RUI: improving the ligo data analysis pipeline from calibration to parameter estimation with streaming frameworks, AI tools, and sophisticated inference methods

The National Science Foundation Division of Physics awarded Kenyon College $200,000 on August 1, 2026, under the Mathematical and Physical Sciences program (CFDA 47.049) to improve the LIGO data analysis pipeline for gravitational-wave detection and parameter estimation. The award funds research spanning three primary areas of the gravitational-wave data analysis pipeline: calibration of detector output into accurately calibrated data, distinguishing true astrophysical signals from instrumental noise, and extracting properties of ultra-dense nuclear matter from detected signals. Research will employ streaming frameworks, AI tools, and sophisticated inference methods to strengthen the process by which raw detector data are converted into scientific results.

Kenyon College will conduct the work in Gambier, Ohio, with performance through July 31, 2028. The award supports research experiences for undergraduate students who will join the LIGO international scientific collaboration and contribute directly to its science. Funding also supports the Radio and Optical Astronomy Research (ROAR) group, which engages first- and second-year undergraduates in astrophysics to promote retention in the physics major, and educational outreach through the Mount Vernon High School Science Club, which delivers hands-on physics, astronomy, and chemistry activities to students in rural central Ohio.

Abstract

Ten years ago, the first direct detection of gravitational waves from two colliding black holes opened an entirely new way of observing the universe. Since then, NSF's laser interferometer gravitational-wave observatory (LIGO) and its international partners have detected hundreds of collisions of black holes and neutron STARS, revealing new populations of black holes and providing unprecedented views into the interiors of neutron STARS, the densest known objects in the universe. The next observing run of these detectors promises the best sensitivity to date and, with IT, discoveries. This award supports research at kenyon college that strengthens nearly every stage of the process by which raw detector data are turned into scientific results: converting detector output into accurately calibrated data, distinguishing true astrophysical signals from instrumental noise, and extracting the properties of ultra-dense nuclear matter from detected signals. This work advances the national interest by promoting the progress of science, maintaining united states leadership in gravitational-wave astrophysics, and developing a highly skilled stem workforce. The award funds research experiences for undergraduate students at kenyon college, who join an internationally renowned scientific collaboration and contribute directly to its science. IT also supports the radio and optical astronomy research (ROAR) group, which engages first- and second-year undergraduates in astrophysics with an eye toward retention in the physics major, and the mount vernon high school science club, which brings hands-on physics, astronomy, and chemistry activities to students in rural central ohio. This award supports research in three primary research areas spanning the gravitational-wave data analysis pipeline. The first area is calibration of the LIGO detectors, including operation of the low-latency calibration pipeline, integration of alerting and additional metrics into the real-time calibration monitoring system, development of a machine-learning method that estimates broadband calibration systematic error from narrow-band photon calibrator measurements, and application of real-time corrections to compensate for measured systematic error in the low-latency strain data. The second area involves improvements to the low-latency data quality identification system known as IDQ, which uses machine learning algorithms to predict the presence of transient noise events, or glitches, in the strain data using only auxiliary channel information. The group will develop a two-stage ensemble neural network trained on specific glitch morphologies and evaluate its impact on the sensitivity of modeled searches for compact binary systems; in particular, blip glitches pose a significant challenge for intermediate mass black hole binary (IMBHB) searches, and a classifier trained to identify blip glitches could enhance the performance of these searches. The third area is inference of the neutron star equation of state (EOS) from binary neutron star mergers. The group will upgrade the robustness and sophistication of the current parameterized EOS inference framework, including implementation of an overhauled tolman-oppenheimer-volkoff solver capable of handling multi-branch EOS models with sharp phase transitions, compatible upgrades to the bilby parameter estimation software with sampling in component central pressures, and implementation of a reversible jump markov chain monte carlo method that allows the data to determine the appropriate number of EOS model parameters. The group will also contribute to EOS inference efforts on any binary neutron star signals detected during the fifth observing run. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the foundation's intellectual merit and broader impacts review criteria.- subawards are not planned for this award.