POSEIDON: POSEIDON: Physics-Optimized Seismic Energy Inference and Detection Operating Network

Boris Kriuk, Fedor Kriuk

Keywords: Physics-informed machine learning, Energy-based models, Seismology, Multi-task learning, Interpretable AI

TL;DR: We present POSEIDON, a physics-informed energy-based model and large-scale global dataset that jointly predict seismic events by learning interpretable seismological laws with state-of-the-art performance.

Abstract: Earthquake prediction and seismic hazard assessment remain fundamental challenges in geophysics, with existing machine learning approaches often operating as black boxes that ignore established physical laws. We introduce POSEIDON (Physics-Optimized Seismic Energy Inference and Detection Operating Network), a physics-informed energy-based model for unified multi-task seismic event prediction, alongside the Poseidon dataset---the largest open-source global earthquake catalog comprising 2.8 million events spanning 30 years. POSEIDON embeds fundamental seismological principles, including the Gutenberg-Richter magnitude-frequency relationship and Omori-Utsu aftershock decay law, as learnable constraints within an energy-based modeling framework. The architecture simultaneously addresses three interconnected prediction tasks: aftershock sequence identification, tsunami generation potential, and foreshock detection. Extensive experiments demonstrate that POSEIDON achieves state-of-the-art performance across all tasks, with the highest average F1 score among all compared methods. The learned physics parameters converge to scientifically interpretable values---Gutenberg-Richter b-value of 0.752 and Omori-Utsu parameters $p = 0.835$, $c = 0.1948$ days---falling within established seismological ranges. The Poseidon dataset is publicly available at \url{https://huggingface.co/datasets/BorisKriuk/Poseidon}.

Submission Number: 45