[2021] Pavlos Andreadis and three 4th year students: Calum McMeekin, Christoph Minixhofer, Mark Swan, authored a paper which takes a significant step towards a global drought prediction model and enables future such research by creating the first dataset and models for drought prediction across the Continental USA. The paper was accepted for inclusion in the International Conference on Machine Learning (ICML) 2021 Workshop on Tackling Climate Change with Machine Learning. Image 'DroughtED: A dataset and methodology for drought forecasting spanning multiple climate zones', is a continuation of the students' project coursework for Machine Learning Practical this year. Previous attempts to forecast drought conditions using machine learning have focused on regional models which have two major limitations for national drought management - they are trained on localised climate data, and their architectures prevent them from being applied to new heterogeneous regions. In their paper, researchers from the School of Informatics present a new large-scale dataset for training machine learning models to forecast national drought conditions, named DroughtED. The work's greatest potential impact comes from significantly lowering the bar (in available data and infrastructure) for getting predictions on drought, allowing less developed areas to capitalise on the data gathered in other regions. The paper formed part of the schedule for this year's International Conference on Machine Learning (ICML), one of the premier conferences on machine learning, which draws a wide audience of researchers and practitioners in academia, industry, and related fields. The workshop took place digitally on July 23rd, 2020, featuring 89 posters, 13 spotlight presentations, along with invited speakers and panels. Related links View Paper Pavlos Andreadis' personal page This article was published on 2024-03-18
