The rapid advancements in machine learning (ML) offer transformative potential for improving the prediction and analysis of Earth Orientation Parameters (EOP), which are critical in the application of geodesy, precise navigation, and space missions. To harness this potential, we established the Joint Study Group 3 under GGOS and IAG Commission 3, titled Artificial Intelligence for Earth Orientation Parameter Prediction (AI4EOP Prediction).
This Study group is a collaborative platform for researchers, data scientists, and domain experts to explore innovative ML techniques, share datasets, and develop robust models for EOP prediction and anomaly detection. By integrating ML with traditional geophysical approaches, we aim to enhance prediction accuracy, address current modeling limitations, and foster interdisciplinary collaboration. We invite experts and enthusiasts from diverse backgrounds to join us in advancing the field and shaping the future of EOP research through ML-driven solutions. This presentation provides a summary of the AI4EOP Prediction Study Group's activities to date, as well as plans for future initiatives. We also present the first results of a special internal EOP prediction comparison sub-campaign, which is part of the Second Earth Orientation Parameters Prediction Comparison Campaign - EOP PML. The goal of EOP PML is to further explore the potential of ML in EOP prediction through the regular collection, comparison, and evaluation of EOP predictions developed using various ML approaches. EOP PML follows strict guidelines for input data, forecasting horizons, and evaluation methods to objectively assess ML models. This initiative provides an excellent opportunity for experts in geodesy, AI, and Earth sciences to collaborate and refine EOP prediction methodologies. The EOP PML sub-campaign is open to all interested participants, and new contributors are welcome to join at any time.

Keywords: EOP, Prediction, EOP PML, AI4EOP