Scientists from the University of Exeter have developed a sophisticated new approach in modeling and mapping Earth's natural environment. They created a new artificial intelligence (AI) technique that can recognize intricate terrain features that are beyond the capabilities of traditional methods.
The team says that this new technique can help generate better quality environmental maps and pave the way to unlocking discoveries that may help address climate and environmental issues in the 21st century.
Pioneering AI Technique Accurately Recognize Earth's Natural Features in Detail for Better Environmental Maps
Novel AI Technique Set in Bayesian Statistical Framework
Modeling and creating maps of the environment could be time-consuming, lengthy, and expensive. Usually, the number of observations obtained depends on the cost which impedes the creation of comprehensive and spatially-continuous maps that cannot be done using conventional modeling methods as they rely on users to manually engineer predictive features.
According to Phys.org, the team believes that there are more nuanced relationships at play within natural environments, so they developed an AI approach based on a Bayesian statistical framework that allows them to quantify uncertainties and provide a range of measures that will feed directly into the decision making processes.
The pioneering AI technique contains environmental information extraction to automatically recognize and make sense of the relationships that go unnoticed and unused by humans using conventional modeling methods.
The new AI technique aims to improve map quality and discover new relationships in the natural environment while eliminating huge amounts of trial-and-error experimentation in the modeling process.
Charlie Kirkwood, a postgraduate from the University of Exeter, said that AI serves a crucial role in providing and harnessing information more effectively than traditional approaches allow.
Calcium Provides Challenging Use for the AI
In a similar report from Science Daily, the researchers said the new AI approach was demonstrated using stream sediment calcium observations from the Geochemical Baseline Survey of the Environment (G-BASE) project of the British Geological Survey.
Both geological and hydrological processes control the distribution of calcium in the environment, which is important for soil fertility. Therefore, it provides a challenge for the AI approach because it must be taught how to recognize and utilize features relating to bedrock geology and surface hydrology.
Scientists said that despite only depicting calcium, the method produced a detailed and accurate map that reveals the geology of Britain in a new level of detail thanks to the new AI technique that is made possible by the cooperation of scientists from multiple institutes.
The work is the epitome of how integrating technical knowledge of AI and machine learning with geosciences could produce detailed maps of various environmental hazards that will provide a rich source of information for scientists and decision-makers.
Researchers detailed their new AI technique in the study, titled "Bayesian Deep Learning for Spatial Interpolation in the Presence of Auxiliary Information," published in the journal Mathematical Geosciences.
RELATED ARTICLE: Artificial Intelligence Used in Weather Forecast: How Does This Technology Affect Human Actions?
Check out more news and information on Artificial Intelligence in Science Times.