CASPER is a ResNet-based classifier with FastSHAP explainer that detects gravitational waves in real detector noise at 91% AUC using 260 unaugmented events and recovers chirp morphology in attribution maps.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
gr-qc 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
CASPER: Interpretable ResNet based Classifier with FastShap Explainer for Gravitational Wave Detection
CASPER is a ResNet-based classifier with FastSHAP explainer that detects gravitational waves in real detector noise at 91% AUC using 260 unaugmented events and recovers chirp morphology in attribution maps.