JEDDAH: King Abdullah University of Science and Technology, and SARsatX, a Saudi company specializing in Earth observation technologies, have developed computer-generated data to train deep learning models to predict oil spills.

According to KAUST, validating the use of synthetic data is crucial for monitoring environmental disasters, as early detection and rapid response can significantly reduce the risks of environmental damage.

Matthew McCabe, dean of the Biological and Environmental Science and Engineering Division at KAUST, said one of the biggest challenges in environmental applications of artificial intelligence was the shortage of high-quality training data.

He explained that this challenge can be addressed by using deep learning to generate synthetic data from a very small sample of real data and then training predictive AI models on it.

This approach can significantly enhance efforts to protect the marine environment by enabling faster and more reliable monitoring of oil spills while reducing the logistical and environmental challenges associated with data collection.