New Initiative Will Pit Machine Learning Against Climate Change
The European Commission has announced the funding of a new Innovative Training Network, led by Oxford University, which will train PhD students in Machine Learning Skills to address Climate Change.
iMIRACLI (innovative MachIne leaRning to constrain Aerosol-cloud CLimate Impacts) brings together leading climate and machine learning scientists across Europe with non-academic partners, such as Amazon and the MetOffice, to educate a new generation of climate data scientists.
The project will start in 2020, with students beginning their projects in September 2020, kicking off with a summer school held at Oxford. It will fund 15 PhD student across Europe, with three of them directly supervised in Oxford (two in Physics, one in Statistics). Oxford is the overall lead of the project and Philip Stier, Professor of Atmospheric Physics at Oxford University, is the lead PI.
Each student will have a climate science and a machine learning supervisor as well as an industrial advisor. All students will have secondments to their industrial partners as well as to the co-supervisor.
While there is now general acceptance that Climate Change is being influenced by human activity, with historic agreements such as the Paris agreement aiming to keep global mean temperature rise below 2oC of pre-industrial levels, the understanding of climate change at a quantitative level is still subject to large uncertainties. This is to a large extent because of the uncertain role of clouds in the climate system.
It is this quantitative understanding of the role of clouds for climate change that the consortium of nine Universities, led by the University of Oxford, will address with an innovative programme of study that provides each PhD student with co-mentorship and supervision from a climate and data scientist.
Machine Learning, underpinned by Artificial Intelligence, has undergone rapid advances in recent years and offers new tools to study, analyse and learn from the mass of data being collected by Earth Observations.
Modern satellites, airborne and ground-based instruments provide unprecedented observational data that coupled with Machine Learning techniques can enable the coupling of predictions and real-world observations.
Prof Philip Stier, said: ‘Machine Learning has the potential to unlock unique in-depth understanding of the climate system from vast climate datasets. However, this requires a new generation of experts with substantial knowledge of both climate and data science. We will train and shape a new generation of climate data scientists, with a solid foundation in climate science and a competence in the latest machine learning techniques.’
This article has been republished from the following materials. Note: material may have been edited for length and content. For further information, please contact the cited source.