Contents 1. Introduction
Future high data volume NASA Earth science missions are driving unprecedented growth of archived data. The Earth Science Data Systems (ESDS) program needs to evolve to address numerous challenges that are presented by these missions. The evolution of the data system needs to consider managing each step of the full data lifecycle [1]. Furthermore, advances in data-driven technologies are driving user needs for additional data systems capabilities. One such technology is artificial intelligence (AI). AI refers to a class of algorithms that are capable of processing information and performing tasks that are comparable to human intelligence. Over the years, AI itself has evolved. AI transformation can be summarized as shown in Figure 1. AI accelerated hardware along with benchmark datasets fostered advanced AI algorithm development. Now, accessibility of these advanced algorithms is being democratized using cloud computing and open-source frameworks. The next transformation that we are seeing is the move towards service-centric and data-centric AI [2]. Data-centric AI approaches focus on making it easier for AI practitioners to iterate on datasets by improving characteristics of datasets instead of improving the models. These approaches combined with service-centric AI are foundations of taking AI and putting them into practice for real world applications.
