that helps you do what you do. A good mentor tries to make you replace him. He lets you stand on his shoulders and say 'I want you to be better than me'," Levitt said. His thought of mentoring was echoed by two co-speakers — professor Ma Jianpeng and professor Qi Yuan, both from Fudan University.
Ma, dean of the Multiscale Research Institute for Complex Systems of Fudan University, said: "Biology is becoming more complicated."
Experimental means are limited but data are getting bigger and bigger.
As a result, biology relies on calculation methods. That explains why computational biology has become a leading science and why it is becoming more important.
"It's imperative for us to cultivate talents. The forum is the only one of its kind that takes place in a university and helps to deepen students' understanding of development of computational biology."
Qi Yuan, who was the chief AI scientist and former vice-president of Ant Group, is now dean of the Artificial Intelligence Innovation and Incubation (AI³) Institute of Fudan University.
"The invention of telescope brings the previously unseen universe to human eyes, preluding the discovery of Kepler's laws of planetary motion and Newton's law of universal gravitation. Today, in the era of computing, AI is becoming the telescope and microscope driving new scientific discovery and propelling economic development. AI is integrating with life sciences, chemistry, material sciences, meteorology and other basic research, moving scientific research to a new computation-based paradigm; and this integration will generate original innovations that can, in turn, advance new drug discovery, intelligent health management and carbon neutrality," Qi said.
He added: "To link all the elements together, we need a new innovation system that supports AI for science research and guides the search by critical unmet social and economic needs. I have worked in both academia and industry. A critical lesson I learned is that there is an important gap between a good idea and a great product. We need new types of innovation and incubation centers that link university researchers with engineers, product developers and funding support; we need to cultivate and assemble interdisciplinary talents who know AI, science and business and work together to solve important scientific and engineering problems."