Galaxy Zoo: Clump Scout II

What are clumps?

Clumps are big, bright regions in galaxies where stars are being formed more quickly than usual. They can be anywhere from ten light-years across to over a thousand, and come in a wide range of shapes, colors and properties.

What's the goal of the Clump Scout project?

We're trying to identify all of the clumps in the Sloan Digital Sky Survey (SDSS), which has images of hundreds of thousands of nearby galaxies. We know some of these galaxies have clumps, but we don't know which or how many.

Why are they called "clumps"?

When mid-20th century scientists found a new type of galaxy with bright spots scattered around it, they didn't know quite what to make of them. Astronomers decided to call these galaxies "clumpy" in order to distinguish them from their smoother-looking counterparts. From there, the bright spots themselves became known as "clumps", and the name stuck.

Why are clumps worth studying?

Galaxies in the early universe used to be brimming with clumps, but today, there are very few. Their vanishing act raises many questions: Why did clumps form in the first place, and what happened to make them stop? By learning more about clumps, we're trying to understand how galaxies and the universe as a whole have changed over time.

We can also use clumps to test existing theories about how galaxies work. For example, some theories predict that galaxies are stable enough for clumps to live for billions of years; others predict that galaxies are less stable, and clumps are constantly being ripped apart and reformed. We can test both of these theories, and others like them, by finding real clumps to study.

What if I'm not sure if something marked by the machine is clump?

You can read the Field Guide on the far right-hand side of the page at any time, which shows many examples of normal clumps and odd clumps that should be marked.

If you don't think that an object marked by the machine model is a normal clump or an odd clump, you can delete the marking. Instructions on how to do this are available in the Field Guide.

After I submit my marks, what happens to them?

The first step is that marks are "aggregated". This means we take all of the marks made by all of the volunteers who classified the image and form a consensus shape. We rely on crowd agreement, areas of the image that many volunteers mark are designated as clumps, while those marked by only a few volunteers tend to be ignored.

Once we're certain that people's marks agree, we retire the image and it is no longer shown to users. The locations of its clumps become official scientific data.

If a computer can do this, why do you need humans to help?

Thanks to the volunteers who contributed to the first Galaxy Zoo: Clump Scout project, we now have a deep learning model that is fairly good at identifying clumps. However, as you will see while working on the project, our model is far from perfect. Identifying clumps may seem easy to us, but computers (even sophisticated deep learning models) have a lot more trouble. The reason for this is that clumps appear in images that are full of extraneous information. Computers have to sift through other galaxies in the background, bright stars in the foreground, and the complexities of the subject galaxy itself before they can isolate clumps from all these competing signals. For computers to get better, they need to to be corrected by humans when they get something wrong. That's where you and this project come in!

Human volunteers are also better at drawing attention to clumps that are interesting, unexpected, or just plain weird. Based on your feedback, which you can submit via the Talk boards, we can go back and examine these clumps in more detail. This is how citizen scientists in Galaxy Zoo project discovered Green Peas, an entirely new class of galaxy we weren't expecting.