Google DeepMind AI boosts gravitational wave detection by 100x

Google DeepMind has developed Deep Loop Shaping, a novel AI method that reduces noise in gravitational wave observatories by 30 to 100 times, significantly improving their ability to detect cosmic events. Published in Science and tested at LIGO’s Louisiana facility, this breakthrough could help astronomers detect hundreds more black hole mergers and neutron star collisions annually, potentially revealing intermediate-mass black holes that represent the “missing link” in understanding galaxy evolution.

Why this matters: LIGO’s extraordinary sensitivity—measuring distances 1/10,000th the size of a proton—makes it vulnerable to any vibration, even ocean waves 100 miles away, limiting its ability to peer deeper into the universe.

How it works: Deep Loop Shaping uses reinforcement learning with frequency domain rewards to train controllers that suppress vibrations without amplifying harmful noise.
• The AI system learns through repeated interaction to stabilize LIGO’s mirrors while keeping control noise below quantum fluctuation levels.
• Unlike traditional linear control methods, the AI approach eliminates the controller itself as a meaningful source of noise.
• Testing showed the method works equally well in simulation and on actual LIGO hardware over prolonged periods.

The technical challenge: LIGO’s detector mirrors must remain extraordinarily still to measure gravitational waves, requiring a delicate balance in control systems.
• Too little control causes mirrors to swing wildly, making measurements impossible.
• Too much control amplifies vibrations instead of suppressing them, drowning out signals in critical frequency ranges.
• This “control noise” has been a major barrier to improving LIGO’s cosmic reach.

Real-world impact: The method successfully eliminated the most unstable feedback loop at LIGO as a meaningful noise source for the first time.
• Applying Deep Loop Shaping to all of LIGO’s mirror control loops could enable detection of hundreds more cosmic events per year.
• The improved sensitivity particularly benefits detection of intermediate-mass black holes, which astronomers have rarely observed.
• Enhanced measurements could provide far greater detail about neutron star collisions, black hole formation, and heavy element creation.

What they’re saying: “Studying the universe using gravity instead of light, is like listening instead of looking. This work allows us to tune in to the bass,” said Rana Adhikari, Professor of Physics at Caltech.

Beyond astronomy: Deep Loop Shaping’s applications extend far beyond gravitational wave detection.
• The method could address vibration suppression and noise cancellation problems in aerospace engineering.
• Robotics and structural engineering applications involving highly dynamic or unstable systems could benefit.
• Future space-based gravitational wave observatories may incorporate these AI control methods from the design phase.

The collaboration: DeepMind developed the method with LIGO, operated by Caltech, and Gran Sasso Science Institute, proving the approach at the Livingston, Louisiana observatory.