Mathematicians in China and the United States say they have developed a “model of everything” that simplifies complex systems in the body, society and the environment by revealing their patterns and interactions.
The team, led by renowned Chinese-American mathematician Yau Shing-Tung, says their model can be used to analyse many scenarios.
They range from how gut microbes interact and trigger inflammation, to how increased carbon dioxide and pollutants cause climate warming, and how firms compete and cooperate to boost the economy.
In biological scenarios, the model’s analysis could be used to help pharmaceutical professionals develop treatments – for example, for cancer by decoupling cells that are identified to be working together and giving rise to abnormal cell growth.
The model “provides a generic tool for unveiling hidden patterns in complex systems across a wide spectrum of physical and biological scenarios”, the team wrote in a paper published in the peer-reviewed journal Proceedings of the National Academy of Sciences last month.
The researchers are from the Beijing Institute of Mathematical Sciences and Applications, Tsinghua University, Pennsylvania State University and Creighton University School of Medicine.
Lead author of the study Wu Rongling, a researcher at the Beijing Institute, called it a “model of everything” since it could be used to explain the intrinsic patterns of social and natural phenomena.
Wu, who is also a chair professor at Tsinghua’s Yau Mathematical Sciences Centre, said the model provided a more advanced mathematical foundation for developing artificial intelligence that closely mimics the human brain.
Both institutes in the Chinese capital are directed by Yau Shing-Tung, a leading geometry expert who retired from Harvard University in 2022 to teach full-time at Tsinghua. He said he aimed to help cultivate talented young mathematicians so that China can become a maths powerhouse.
In the paper, the team said their model was “mechanistic in the sense that it incorporates evolutionary game theory and behavioural ecology principles into a unified statistical mechanics framework”.
Wu explained evolutionary game theory – which looks at how players strategise and make decisions – using the example of two cats wanting to eat.
“When two cats are together, they can either compete or cooperate as they seek to maximise their food intake,” he said. “They will choose their strategy based on their strength and how it could be influenced by the other cat.”
In the statistical framework, the cats are agents modelled as nodes and their interactions are seen as links to be coded into hypernetworks. The network expands as the number of nodes increases, he said.
In a biological scenario involving tumours, Wu said the different types of cells that give rise to tumours are analogous to cats.
“For a cell to grow, it needs to obtain nutrients. It can either compete or cooperate with other cells, just like the cats. By applying medical data on cells, the model can uncover the relationships between them,” he said.
“If the cells grow by joining forces, a medication could be designed to weaken their cooperation,” he said, adding that the immune system, as a node in the hypernetwork, could also be directed to halt this cooperation.
“If the cells compete with each other and cancer cells proliferate as a result, we should limit that competition to prevent one party from outgrowing the other.”
Wu said the mathematicians had worked with biologists to test the model, for example, for the gut microbiome.
“Using the model, we found that two types of gut microbes can cause inflammation of the gastrointestinal tract when they cooperate. In an experiment, biologists introduced the two microbes into mice and observed inflammation, confirming the model’s predictions,” he said.
In climate scenarios, Wu said the relationships between carbon dioxide and pollutant concentrations and other factors could be used to predict changes in temperature and rainfall, for example.
As for social phenomena, he said the model could analyse scenarios such as how companies and industries drive economic development through competition and cooperation, how managers can coordinate different parties to optimise efficiency, how an individual’s perspective can be influenced by others’ opinions, and how school subjects interact – such as whether learning languages helps students excel in maths.
Wu said the model could also serve as an underlying framework for AI and lay a mathematical foundation that was more complex and closer to the human brain than existing algorithms.
“Mathematics is at the core of AI. The current neural network algorithm is based on a relatively simple structure,” he said.
“With this more complex network carrying a large amount of data, we can boost AI development by closely mimicking how neurons in the human brain communicate to perform more efficient computing and achieve more accurate results.”
