Partnership between Climate AI Nordics and Klimatkollen
We are thrilled to share our partnership with Klimatkollen 🌍📈!
Read more on our partner page! Also check out this Swedish text about the partnership.
You May Also Enjoy
Understanding the complex Earth system with Machine Learning and Hybrid Modelling
Event date:
Webinar with Markus Reichstein, Max-Planck-Institute for Biogeochemistry and ELLIS Unit Jena. The Earth system is a complex, dynamic and strongly interconnected system, shaped by interactions between climate, ecosystems, biogeochemical cycles and human activities. Rapidly growing streams of satellite, in-situ and experimental observations, together with advances in machine learning, offer new opportunities to detect patterns, infer processes and improve prediction across scales. Yet purely data-driven approaches often lack physical consistency and interpretability, while classical process-based models remain limited by uncertain parameterizations and incomplete representations of complex feedbacks. In this talk I will discuss how machine learning and hybrid modelling can help bridge this gap. By combining the versatility of data-driven methods with the constraints and explanatory power of mechanistic understanding, hybrid approaches can support more robust, interpretable and physically consistent models of the Earth system. Examples from the terrestrial biosphere, land-atmosphere exchange, carbon and water cycles, and climate extremes will illustrate how such approaches can contribute not only to improved prediction, but also to deeper scientific understanding of Earth system dynamics.
EarthShift: A benchmark for real-world ditribution shifts in Earth observation
Event date:
Webinar with Kelsey Doerksen, University of Cape Town and Arizona State University. Geospatial Foundation Models claim to offer powerful solutions to simplify and accelerate real-world problems, enabling the capabilities to monitor, analyze, and predict changes on our planet. Current Earth Observation benchmarks to quantify the performance of these models focus on measuring performance on diverse tasks and applications, typically measuring generalization in-distribution. However, when models are deployed, they must generalize to many out-of-distribution scenarios, such as new time periods, geographies, and sensors; and in many contexts, these models are brittle. We introduce EarthShift: the first public testbed for benchmarking robustness across multiple realistic distribution shifts encountered in remote sensing. EarthShift enables users to measure distributional robustness by comparing performance in- and out-of-distribution using datasets from paired data sources, temporal windows, geographic locations, and sensors. EarthShift provides a testbed to guide future work to create foundation models that are robust and reliable in real-world applications.
A Critical Look at Explainable AI
Event date:
Webinar with Gustau Camps-Valls, University of Valencia. I will give a sarcastic and quite op-ed tour trying to explain why XAI is misleading us. Everything from SHAP plots to counterfactuals may look trustworthy, but underneath, they're often driven by correlations, not causation. In fields like climate, neuroscience and social sciences, that's a serious risk. Inspired by philosophy of science, I argue that explanations must go beyond surface patterns. Fortunately, the frontier is moving fast: causal‐informed SHAP, meaningful counterfactuals (you can't go younger), causal certification in explanations, and structural causal modeling are all promising. Yet, it's time we treat XAI not just as a cosmetic fix, but as a tool grounded in truth: seamful, thought-provoking, and scientifically defensible. And if time allows I'd like to say a few words about why AI needs a new philosophy of science.
More is different: emergent social conventions and tipping points in AI populations
Event date:
Webinar with Ariel Flint Ashery, City St George’s, University of London. Social conventions are the foundation of social coordination, shaping how individuals come together to form a society. In this talk, I will present theoretical and experimental findings that demonstrate the spontaneous emergence of social norms in LLM populations, as well as the existence of tipping points in social convention. I will show that agentic AI populations can establish social conventions and highlight how collective biases can emerge even when individual agents appear unbiased. I will conclude by stressing how the ability of AI agents to develop norms without explicit programming has significant implications for designing AI systems that align with human values and societal goals.