g2net WG3 training school on Machine Learning for Advanced Control Techniques

Andrea Chincarini (INFN Genova), Luigia Petre (Åbo Akademi University)

WG3 explores the use of Machine Learning (ML) techniques in the control and noise mitigation strategies of scientific experiments, specifically for Gravitational Wave (GW) detectors. GW detectors, both those currently running and those foreseen to be spaceborne, are uniquely complex instruments with specific and new challenges in terms of control and noise issues. These challenges call for significant adaptation and ingenuity in the ML approaches, which are seldom used as textbook cases and are often coupled with simulations and burden with heavy experimental constraints. These developments need diverse expertise and interaction, which is the benefit of the current COST action. This working group’s goal is to develop ML algorithms as part of the detectors’ feedback-control systems as well as for the feed-forward cancellation of noise.

In the training school we address the following topics:

  • Einstein Telescope
  • LISA
  • Deep learning taxonomy
  • Newtonian noise cancellation with machine learning
  • Fractal analysis for controlling interferometers
  • Robotics for interferometers

The workshop is organised during August 30-September 2, 2021, ONLINE, see below for the zoom coordinates. 

The following speakers have kindly agreed to join us:

  • Dr Francesca Badaracco, UCLouvain, Belgium: "Newtonian Noise Cancellation Strategies and Optimisation Problems"
  • Prof Marco Cavaglia, Missouri Univeristy of Science and Technology, US: "Fractal analysis for interferometer control"
  • Dr Razvan Pascanu, DeepMind, UK: "Deep learning Taxonomy"
  • Dr Natalia Korsakova, Artemis, France : "LISA talk on disentanglement"
  • Dr Fabio Bonsignorio, HeronRobotics: "Robotics for interferometers"
  • Dr Luca Naticchioni, INFN, Italy: "Einstein Telescope site characterisation measures and their impact on the third generation GW detectors"
  • Prof Stefan Hild, Maastricht University, The Netherlands: "Einstein Telescope: technical & scientific challenges for the future GW detectors"

These speakers will also oversee a round table discussion on Thursday, September 2, mainly on investigating the machine learning potential in their topics, but also discussing other topics of interest.

Please contact either/both Luigia and Andrea (lpetre@abo.fi, andrea.chincarini@ge.infn.it) if you have questions.


Luigia Petre is inviting you to a scheduled Zoom meeting.

Topic: WG3's Training School, August 30- September 2, 2021
Time: This is a recurring meeting Meet anytime

Join Zoom Meeting

Meeting ID: 636 3418 8394

Join by SIP

Join by H.323
Meeting ID: 636 3418 8394

  • Agata Trovato
  • Akanksha Ahuja
  • Andrei Utina
  • Arash Gavgani
  • Argyro Sasli
  • Chaganti Rama Durga Santosh Kumar
  • Conor Muldoon
  • Dimitrios Pesios
  • Elahe Khalouei
  • Fabio bonsignorio
  • Filip Morawski
  • Francesca Badaracco
  • Harish Singh
  • Hector Hortua
  • Ian Baldwin
  • Isabel Cordero-Carrión
  • Jacopo Soldateschi
  • Lavinia Longo
  • Luca Naticchioni
  • Luigia Petre
  • Marco Cavaglia
  • Marta Budroni
  • Mateusz Bawaj
  • Márcio Ferreira
  • Natalia Korsakova
  • Nikolaos Karnesis
  • Panagiotis Iosif
  • Paula Pons Muñoz
  • prateek upadhyay
  • Rafel Jaume Amengual
  • Razvan Pascanu
  • Sepideh Ghaziasgar
  • Shruti Sharma
  • Tomislav Andric
  • tushar vaidya
  • Velimir Ilic
  • Vijayasri Iyer
  • Wathela Alhassan
  • Xianjia Yu
    • 12:00 PM 1:00 PM
      Lunch 1h
    • 1:00 PM 1:30 PM
      Opening 30m
      Speakers: Elena Cuoco, Andrea Chincarini, Luigia Petre (Åbo Akademi University)
    • 1:30 PM 3:30 PM
      Newtonian Noise Cancellation Strategies and Optimisation Problems 2h

      The recording of Francesca's presentation can be found here:

      Speaker: Dr Francesca Badaracco (UCLouvain)
    • 3:30 PM 4:00 PM
      Coffee break 30m
    • 4:00 PM 6:00 PM
      Einstein Telescope site characterisation measures and their impact on the third generation GW detectors 2h

      Presenting the telescope and problematic. The recording of Luca's lecture is here:

      Speaker: Dr Luca Naticchioni
    • 9:00 AM 11:00 AM
      Einstein Telescope (ET): technical & scientific challenges for the future GW detectors 2h

      What data analysis is done and how we could employ ML. The recording of Stefan's presentation follows below:

      Speaker: Prof. Stefan Hild (Maastricht University)
    • 11:00 AM 11:30 AM
      Coffee break 30m
    • 11:30 AM 1:00 PM
      Deep learning taxonomy - part 1 1h 30m

      The recording of Razvan's talk follows below:

      Speaker: Dr Razvan Pascanu (DeepMind)
    • 1:00 PM 2:00 PM
      Lunch 1h
    • 2:00 PM 2:30 PM
      Deep learning taxonomy - part 2 30m
      Speaker: Dr Razvan Pascanu (DeepMind)
    • 2:30 PM 3:30 PM
      Fractal analysis for interferometer control - part 1 1h

      Slides, part 1: https://drive.google.com/file/d/1M0MBBTrjLDbu_zcZBO7pIiC9SuJ7AlF6/view

      Slides, part 2: https://drive.google.com/file/d/1eRlKqzAM7XntED9AHv8HnHx_UO8jQF2C/view'

      The recording of Marco's lecture follows below:

      Speaker: Prof. Marco Cavaglia (Missouri University of Science and Technology)
    • 3:30 PM 4:00 PM
      Coffee break 30m
    • 4:00 PM 5:00 PM
      Fractal analysis for interferometer control - part 2 1h

      Instructions for hands-on activity here:


      Speaker: Prof. Marco Cavaglia (Missouri University of Science and Technology)
    • 9:00 AM 11:00 AM
      LISA talk on disentanglement of GW 2h

      We will detect tens/hundreds GWs per day and need to determine if more than one GW occurs at the same time; opportunity for multi messenger cooperation.

      Here are links to the two google colab notebooks that Natalia created for practicing the concepts from her lecture:



      The recording of Natalia's presentation follows below:

      Speaker: Dr Natalia Korsakova (SYRTE/Paris Observatory)
    • 11:00 AM 11:30 AM
      Coffee break 30m
    • 11:30 AM 1:00 PM
      Possible Utilization of Intelligent Robotics Technologies in GW detection and in the Einstein Telescope - part 1 1h 30m

      Fabio's slides can be found here:

      The recording of Fabio's presentation follow bellow:

      Speaker: Fabio Bonsignorio
    • 1:00 PM 2:00 PM
      Lunch 1h
    • 2:00 PM 2:30 PM
      Possible Utilization of Intelligent Robotics Technologies in GW detection and in the Einstein Telescope - part 2 30m

      For the participants

      1) (needed) Sign up on this website:https://robotbenchmark.net
      2) (nice to have) You can download webots from here: https://cyberbotics.com/doc/guide/installing-webots
      and install python3 if they didn't already
      3) (nice to have) I will walk the participants through a real world research ROS-Gazebo installation from scratch; you can do the same on your computer; you will need an Ubuntu 18.04 LTS OS either on their computer or in a virtual machine (not recommended). Virtual machine software can be downloaded from here https://www.virtualbox.org/wiki/Downloads. Ubuntu 18.04 LTS can be downloaded from here: https://ubuntu.com/download/alternative-downloads

      Speaker: Fabio Bonsignorio
    • 9:00 AM 12:00 PM
      Roundtable discussion: ML for advanced control 3h
      Speakers: Fabio Bonsignorio (HeronRobotics), Francesca Badaracco (UCLouvain), Luca Naticchioni (INFN), Prof. Marco Cavaglia (Missouri University of Science and Technology), Dr Natalia Korsakova (SYRTE/Paris Observatory), Dr Razvan Pascanu (DeepMind), Prof. Stefan Hild
    • 12:00 PM 1:00 PM
      Lunch 1h