Deep learning approach for EMRI detection achieving 94.2% TPR at 1% FPR, demonstrating the potential for efficient signal detection in space-based gravitational wave detectors.
(This article belongs to the [Special Issue Newest Results in Gravitational Waves and Machine Learning](https://www.mdpi.com/journal/universe/special_issues/48U1E55JLC))
Science-driven multi-stage deep neural network for space-based gravitational wave detection and extraction achieving >99% accuracy and ≥95% signal similarity, with strong generalization and interpretability demonstrated on synthetic LISA data.