It's been a long few years for everyone, but we're excited to get back to searching the universe for gravitational waves! The LIGO/Virgo/KAGRA detectors, which were down for commissioning and upgrades since April 2020, have just started the next observing run. The LIGO detectors started the 4th observing run (O4) on May 24 2023, which will run for a year and a half. This also means new glitches are starting to roll in! See here more information. And in Gravity Spy 2.0, you can explore some of the other data the detectors are recording.

It's been a long few years for everyone, but we're excited to get back to searching the universe for gravitational waves! The LIGO/Virgo/KAGRA detectors, which were down for commissioning and upgrades since April 2020, have just started the next observing run. The LIGO detectors started the 4th observing run (O4) on May 24 2023, which will run for a year and a half. This also means new glitches are starting to roll in! See here more information. And in Gravity Spy 2.0, you can explore some of the other data the detectors are recording.

Research

Gravitational-wave astronomy

Astronomers try to understand our Universe by making observations of the cosmos. Since the Universe is so vast, we cannot perform experiments where we control systems and see how they behave when we change something. Instead, we must observe what the Universe provides us with, and determine how things work from these observations. This makes each observation valuable, as it is potentially another piece in the puzzle of understanding our Universe.

The first astronomers used their eyes to observe. Then we developed telescopes to see fainter objects and resolve finer details. Throughout the 20th century, we expanded what we could see, by developing technology that could detect light outside of the range picked up by our eyes. Each advance in detector technology enabled new discoveries to be made.

The 21st century saw the start of gravitational-wave astronomy. Instead of using light to make observations, we use gravitational waves—ripples in spacetime created by accelerating masses. GRavitational waves enable us to observe systems that would be difficult or impossible to see using light, such as merging black holes.

The first observation of gravitational waves was made in 2015 by the Advanced LIGO detectors. The signal, known as GW150914 as it was observed on 14 September 2015, came from the merger of two black holes each about 30 times the mass of our Sun. Since then many more gravitational-wave signals have been identified, and the international gravitational-wave detector network has been expanded to include Virgo in Italy and KAGRA and in Japan. The fourth observing run of this network began in May 2023.

Laser interferometers

Key to being able to observe gravitational waves is having detectors sensitive enough to measure the tiny distortion in space created by a passing gravitational wave. This has required the development of sophisticated detectors that are isolated from unwanted (non-gravitational-wave) disturbances, referred to as noise.

The basic idea of our current gravitational-wave detectors is to measure the stretching and squeezing of space caused by a passing gravitational wave. To do this, we need some objects that are free to move with the stretching and squeezing of space, we call these test masses, and we need some way to measure the distance between them, which we do by timing how long it takes for light to travel between them. The overall design is a laser interferometer. Our test masses are mirrors at the end of 4 kilometre arms, we bounce lasers between them, and use the interference between the laser beams to measure the relative change in time it takes for light to travel along each arm.

While a basic interferometer only requires a laser, a beam splitter and a couple of mirrors, the design of the Advanced Laser Interferometer Gravitational-Wave Observatory (LIGO) detectors is much more complicated in order to provide the performance needed to measure gravitational waves. For example, sophisticated suspension systems are needed to hold the test masses, to ensure they do not move due to disturbances of the ground; extra mirrors are added to boost the amount of laser light traveling in the arms, and additional mirrors are added to ensure that the laser beam has the correct shape. Scientists have worked for many decades to minimize sources of noise.

No detector is perfect though. One particularly troublesome type of noise are glitches: bursts of noise that complicate the analysis of gravitational-wave data.

Glitches

Glitches come in a variety of different types, and may have a variety of different causes. Some may be the result of environmental effects, such as ground vibrations, while others could have instrumental effects, such as irregularities in laser output. Some glitches may be a combination of both, such as those resulting from scattered light, where a bit of laser light gets unintentionally reflected, and the pattern observed in the detector output depends on the motion of the reflecting component and hence ground motion. Ideally, we would like to eliminate glitches, but this requires understanding them.

The original idea for the Gravity Spy project was to classify glitches. This helped detector scientists in two ways. First, it provided a large set of glitches that could be studied. This made it easier to look for any common patterns that might hint as a cause. Second, it enabled identification of new classes of glitch, indicating a change in the data from the instruments. This was particularly important as it might mean that analysis of the data needs to be updated. Our expanded project now seeks to go further in identifying the causes of glitches, by looking at correlations between the output of the detectors (the gravitational-wave strain data) and monitors of the state of the detector and its environment (auxiliary data).

With your help, we may identify the origins of glitches, and how to improve our detectors such that we can observe more of the Universe.

Machine learning

While Zooniverse volunteers may be fantastically hard working, there are too many glitches for humans to study all of them. We want to be able to use computers to analyze glitches as they are much better at dealing with large amounts of data. Machine learning methods have been effective in classifying glitches. However, we require training data to build these. One of the key outputs of Gravity Spy are glitch data sets that can be used to train machine learning algorithms to automatically classify glitches.

Furthermore, it is difficult to use computers to identify new types of glitch, and to identify new patterns in the data, such as a correlation with an auxiliary channel that provides a hint as to the cause of the glitch. Hence, we are relying on humans to look through the data to find these patterns. Using these human insights, we will be able to build improved machine learning algorithms that will help LIGO scientists to better understand their data.

Combining the strengths of both humans and computers, the Gravity Spy project will enable LIGO to make new advancements in gravitational-wave science.