Big Data and Artificial Intelligence Analytics in Geosciences: Promises and Potential

Abstract

Big data and machine learning are IT methodologies that are bringing substantial changes in the analysis and interpretation of scientific data. By adding GPU processing resources to the typical equipment of a server host, it is possible to speed up queries performed on large databases and reduce training time for deep learning architectures. A recent pairing of the big data technologies , applied to old and new data, and artificial intelligence techniques has enabled a team of scientists to create an interactive virtual globe that shows a color mosaic of the seabed geology. This interactive model allows us to obtain robust reconstructions and predictions of climate changes and their impacts on the ocean environment. We suggest a possible evolution of such a model by means of the expansion of functionalities and performance improvements. We refer respectively to the implementation of isochronic layers of seabed lithologies and the addition of GPU resources to speed up the learning phase of the support vector machine (SVM) model. These additional features would allow us to establish broader correlations and extract additional information on large-scale geological phenomena. INTRODUCTION The Earth system generates continuous data, and our acquisition capacity has significantly increased over time. The growing availability of acquired geological data and the methods developed in the field of information technology make it possible to identify associations and understand patterns and trends within data (Big Data), solve difficult decision problems (artificial intelligence), and provide acceleration to data processing (GPU computing). Big Data is a term that indicates very large databases (often by order of zettabytes, i.e., billions of terabytes) that