Erlangen 2026 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
GR: Fachverband Gravitation, Relativistische Astrophysik und Kosmologie
GR 9: Black Holes II
GR 9.5: Vortrag
Mittwoch, 18. März 2026, 14:45–15:00, KH 02.012
Forecasting Ground Motion with Large Sequence Models: Implications for the Einstein Telescope — •Waleed Esmail1, Alexander Kappes1, Stuart Russell2, and Christine Thomas2 — 1Institut für Kernphysik, Universität Münster, Wilhelm-Klemm-Straße 9, 48149, Münster — 2Institut für Geophysik, Universität Münster, Wilhelm-Klemm-Straße 9, 48149, Münster
Seismic forecasting is entering a new phase in which data-driven sequence models can learn the complex temporal structure of ground motion directly from waveform records. In this talk, I will discuss a transformer-based approach that treats seismic waveforms as a generative sequence prediction problem, allowing short-horizon forecasting. A key motivation comes from the Einstein Telescope, where seismic noise, especially at low frequencies, can limit detector sensitivity. Accurate, low-latency forecasting enables active noise cancellation, where the detector can anticipate incoming noise and suppress it before it couples into the interferometer. I will show how transformer-based models can support this vision and why a seismic foundation model may become a core component of the next generation of gravitational-wave observatories.
Keywords: Einstein Telescope; Seismic Noise; Deep Learning; Gravitational Waves Detection