Outside of science, and until the pandemic, most ordinary people were not aware of scientific modeling.
After two years of hearing predictions about which way virus variations were headed, and the way researchers come up with these numbers, that’s changed. Computational models look at the course of the pandemic regionally and globally. They’re the basis of policy decisions on public health.
Computer models are also used in climate science and almost every other type of science.
“You're doing some level of computation if you're doing advanced research,” said Michael Barton, an Arizona State University geoarchaeologist and anthropologist.
But how do scientists share, integrate and reproduce each other’s work? There’s never been a unified framework — until now.
After seven years of work led by Barton, 66 participants representing 50 organizations supporting modeling science on five continents have established the Open Modeling Foundation.
The new foundation will administer international community standards for computational modeling across social, ecological, environmental and geophysical modeling — enabling open, reproducible and integrated modeling science.
“This was a historic moment that brought the international modeling community together as never before and will hopefully play an important role in supporting sophisticated, next-generation modeling in the future,” said Barton, a professor in the School of Human Evolution and Social Change.
As Barton says, the foundation will help provide the infrastructure to better plan for the future.
“Computation is ubiquitous in science,” said Barton, who is also associate director of the School of Complex Adaptive Systems. “It used to be you'd run a program to do some statistics or an analysis that maybe have a word processor, and now it it's almost any field. Increasingly the computation is the science.
"Probably the two best known categories of those models for the general public today are climate models. So these are large computational models that scientists have set up. What they do is then they can generate outcomes that look like the way the atmosphere and the land and the ocean might be in the future. The research is in the model. That's where all the science takes place.”
Science is like scaffolding. You take someone else’s work and enhance it. Transmitting that knowledge is traditionally done in scientific journals. (And sometimes through the popular press.)
“And that's great for describing many kinds of research, how we collect data, analyze data scientific insights, observations and things like that,” Barton said. “But for computation and models, that's not enough because just describing the result you got doesn't help anybody learn how you did it and be able to replicate it, to build on it or even check it.”
As your sixth grade math teacher said, show your work. But doing that with computational modeling doesn’t work because it doesn’t fit into the journal world. There aren’t any norms or standard practices.
“There's a good chance that no one will ever cite the code that you just published, or if they do cite it, it's almost certain that it will never show up in your citation,” Barton said.
What the Open Modeling Foundation has done is establish common standards that span multiple disciplines — social sciences, ecological sciences, environmental sciences, geophysical sciences, to name a few — that people can point to on the best way to make your code available, document it and package it so it’s reproducible by others.
“So you can link your model together with someone else's model or someone else could link your model and someone else's model, and be able to look at the dynamic interactions of different aspects of human systems and natural systems,” Barton said. “It's not so exciting as doing the research, writing the models, but standards are needed."
Top image courtesy of Pixabay
More Science and technology
How ASU is reshaping manufacturing training
The manufacturing sector faces a persistent challenge: a shortage of skilled workers.As industries adopt emerging technologies, traditional training methods struggle to meet the demand for skilled…
Podcast explores the future in a rapidly evolving world
What will it mean to be human in the future? Who owns data and who owns us? Can machines think?These are some of the questions pondered on a newly launched podcast titled “Modem Futura.” Co-…
New NIH-funded program will train ASU students for the future of AI-powered medicine
The medical sector is increasingly exploring the use of artificial intelligence, or AI, to make health care more affordable and to improve patient outcomes, but new programs are needed to train…