A key problem in predicting aerosol formation is the enormous number of compounds, reactions and processes involved as well as the complexity of the these compounds and the difficulty of measuring the processes. The VILMA project aims to understand aerosol formation by establishing a set of virtual instruments and models.
VILMA, or the Virtual Laboratory for Molecular-Level Atmospheric Transformations, is a Centre of Excellence funded by the Academy of Finland in the CoE Programme 2022–2029. The virtual laboratory provides researchers with the opportunity to tackle many unsolved problems in atmospheric sciences, including the identification of reactions responsible for the formation and growth of organic nanoparticles.
VILMA combines three areas: the theory and modelling of atmospheric phenomena, experiments, and machine learning and artificial intelligence.
– In the AI component of the project, our goal is to create digital twins of atmospheric processes combining physical simulators and a range of AI and machine learning models. Subsequently, we intend to investigate these processes with various AI-based tools. This is what we call a virtual laboratory, says Associate Professor Kai Puolamäki from the University of Helsinki. Puolamäki also coordinates the Atmospheric AI program at FCAI.
One problem is that the physical simulations currently in use are computationally cumbersome. A solution to this is to develop machine learning models that are able to quickly emulate slower physical simulations.
In other words, VILMA will pilot novel research methods whose development takes years. There are no ready-made solutions – instead, everything must be developed by the researchers. With Academy funding, new ideas can now be tested and developed in the long term.
Puolamäki’s focus is on explainable AI, such as how digital twins function, or whether there are simpler but potentially more explainable models that could be used to describe the relevant processes.
– This is also about how we can build these digital twins in an interactive manner, while understanding how they work. Managing uncertainty is another big issue. Can we trust the results, how can we calculate confidence intervals for the results, and is the model used outside its scope of application? These are surprisingly difficult problems, especially in the case of complex machine learning models, Puolamäki says.
In the service of sustainability goals and future generations of researchers
In terms of both the topic and goals, the project complies with the goals of sustainability, even though sustainability is not directly investigated in VILMA.
It’s of course clear that this is basic research. The results of our work relate to, for example, air quality modelling. However, we will come up with methods that can be applied in other fields as well in the acquisition of research-based knowledge, says Puolamäki.
Such new methods can have a significant impact on the natural sciences and experimental research.
– We are solving key problems in atmospheric sciences, for which new methods are needed. Then again, from the perspective of computer science we are investigating how methods based on artificial intelligence can be applied in the natural sciences, which may have a considerable effect on the future conduct of such research based on measurements, Puolamäki says.
Ultimately, the goal is to gain the ability to model the atmosphere and the world more broadly also outside molecular transformations, VILMA’s current focus.
The aim is to disseminate results and AI methods, for example, by publishing open-access software and tools for the virtual laboratory. Doctoral and postdoctoral researchers also are being recruited for the project.
The Centre of Excellence is also taking the general public into consideration. In the future, versions of the VILMA virtual laboratory tailored for science communication will also offer schoolchildren and the general public the chance to gain insights concerning not only the atmospheric sciences, but also the scientific method in general.
VILMA is a Centre of Excellence in research funded by the Academy of Finland and carried out by the University of Helsinki, the University of Eastern Finland, Tampere University and Aalto University.
This story was originally published on FCAI´s website.