Scientists use artificial intelligence to detect gravitational waves

Scientific visualization of numerical relativity simulations explaining the collision of two black holes consistent with the merger of black hole binaries GW170814. The simulation was run on the Theta supercomputer using the open source Numerical Relativity community software Einstein Toolkit ( Credits: Argonne Leadership Computing Facility, Visualization and Data Analytics Group [Janet Knowles, Joseph Insley, Victor Mateevitsi, Silvio Rizzi].. )

When the gravitational wave was first detected by the Advanced Laser Interferometer Gravitational Wave Observatory (LIGO) in 2015, it confirmed another theory of Einstein and showed the birth of gravitational wave astronomy, causing ripples in the scientific world. It was. Five years later, a number of gravitational wave sources were detected, including the first observations of two neutron stars colliding with gravitational waves and electromagnetic waves.

LIGO and its international partners increase the sensitivity of the detector Gravitational waves, They will be able to scrutinize the larger volume of the universe, thereby causing daily detection of gravitational sources. The Flood of this discovery begins an era of precision astronomy that takes into account extrasolar messenger phenomena such as electromagnetic radiation, gravitational waves, neutrinos, and cosmic rays. However, achieving this goal requires a radical rethinking of the existing methods used to search and find gravitational waves.

Recently, he was a computational scientist, leader of Argonne National Laboratory’s translation artificial intelligence (AI) Eliu Huerta, and collaborators at Argonne, the University of Chicago, the University of Illinois at Urbana Champagne, and NVIDIA. We are cooperating. And IBM has developed a new production-scale AI framework that enables accelerated, scalable, and reproducible detection of gravitational waves.

This new framework shows that AI models can be as sensitive as traditional template matching algorithms, but they are orders of magnitude faster. In addition, these AI algorithms require only inexpensive graphics processing units (GPUs), such as those found in video game systems, to process advanced LIGO data faster than in real time.

The AI ​​ensemble used in this study processed the entire month of advanced LIGO data (August 2017) in less than 7 minutes and distributed the dataset to 64 NVIDIA V100 GPUs. The AI ​​ensemble used by the team for this analysis identified all four previously identified binary black hole merges in the dataset and did not report any misclassification.

“As a computer scientist, I’m excited about this project to show how to naturally integrate AI techniques with the right tools. Scientists’ workflows replace human intelligence. Instead, it allows you to work faster and better by augmenting it. “

Leveraging a variety of resources, this interdisciplinary, multi-institutional collaborator team published the following treatise: Nature Astronomy Introducing a data-driven approach that combines the collective supercomputing resources of the team to enable reproducible and accelerated AI-driven gravitational wave detection.

“This study used a combination of AI and supercomputing to solve timely and relevant big data experiments. See if AI can now offer new solutions to epic challenges. Not only does it make AI research perfectly reproducible, “Ferta said.

Based on the interdisciplinary nature of this project, the team is looking forward to a new application of this data-driven framework that goes beyond the challenges of big data in physics.

“This work highlights the important value of data infrastructure. Scientific communityBen Blaiszik, a research scientist at Argonne and the University of Chicago, said: “A set of building blocks with long-term investments by DOE, the National Science Foundation (NSF), the National Institute of Standards and Technology, etc. We will bring these building blocks together in a new and exciting way to extend this analysis and in the future. You can help provide these features to others. “

Huerta and his research team are working on a new framework through support for NSF, Argonne’s Institute-led Research and Development (LDRD) program, and an innovative and innovative computational impact (INCITE) program on DOE theory and experimentation. Was developed.

“These NSF investments include creative and innovative ideas with great potential to transform the way scientific data arrives in high-speed streams. Accelerated by planned activities. Heterogeneous computing technologies are being brought to many scientific practice communities, “says NSF’s Advanced Cyber ​​Infrastructure Office.

Scientists pioneered the use of deep learning for real-time gravitational wave discovery

For more information:
EA Huerta et al, accelerated, scalable and reproducible AI-driven gravitational wave detection, Nature Astronomy (2021). DOI: 10.1038 / s41550-021-01405-0

Quote: Scientists use artificial intelligence to obtain gravitational waves on July 12, 2021 from https: // (2021) July 7) will be detected

This document is subject to copyright. No part may be reproduced without written permission, except for fair transactions for personal investigation or research purposes. The content is provided for informational purposes only.

Scientists use artificial intelligence to detect gravitational waves

Source link Scientists use artificial intelligence to detect gravitational waves

Related Articles

Back to top button