Reproduce “Towards Label-Efficient Single-Station Seismic Sensing Using Hybrid Self-Supervised Learning” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Ashutosh Kumar Sinha, Himanshu Raj, Rahul Mishra
SeisLoc-FM estimates earthquake back-azimuth, epicentral distance, and P-arrival time from a single station’s three-component waveform. It combines waveform reconstruction, short-time Fourier transform prediction, and cross-station event-invariant learning before fine-tuning with labels. On INSTANCE at 1% labeled events, its three errors are 25.2°, 9.0 km, and 0.439 s, versus 49.5°, 16.0 km, and 0.679 s with masked-only pretraining; this is an in-dataset result.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.