Rossella Arcucci

Information

Bio

Rossella Arcucci is a mathematician, machine learner, and societal engineer. She obtained her Master’s degree in Applied Mathematics in 2008 from the University of Naples Federico II and her PhD in Computational and Computer Science in 2012.

Professor Arcucci has been actively involved in operational research since her PhD, continuing through her first postdoctoral position at the Euro-Mediterranean Centre on Climate Change.

She is currently an Associate Professor in Data Learning and AI for Good at Imperial College London, where she also serves as Director of Research at the Imperial Data Science Institute and Director of the Ada Lovelace Academy. In 2018, she founded and now leads the Data Learning Group, a research team dedicated to advancing Data Assimilation and Machine Learning through interdisciplinary and international collaboration.

Her research addresses fundamental questions on how to effectively use big data, improve its reliability, and reduce uncertainty in predictive model development to extract meaningful features and actionable outcomes. She pioneered the integration of Data Assimilation with Machine Learning, creating the field of Data Learning. Her work spans diverse areas including weather prediction, wildfire modelling, flood nowcasting, healthcare, and epidemic control, all with a focus on saving human lives in crises.

Arcucci’s expertise spans Data Assimilation, Artificial Intelligence for Earth Observation, and Physics-Informed Machine Learning, focusing on integrating heterogeneous data sources with physical models to produce interpretable and reliable predictions. Shee aims to advance AI methodologies that bridge data-driven intelligence with scientific knowledge, contributing to the development of trustworthy, explainable systems for climate, environmental, and societal monitoring aligned with ESA Φ-lab’s mission.

Areas of Expertise

My expertise lies in several key areas of Data Science and Machine Learning, including Data Assimilation, AI for Earth Observation, Computational Modelling, Physics-Informed Machine Learning (PIML) and Large Language Models (LLMs). These specialisations enable me to design innovative algorithms that integrate heterogeneous data sources, respect physical laws, and deliver interpretable and reliable predictions. The data I work with span a wide spectrum, from social media streams to satellite observations, sensor networks, and other real-world sources. By combining these diverse datasets with advanced AI and physics-aware methods, I can extract meaningful insights, enhance predictive models, and build systems that are both data-driven and scientifically sound. Such expertise can significantly benefit ESA Φ-lab projects by enabling the fusion of multi-source data with physical models, creating trustworthy AI tools for climate, environmental, and societal monitoring, and supporting the next generation of physics-informed AI solutions aligned with ESA’s mission of space innovation and Earth system science.

Vision for the Future

Driven by a vision of a future where AI and Physics-Informed Machine Learning work seamlessly with diverse data sources (from satellites to social media) to create trustworthy, interpretable, and impactful digital twins of our planet, I am committed to advancing methodologies that bridge data-driven intelligence with scientific knowledge. My long-term goal is to enable AI systems that not only analyse but also explain and predict complex environmental and societal phenomena, supporting decision-making at local, national, and global scales. By combining LLMs, multi-source data fusion, and physics-informed approaches, my work aims to projects at the ESA Φ-lab to accelerate innovation in Earth observation, climate monitoring, and sustainable development