Machine Learning for Earth Observation with the Environmental Intelligence Network

Recent years have witnessed a dramatic increase in the acquisition of remote sensing observations from satellite, aircraft and drone-based sensors, and in-situ devices. Technological advances have led to improvements in measurement resolution and precision, which is shifting the paradigm of Earth observation from data scarcity to data abundance. While these data have enormous potential for helping us achieve a range of United Nations Sustainable Development Goals, identifying the optimal approaches for handling and analysing these large datasets remains a challenge for both academia and industry. Recent breakthroughs in AI/ML offers promising solutions to these challenges, including automated identification and extraction of key observations, predicting future trends, identifying key environment factors, and dealing with noisy signals under uncertainties.