This session highlights recent advances in interdisciplinary research on pre-earthquake processes that combine space- and ground-based observations, physical modeling, and artificial intelligence (AI). Diverse observables, such as deformation (GPS, InSAR), electromagnetic, geochemical, hydrogeological, and ionospheric changes, have been linked to stress evolution in the lithosphere before large earthquakes. The emergence of large language models (LLMs), generative AI, and multi-modal learning frameworks is transforming this field. These tools enable fusion of heterogeneous datasets, generation of synthetic precursors for model testing, and real-time interpretation of complex signals through in-context and physics-informed learning. We invite contributions on observational and modeling results; machine learning and AI-enhanced methods for earthquake forecasting and nowcasting; multi-modal data integration; explainable and physics-informed AI; and advanced sensing and computational technologies. The session aims to connect geophysical, geochemical, and computational communities to accelerate progress toward more reliable, interpretable, and physically grounded earthquake forecasting systems.

This session highlights recent advances in interdisciplinary research on pre-earthquake processes that combine space- and ground-based observations, physical modeling, and artificial intelligence (AI). Diverse observables, such as deformation (GPS, InSAR), electromagnetic, geochemical, hydrogeological, and ionospheric changes, have been linked to stress evolution in the lithosphere before large earthquakes. The emergence of large language models (LLMs), generative AI, and multi-modal learning frameworks is transforming this field. These tools enable fusion of heterogeneous datasets, generation of synthetic precursors for model testing, and real-time interpretation of complex signals through in-context and physics-informed learning. We invite contributions on observational and modeling results; machine learning and AI-enhanced methods for earthquake forecasting and nowcasting; multi-modal data integration; explainable and physics-informed AI; and advanced sensing and computational technologies. The session aims to connect geophysical, geochemical, and computational communities to accelerate progress toward more reliable, interpretable, and physically grounded earthquake forecasting systems.