Detecting continuous gravitational-wave signals: a Kaggle competition

More info:  external link

As a collaboration between the University of Glasgow and the University of the Balearic Islands, with the support of the G2Net COST Action CA17137, we have launched a Kaggle competition for the detection of continuous gravitational waves. In this competition, participants are asked to develop a machine-learning-based solution to identify data samples in which a simulated continuous wave is present. The competition lasts for three months (October 2022 – January 2023) and there is a prize of $25,000 generously donated by Google to be split among the three winners of the competition.

Continuous gravitational waves are a long-lasting form of gravitational radiation, yet to be detected by the current generation of interferometric detectors operated by the LIGO-Virgo-KAGRA collaboration. They are expected to be emitted by non-axisymmetric rapidly-spinning neutron stars and they would allow us to probe the physics of such compact objects, as well as accessing the electromagnetically-quiet population of galactic neutron stars, otherwise inaccessible. Their detection would constitute the next major milestone of gravitational-wave astronomy.

Current continuous-wave searches are computationally limited, as the most optimal methods incur a computational cost orders of magnitude greater than typical budgets nowadays available. Latest advancements in machine learning, on the other hand, suggest that other, more efficient methods, could be used to develop more sensitive searches.

Kaggle is an open online platform associated with Google which hosts data-analysis competitions for different entities. Any individual can take part in this competition regardless of their background at no cost. Upon completion, the source code of winning solutions should be disclosed in order to be eligible for a monetary prize.

Join the competition and find all the info at:

Rodrigo Tenorio

Michael Williams

Chris Messenger