DeepExtractor: Time-domain reconstruction of signals and glitches in gravitational wave data with deep learning

2025-01-30

AI Summary

Distinguishing genuine astrophysical gravitational-wave signals from transient non-Gaussian noise artifacts (glitches) in detector data remains a critical challenge. Method: We propose DeepExtractor, an end-to-end deep learning framework that decouples noise modeling from residual signal reconstruction—departing from template-based or physics-informed waveform mapping paradigms. Leveraging the Gaussian stationary noise assumption, it integrates time-domain convolutional architectures with residual learning and enables real-time inference on CPU. Contribution/Results: On simulated data, DeepExtractor achieves a median mismatch of only 0.9% in glitch reconstruction. It outperforms BayesWave in accuracy and accelerates per-event processing by 3600× (0.1 s vs. 1 h), while demonstrating robustness on real LIGO data. Its core innovation lies in establishing a model-free, highly generalizable, high-fidelity, and real-time deployable paradigm for joint time-domain reconstruction of gravitational-wave signals and glitches.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Natural Language Processing: Information Extraction
• Intelligent Robots: State Estimation

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Gravitational wave (GW) interferometers, detect faint signals from distant astrophysical events, such as binary black hole mergers. However, their high sensitivity also makes them susceptible to background noise, which can obscure these signals. This noise often includes transient artifacts called"glitches"that can mimic astrophysical signals or mask their characteristics. Fast and accurate reconstruction of both signals and glitches is crucial for reliable scientific inference. In this study, we present DeepExtractor, a deep learning framework designed to reconstruct signals and glitches with power exceeding interferometer noise, regardless of their source. We design DeepExtractor to model the inherent noise distribution of GW interferometers, following conventional assumptions that the noise is Gaussian and stationary over short time scales. It operates by predicting and subtracting the noise component of the data, retaining only the clean reconstruction. Our approach achieves superior generalization capabilities for arbitrary signals and glitches compared to methods that directly map inputs to the clean training waveforms. We validate DeepExtractor's effectiveness through three experiments: (1) reconstructing simulated glitches injected into simulated detector noise, (2) comparing performance with the state-of-the-art BayesWave algorithm, and (3) analyzing real data from the Gravity Spy dataset to demonstrate effective glitch subtraction from LIGO strain data. DeepExtractor achieves a median mismatch of only 0.9% for simulated glitches, outperforming several deep learning baselines. Additionally, DeepExtractor surpasses BayesWave in glitch recovery, offering a dramatic computational speedup by reconstructing one glitch sample in approx. 0.1 seconds on a CPU, compared to BayesWave's processing time of approx. one hour per glitch.

Problem

• Gravitational Wave Data
• Astronomical Event Detection
• Noise Discrimination

Innovation

• DeepExtractor
• Gravitational Wave Data Processing
• Noise Reduction