16-20 November 2020
Nagoya University KMI
Japan timezone
Breakout room #4: Detection and parameter estimation for GW-burst signal with machine learning
Presented by Yuto KIMURA
on
17 Nov 2020
from
14:20
to
15:50
Session:
Virtual Poster Session
Description
In recent years, machine learning(ML) has begun to be used to find out small signal of gravitational wave(GW) from noisy data and to estimate physical parameters of it. The approach is an alternative to using templates, by which the parameters are estimated by matching with theoretical models. George and Huerta (2018) demonstrated the ML method for GW from binary black holes. It is important to explore the possibility of detection and estimation for different types of GW signals. We consider the possibility of detection and parameter estimation for GW burst by ML. Since the wave forms, e. g., driven by magnetar giant flares, are uncertain at present, we model them and explore the ability of the ML approach. We use the same algorithms based by convolution neural network used in binary black hole merger by George and Huerta (2018). In this poster, we discuss accuracy of detection and how much error we can estimate parameters.
Place
Location: Nagoya University KMI
Room: Online