Speaker
Nim Ki Wong
(National Yang Ming Chiao Tung University)
Description
We present an automated noise subtraction pipeline combining Coherence Monitor (CohMon) and DeepClean for KAGRA’s O4a data. CohMon evaluates PEM sensors via frequency-domain coherence to select optimal witness channels. These feed into DeepClean, a 1D CNN, to model and subtract non-linear noise. Applied to O4a strain, it removes artifacts without altering astrophysical signals. Evaluated by ASD ratios, this provides a automated denoising framework for KAGRA.
| Participate the oral/poster presentation award competition | Yes |
|---|