Fourth Science Run Research Articles

Application of machine learning algorithms to the study of noise artifacts in gravitational-wave data

The sensitivity of searches for astrophysical transients in data from the LIGO is generally limited by the presence of transient, non-Gaussian noise artifacts, which occur at a high-enough rate such that accidental coincidence across multiple detectors is non-negligible. Furthermore, non-Gaussian noise artifacts typically dominate over the background contributed from stationary noise. These "glitches" can easily be confused for transient gravitational-wave signals, and their robust identification and removal will help any search for astrophysical gravitational-waves. We apply Machine Learning Algorithms (MLAs) to the problem, using data from auxiliary channels within the LIGO detectors that monitor degrees of freedom unaffected by astrophysical signals. The number of auxiliary-channel parameters describing these disturbances may also be extremely large; an area where MLAs are particularly well-suited. We demonstrate the feasibility and applicability of three very different MLAs: Artificial Neural Networks, Support Vector Machines, and Random Forests. These classifiers identify and remove a substantial fraction of the glitches present in two very different data sets: four weeks of LIGO's fourth science run and one week of LIGO's sixth science run. We observe that all three algorithms agree on which events are glitches to within 10% for the sixth science run data, and support this by showing that the different optimization criteria used by each classifier generate the same decision surface, based on a likelihood-ratio statistic. Furthermore, we find that all classifiers obtain similar limiting performance, suggesting that most of the useful information currently contained in the auxiliary channel parameters we extract is already being used.

A coherent triggered search for single-spin compact binary coalescences in gravitational wave data

In this paper, we present a method for conducting a coherent search for single-spin compact binary coalescences in gravitational wave data and compare this search to the existing coincidence method for single-spin searches. We propose a method to characterize the regions of the parameter space where the single-spin search, both coincident and coherent, will increase detection efficiency over the existing non-precessing search. We also show example results of the coherent search on a stretch of data from Laser Interferometer Gravitational-wave Observatory's fourth science run, but note that a set of signal-based vetoes will be needed before this search can be run to try to make detections.

Targeted coherent search for gravitational waves from compact binary coalescences

We present the details of a method for conducting a targeted, coherent search for compact binary coalescences. The search is tailored to be used as a followup to electromagnetic transients such as Gamma Ray Bursts. We derive the coherent search statistic for Gaussian detector noise and discuss the benefits of a coherent, multi-detector search over coincidence methods. To mitigate the effects of non-stationary data, we introduce a number of signal consistency tests, including the null SNR, amplitude consistency and several $\chi^{2}$ tests. We demonstrate the search performance on Gaussian noise and on data from LIGO's fourth science run and verify that the signal consistency tests are capable of removing the majority of noise transients and the search gives an efficiency comparable to that achieved in Gaussian noise.

First LIGO search for gravitational wave bursts from cosmic (super)strings

We report on a matched-filter search for gravitational wave bursts from cosmic string cusps using LIGO data from the fourth science run (S4) which took place in February and March 2005. No gravitational waves were detected in 14.9 days of data from times when all three LIGO detectors were operating. We interpret the result in terms of a frequentist upper limit on the rate of gravitational wave bursts and use the limits on the rate to constrain the parameter space (string tension, reconnection probability, and loop sizes) of cosmic string models. Many grand unified theory-scale models (with string tension Gμ/c2≈10-6) can be ruled out at 90% confidence for reconnection probabilities p≤10-3 if loop sizes are set by gravitational back reaction.

Search for gravitational-wave bursts in LIGO data from the fourth science run

The fourth science run of the LIGO and GEO 600 gravitational-wave detectors, carried out in early 2005, collected data with significantly lower noise than previous science runs. We report on a search for short-duration gravitational-wave bursts with arbitrary waveform in the 64–1600 Hz frequency range appearing in all three LIGO interferometers. Signal consistency tests, data quality cuts and auxiliary-channel vetoes are applied to reduce the rate of spurious triggers. No gravitational-wave signals are detected in 15.5 days of live observation time; we set a frequentist upper limit of 0.15 day−1 (at 90% confidence level) on the rate of bursts with large enough amplitudes to be detected reliably. The amplitude sensitivity of the search, characterized using Monte Carlo simulations, is several times better than that of previous searches. We also provide rough estimates of the distances at which representative supernova and binary black hole merger signals could be detected with 50% efficiency by this analysis.

First cross-correlation analysis of interferometric and resonant-bar gravitational-wave data for stochastic backgrounds

Data from the LIGO Livingston interferometer and the ALLEGRO resonant bar detector, taken during LIGO's fourth science run, were examined for cross-correlations indicative of a stochastic gravitational-wave background in the frequency range 850-950 Hz, with most of the sensitivity arising between 905 Hz and 925 Hz. ALLEGRO was operated in three different orientations during the experiment to modulate the relative sign of gravitational-wave and environmental correlations. No statistically significant correlations were seen in any of the orientations, and the results were used to set a Bayesian 90% confidence level upper limit of Omega_gw(f) <= 1.02, which corresponds to a gravitational wave strain at 915 Hz of 1.5e-23/rHz. In the traditional units of h_100^2 Omega_gw(f), this is a limit of 0.53, two orders of magnitude better than the previous direct limit at these frequencies. The method was also validated with successful extraction of simulated signals injected in hardware and software.

Searching for a Stochastic Background of Gravitational Waves with the Laser Interferometer Gravitational-Wave Observatory

The Laser Interferometer Gravitational-wave Observatory (LIGO) has performed the fourth science run, S4, with significantly improved interferometer sensitivities with respect to previous runs. Using data acquired during this science run, we place a limit on the amplitude of a stochastic background of gravitational waves. For a frequency independent spectrum, the new limit is $\Omega_{\rm GW} < 6.5 \times 10^{-5}$. This is currently the most sensitive result in the frequency range 51-150 Hz, with a factor of 13 improvement over the previous LIGO result. We discuss complementarity of the new result with other constraints on a stochastic background of gravitational waves, and we investigate implications of the new result for different models of this background.