Searches for stochastic gravitational waves with LIGO/Virgo. Stochastic gravitational-wave backgrounds can be created in the early universe from amplification of vacuum fluctuations following inflation, phase transitions in the early universe, cosmic strings and pre-Big Bang models. Stochastic gravitational-wave foregrounds, meanwhile, can be created from the superposition of astrophysical sources such as core-collapse supernovae, protoneutron star excitations, binary mergers and the persistent emission from neutron stars. We use data from the Advanced LIGO and Advanced Virgo interferometers in order to search for stochastic gravitational waves. Based on the the first three observation runs of these detectors, we have established upper limits on isotropic and on anisotropic stochastic gravitational-wave background. These limits improve on indirect limits inferred from Big Bang nucleosynthesis and measurements of the cosmic microwave background, and they approach the expected gravitational-wave background due to compact binary coalescences across the universe.

Searches for long gravitational-wave transients. Gravitational-wave transients lasting from seconds to weeks may be associated with sources such as young neutron stars following core-collapse supernovae, flares associated with isolated neutron stars and binary systems. We study the properties of such sources of long transients and look for their signatures in data from the LIGO and Virgo interferometers. We have run searches for long transients associated with GRBs in LIGO S5 data, an all-sky search for long transients in S5 and S6 initial LIGO data, in O1 Advanced LIGO data, and in O2 Advanced LIGO data. We also searched for very long transients (time-scales longer than 1 hour) emitted post-merger in GW170817 binary neutron star coalescence.

Searching for stochastic gravitational waves with LISA. LISA is a satellite-based gravitational wave detector, expected to be launched in 2034. It is a joint ESA-NASA project, consisting of three satellites separated by 2.5 million kilometers. LISA will explore a wide variety of gravitational wave sources in the milliHertz frequency band. We have developed a Bayesian pipeline to search for the stochastic background with LISA, including both isotropic and anisotropic approaches.

Correlating gravitational-wave and electromagnetic sky-maps. Maps of the gravitational-wave energy density across the sky carry information about the distribution of compact binaries throughout the universe, as well as signatures of the early universe physics. We are developing techniques to cross-correlate these sky-maps with similar maps of the sky obtained using electromagnetic observations (galaxy counts, gravitational lensing, CMB), including approaches based on coherence estimates and on N-point correlation functions. Our goal is to use these correlations to constrain models of structure formation in the universe and models of early universe physics.

The Bayesian Search for the stochastic background due to binary black hole coalescences. We are involved in the development of a Bayesian technique for searching for the background of gravitational waves produced by coalescences of binary black hole systems, most of which are not detectable individually. If successful, this approach will be significantly more sensitive than traditional searches for the stochastic gravitational-wave background, enabling new studies of the population of the black hole binaries in the universe.

Deep learning approach to cleaning LIGO data. Advanced LIGO detectors are sensitive to length fluctuations at the level of 1e-19 meters (10,000 times smaller than a proton). At this sensitivity, a variety of instrumental and environmental effects can couple into the detectors, increase detector noise, and mask gravitational waves. We have developed DeepClean, a deep learning method that uses a series of environment monitoring sensors (accelerometers, magnetometers, microphones etc.) as witness channels to remove environmental contamination from LIGO data. We are currently working on using this technique to produce cleaned Advanced LIGO data in low latency.

Studying gravity-gradient noise at Homestake. Seismic noise and fluctuations in the local gravitational field (called Newtonian, or gravity gradient noise) are large enough to limit the sensitivity of gravitational-wave detectors that operate on the surface. For this reason, it is likely that the next generation of detectors will be built underground. We have operated an array of 25 seismometers in and above the Homestake mine, South Dakota and collected data for nearly two years. We used the data to understand the behavior of seismic waves underground and their composition. The results of this study will be folded into the design of the next generation of gravitational-wave detectors.

Instrumental development. The sensitivity of (terrestrial) gravitational-wave interferometers at low frequencies (below ~10 Hz) are limited in part by seismic noise. However, it is desirable to extend the operating band of detectors to lower frequencies (0.1-10 Hz), where a large number of gravitational-wave sources is expected to be. We are studying the feasibility of new techniques to limit seismic noise at low frequencies. We are also developing quantum tunneling accelerometers that may allow easier tracking of the seismic noise at or near gravitational wave detectors.