Within the European Project InterTwin  we are about to open a position  (Assegno di Ricerca) for 1+1 years on the topic:

Characterisation of detector noise for Gravitational Wave experiments using Generative Adversarial Networks. 

The sensitivity of GW interferometers is limited by noise. We will use Generative Adversarial Networks (GANs) to produce a Digital Twin (DT) of the Virgo interferometer to realistically simulate transient noise (glitches) in the detector. In a first phase, we will use the GAN-based DT to generate synthetic strain data (the physical signal channel) from auxiliary channels only. Auxiliary channels collect data from a variety of instrumental and environmental sensors and are not sensitive to the astrophysical signal, so they allow us to characterise noise. In a second phase, also in the perspective of the Einstein Telescope, we will use the trained model to characterise glitches and optimise the use of auxiliary channels in vetoing and de-noising the signal in low-latency (few seconds) searches. This will allow the low-latency searches  to send out more reliable triggers to observatories for multi-messenger astronomy.

We are looking for a graduate candidate with strong enthusiasm for Machine Learning techniques and possibly some basic knowledge of Python.

Contact: sara.vallero@to.infn.it