Transformer networks for continuous gravitational-wave searches
Wide-parameter-space searches for continuous gravitational waves using semicoherent matched-filter methods require enormous computing power, which limits their achievable sensitivity. Here we explore an alternative search method based on training neural networks as classifiers on detector strain data with minimal preprocessing. Contrary to previous studies using convolutional neural networks (CNNs), we investigate the suitability of the transformer architecture, specifically the vision transformer (ViT). We establish sensitivity benchmarks using the matched-filter <inline-formula><mml:math><mml:mi>F</mml:mi></mml:math></inline-formula>-statistic for ten targeted searches over a ten day timespan, and ten directed and six all-sky searches over a one day timespan. We train ViTs on each of these benchmark cases. The trained ViTs achieve essentially matched-filter sensitivity on the targeted benchmarks, and approach the <inline-formula><mml:math><mml:mi>F</mml:mi></mml:math></inline-formula>-statistic detection probability of <inline-formula><mml:math><mml:mrow><mml:msub><mml:mrow><mml:mi>p</mml:mi></mml:mrow><mml:mrow><mml:mi>det</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mn>90</mml:mn><mml:mo>%</mml:mo></mml:mrow></mml:math></inline-formula> on the directed (<inline-formula><mml:math><mml:mrow><mml:msub><mml:mrow><mml:mi>p</mml:mi></mml:mrow><mml:mrow><mml:mi>det</mml:mi></mml:mrow></mml:msub><mml:mo>≈</mml:mo><mml:mn>85</mml:mn><mml:mi>--</mml:mi><mml:mn>89</mml:mn><mml:mo>%</mml:mo></mml:mrow></mml:math></inline-formula>) and all-sky benchmarks (<inline-formula><mml:math><mml:mrow><mml:msub><mml:mrow><mml:mi>p</mml:mi></mml:mrow><mml:mrow><mml:mi>det</mml:mi></mml:mrow></mml:msub><mml:mo>≈</mml:mo><mml:mn>78</mml:mn><mml:mi>--</mml:mi><mml:mn>88</mml:mn><mml:mo>%</mml:mo></mml:mrow></mml:math></inline-formula>). Unlike the CNNs in our previous studies, which required extensive manual design and hyperparameter tuning, the ViT achieves better performance with a standard architecture and minimal tuning.
展开 ▾用于连续引力波搜索的Transformer网络 · 该论文探索用视觉Transformer(ViT)直接分类探测器应变数据以搜索连续引力波,在目标、定向和全天基准上训练后,其灵敏度接近匹配滤波F统计量,且比CNN所需人工设计与调参更少。
预印本 2025-09-13 · 刊出 2025-12-29 · 收录 2026-10-12