Machine Learning in Gravitational Wave search: g2net next challenges

Europe/Rome
Auditorium (European Gravitational Observatory)

Auditorium

European Gravitational Observatory

via E. Amaldi 5, 56021 S. Stefano a Macerata, Cascina (PI)
Elena Cuoco (EGO &SNS)
Description

g2net

The breakthrough discovery of Gravitational Waves (GWs) on September 14, 2015, was possible through the synergy of techniques drawing from expertise in physics, mathematics, information science and computing. The community nurtured by the CA17137 G2net COST Action is exploring, building on past work, the tremendous opportunity of the systematic application of Machine Learning (ML), Artificial Intelligence (AI) and Robotics to GW detection and Geophysics.

 

In this workshop we will show the results obtained so far by the working groups of the COST Action CA17137. Furthermore, we will discuss the state-of-the-art, as well as future challenges, of Machine Learning applied to gravitational wave research. The program will also include contributions from machine learning experts in different research areas. 

There is no fee for participation. The workshop will be held in hybrid format, with participation in presence at the European Gravitational Observatory and remote connection with Zoom.

Link for remote connection: Zoom link.

 

 

 

 

 

 

Registration
Participants
Participants
    • 10:00 11:00
      Welcome coffee and registration 1h
    • 11:00 11:30
      Welcome
      • 11:00
        Welcome from EGO Director 15m
        Speaker: Stavros Katsanevas
      • 11:15
        Welcome from Action Chair 15m
        Speaker: Elena Cuoco
    • 11:30 12:20
      Invited contributions
      Convener: Marco Cavaglia
      • 11:30
        Discovering gravitational waves with Machine Learning 50m
        Speaker: Marco Cavaglia
    • 12:20 13:00
      G2net result presentations
      • 12:20
        Searches for Mass-Asymmetric Compact Binary Coalescence Events using Neural Networks in the LIGO/Virgo Third Observation Period 20m
        Speaker: Mario Martinez
      • 12:40
        Deep Residual Networks for GW Astronomy 20m
        Speaker: Paraskevi Nousi (Aristotle University of Thessaloniki)
    • 13:00 14:15
      Lunch break 1h 15m
    • 14:15 15:00
      Invited contributions
      • 14:15
        What's happening in AI research? 45m

        The field of AI research finds itself at an interesting decision point: to exploit promising solutions that rose to prominence over the past few years and scale them up to more and more powerful generalist models, or to explore new unknown areas? Do we already have all the basic ingredients we need? 
This talk will describe some of the currently most popular architectures (Transformers, Perceivers, Graph Neural Nets, Diffusion, etc.), discuss the multimodal frontier, and highlight some success stories in technical application domains such as code generation, algorithmic reasoning, and the physical science, and explore results and open questions in the fundamental understanding of these models (e.g. scaling laws, interpretability, compression).
This overview aims to provide an introduction to exciting topics in AI research that may be relevant to this community, under the lens of a data-driven scaling approach.

        Speaker: Michela Paganini (DeepMind)
    • 15:00 16:00
      G2net result presentations
    • 16:00 16:15
      Coffee break 15m
    • 16:15 16:55
      G2net result presentations
    • 16:55 17:15
      Open Discussion
    • 09:10 10:40
      Contributed talks
    • 10:40 11:10
      Coffee break 30m
    • 11:10 11:30
      G2net result presentations
      • 11:10
        Outsourcing astrophysics data analysis to the real experts 20m
        Speaker: Joe Bayley (University of Glasgow)
    • 11:30 13:00
      Invited contributions
      Conveners: Denis Kanonik (Stealth Startup), Melissa Lopez Portilla
      • 11:30
        G2Net Gravitational Wave Detection Kaggle Competition Winner talk 45m
        Speaker: Denis Kanonik (Stealth Startup)
      • 12:15
        Towards the next generation of transient gravitational wave searches 45m

        The state-of-the-art of gravitational wave (GW) search techniques for transient signals have been extremely successful, but their sensitivity continues to be hindered by the presence of transient noise artifacts in the detectors, known as glitches. Glitches happen at a rate of 1 per min reducing the amount of scientific data available, as well as masking or mimicking GW signals. Therefore, there is a need for better modeling and inclusion of glitches, as well as improving the robustness of future GW searches. In this presentation we tackle two different challenges employing Machine Learning techniques: firstly we further analyze glitches populations with Generative Adversarial Networks, and secondly we ameliorate glitches in GW searches by analysing pipelines triggers with Gaussian Process classifier.

        Speaker: Melissa Lopez
    • 13:00 14:30
      Lunch break 1h 30m
    • 14:30 15:50
      G2net result presentations
    • 15:50 16:10
      Open Discussion
    • 16:10 16:30
      Coffee break 20m
    • 16:30 18:30
      Virgo tour
    • 18:30 19:30
      G2net Movie: Patterns
    • 19:30 21:30
      Evening event: Social Dinner
    • 09:10 10:10
      Contributed talks
      • 09:10
        A surrogate model for IMRIs in presence of dark matter spikes 20m
        Speaker: Teophanes Karydas (University of Amsterdam)
      • 09:30
        Machine learning for detection of dark matter spike around black holes through gravitational waves 20m
        Speaker: Alessandro Parisi (Scuola Normale Superiore di Pisa)
      • 09:50
        Deep Learning Gravitational-Wave Searches for Cosmic Strings 20m
        Speaker: Quirijn Meijer
    • 10:10 11:00
      Discussion and closing remarks
    • 11:00 11:20
      Coffee break 20m
    • 19:00 21:00
      Evening event: 2022 European Researchers’ Night