Artificial Intelligence (AI) can considerably get better our understanding of the climate and Earth science, says a study by German scientists.
Artificial intelligence may be useful to data related to great events such as fire spreads or hurricane, which are very difficult processes predisposed by local conditions.
It can also be applied to atmospheric and deep-sea transport, soil movement, and vegetation dynamics data – some of the classic topic of Earth system science.
“as of a plethora of sensors, a deluge of Earth system data has to be converted into obtainable, but so far we’ve been covering behind in analysis and interpretation,” thought Markus Reichstein of the Max Planck institution for Biogeochemistry in Jena, Germany.
“Deep learning technique turns out to be a promising tool, beyond classical machine learning application, for example, image recognition, natural language processing or Alpha Go,” added co-author Joachim Denzler, from the Friedrich Schiller University in Jena (FSU).
Though, deep learning approaches are difficult. All data-driven and statistical approach do not warranty physical constancy per se, Artificial intelligence is extremely dependent on data quality, and may experience difficulties with extrapolations, as per a study published in the journal Nature.
Moreover, the requirement for data processing and storage ability is very high.
If both techniques are bringing together, so-called hybrid models are formed. They can, for example, be used for modeling the movement of ocean water to predict sea surface temperature. As the temperatures are model physically, the ocean water group is represented by a machine learning approach.
“The idea is to mingle the better of two worlds, the reliability of physical models with the adaptability of machine learning, to obtain to a great extent improved models,” Reichstein described.
To read out more similar content please stay tuned to igreenik.com