And, um, one second, the term.

Okay. So for, um,

on the other hand, for core collapse of

massive star or instability in neutron star,

we don't know well the waveform as I told you before.

So we cannot use the waveform to extract the signal.

And so what we use are a model search.

So we search only for

excess in amplitude with respect to the background.

In many years, LIGO and Virgo developed

low latency pipeline to detect promptly

this signal using both type of analysis.

So model search, but also model search.

So we are able now after many years to

detect the trigger in really few minutes.

So in less than few minutes,

we know that there is some trigger that are

significant with respect to the background.

And we are able also to have a skylightization.

So when we have the detection,

what I'm sorry, one second that I have my children,

it is easier.

So.

Now.

Sorry. I'm really sorry. Okay. So when we have the detection, we send the alert to the

astronomer and then we send the retraction or confirmation after a few hours. Then we

send also to the astronomer some refined analysis, and in particular, so we make these, run these

parameter estimation code, and we are able to send to the astronomer some update. So one

of the problems of gravitational wave detection are the poor sky localization, and these require

a network of detector. So this is the first example of gravitational waves detected by

three interferometers, and what we use is the time delay in the three interferometers. And I show

you how much is important to have three detector because as you can see here, the sky localization

with three detector is really very reduced with respect to two detector. So the blue shadow in

this sphere is the, so the detection with two detector and the green part is the detection with

three detectors also with light, with Virgo. So what you see here is that we can reduce the sky

localization of a very big, big factor. So when we send to the astronomer the sky localization with

three detector, they can go deeper and they can make a lot of observation that are able also to

detect very faint sources. Okay, so now start the part on the detection of the electromagnetic

counterpart, and in particular, so we have this big sky localization, and what we need to use is

wide field telescope. Another big problem is that when you have, you look at the sky and the variable

sky, you have many, many transients in this survey. So in this big region of the sky, you have many,

many, many contaminants that you need to select, to exclude in order to have a sample of possible

candidate in which you point the larger telescope like the isotelescope, in which you want, you do

typically the spectroscopy, and so you are able to make the characterization to identify the

electromagnetic counterpart. So it's really very hard because you need to cover under the 2000

of square degree in the sky. You have to remove these contaminants, so you need to use, for example,

machine learning a lot to remove these contaminants. In 100 square degree, you have

a 1000 of transients. Many of them are artifacts, so you are able to remove them quickly,

but then you have a lot of also astrophysical transients.

And then you point the larger telescope, the follow-up and the spectroscopy, and then