About the AI & Sustainability Lab

AI & Sustainability Lab

In the AI labs we work on societal challenges from a multidisciplinary perspective. One of the most pressing societal issues is the sustainability of our environment. In the AI & Sustainability Lab, researchers from the Faculty of Geosciences and the department of Information and Computing Sciences of the Faculty of Science will collaborate with other departments, institutes, companies and public organisations. The researchers will use various innovative techniques from AI and data science to address sustainability challenges such as climate change, the energy transition, and urban systems.

Optimising Earth Systems and Digital Twins: Towards sustainable futures

The long-term effects of climate change on earth are challenging to predict. And with that, so are the long-term effects of policies and legislations implemented in order to mitigate climate change. With the use of AI, we will be able to safely test those effects in a digital space using Digital Twins of the Earth: computer-based simulations that try to reproduce, as faithfully and accurately as possible with current high-performance computing power, a system or phenomenon from the real world. This may include naturally-occurring or man-made systems, or relevant combinations of the two.

These tools help us understand complex dynamics and interactions providing possible pathways to more sustainable systems. The high predictive power of the digital twin allows us to test interventions that will push outcomes in a desirable direction, such as climate change mitigation measures. These projects will give decision makers the tools and information they need to take those measures responsibly and efficiently.

Information extraction from Big Geoscience Data

Basic and applied research in the Geosciences are undergoing a fundamental change because of the exponential increase in available data. How do we extract and combine information from a multitude of data sources in a consistent, correct, safe, reliable and efficient manner? 

Social media, automated sensing, and large digital databases provide enormous potential for improving our understanding of Geoscience systems as well as for reconstruction and prediction of spatio-temporal phenomena, essential for decision making and sustainability. To make sense of these massive, complex data sets, AI methods are needed that are capable of extracting useful information from crude data.

Making Geo-data accessible

Geoscience (research) data potentially have an enormous value for research within and beyond the Geoscience domain and for scientifically informed policy and decision making at multiple levels: from the use of geographically referenced data for personal decision making to policy making regarding climate change and environmental sustainability. However, it is still a challenge to spread, archive, and use data in an effective, FAIR manner; especially when it concerns Big Data sets.

The AI & Sustainability Lab aims to develop methods to organise, store and make accessible Geoscience data in databases, methods for data interoperability for exchange of data between databases and between computer models, methods to convert data into meaningful information that can be used by the public, and scientific visualisation routines for Geoscience data. By tackling this challenge, the lab contributes to FAIR data and software, which is a crucial part of Utrecht University’s ambition to work towards Open Science.

Artificial Intelligence for Sustainability

Artificial Intelligence and data science can make an important contribution to these challenges. Key technologies include:

  • Explainable AI - methods, techniques, and tools aimed to help practitioners to understand, steer, and improve the working of AI algorithms by ‘opening the black box’ that many such algorithms actually are. By being able to look into the inner workings of algorithms, in the future we will be able to recognise uncertainties and biases in order to improve policies and legislation based on algorithms.
  • Large scale data-intensive systems - to develop new techniques and algorithms that allow the construction, combination and execution of (very) large scale data-intensive and/or computing-intensive applications.
  • Human-centric computing - to facilitate a better understanding of users of complex computational ‘decision support’ systems and to increase acceptance. Where does the human come into the picture? And where is the human side in decision-making based on AI?

Utrecht University network

Research in the AI & Sustainability Lab is centred around Utrecht University’s strategic theme Pathways to Sustainability. The work will build upon and complement that of Utrecht University’s Digital Humanities Lab and the focus areas Applied Data Science and Human-centered AI, and will work towards solving sustainability issues as well as continuously innovating and developing novel AI and data science techniques.