Training Strategies for Deep Learning Gravitational-Wave Searches
This work addresses computational bottlenecks in gravitational-wave detection for astrophysicists, but it is incremental as it focuses on optimizing training strategies for an existing deep learning approach.
The study tackled the problem of detecting gravitational-wave signals from binary black holes using deep learning, finding that training with low signal-to-noise ratio (SNR) signals leads to faster convergence and that a modification called unbounded Softmax replacement (USR) retains at least 91.5% sensitivity compared to matched-filter searches at low false-alarm rates.
Compact binary systems emit gravitational radiation which is potentially detectable by current Earth bound detectors. Extracting these signals from the instruments' background noise is a complex problem and the computational cost of most current searches depends on the complexity of the source model. Deep learning may be capable of finding signals where current algorithms hit computational limits. Here we restrict our analysis to signals from non-spinning binary black holes and systematically test different strategies by which training data is presented to the networks. To assess the impact of the training strategies, we re-analyze the first published networks and directly compare them to an equivalent matched-filter search. We find that the deep learning algorithms can generalize low signal-to-noise ratio (SNR) signals to high SNR ones but not vice versa. As such, it is not beneficial to provide high SNR signals during training, and fastest convergence is achieved when low SNR samples are provided early on. During testing we found that the networks are sometimes unable to recover any signals when a false alarm probability $<10^{-3}$ is required. We resolve this restriction by applying a modification we call unbounded Softmax replacement (USR) after training. With this alteration we find that the machine learning search retains $\geq 91.5\%$ of the sensitivity of the matched-filter search down to a false-alarm rate of 1 per month.