According to foreign media reports, Google DeepMind, in collaboration with the LIGO (Laser Interferometer Gravitational-Wave Observer) team and the Gran Sasso Scientific Institute (GSSI), has achieved a major breakthrough in Deep Loop Shaping technology, successfully resolving the low-frequency noise challenge that has plagued gravitational wave detection for years. The research, published in Science, marks a revolutionary application of AI in fundamental physics research.
The LIGO team, which won the 2017 Nobel Prize in Physics for its gravitational wave observations, had struggled to overcome noise in its detectors in the low-frequency range of 10-30 Hz. Using AI, researchers have reduced noise intensity to 1/30th of that of traditional methods, and in some sub-bands to 1/100th of the original, even surpassing the theoretical design target of the quantum limit. This breakthrough opens the possibility of detecting even fainter cosmic signals, such as the early stages of black hole mergers.
The core of the technology lies in AI-powered real-time dynamic analysis of interferometer data. The DeepMind-developed algorithm accurately identifies and cancels out interference factors such as environmental vibrations and thermal noise, whereas traditional physical noise reduction methods are limited by response speed and computational complexity. The partners said that the technology has entered the actual testing stage and may be applied to the next generation of gravitational wave detectors (such as the LISA space project) in the future, further expanding the boundaries of human understanding of the universe.
