Bouman, 29, a postdoctoral researcher at the Harvard-Smithsonian Center for Astrophysics, had been working on such an algorithm for nearly six years, as she was a graduate student at MIT. She was one of about three dozen computer scientists who used algorithms to process data collected by the Event Horizon Telescope project, a worldwide collaboration between astronomers, engineers, and mathematicians.
Video: Katie Bouman was a postdoctoral student at MIT when she led a team that designed one of the algorithms that helped analyze data that led to the first images of a black hole. (Adriana Usero / Washington Post)
Telescopes around the world gathered high-frequency radio waves from the vicinity of Messier 87, a supermassive black hole 54 million light-years away. But atmospheric disturbances and metric awareness meant "an infinite number of possible images" could explain data, Bouman said. Well-designed algorithms had to break through the chaos.
The picture was shared Wednesday, which resembled a molten monk or Sauron's eyes or even a Rembrandt, a composition of several such reconstructions. "We blurred two pictures and then averaged them to the other to get the picture we showed today," Bouman said. The ring of material surrounding Messier 87, which has a mass of 6.5 billion suns, "is something we were incredibly convinced of."
The Washington Post spoke to Bouman after the image was unveiled. The following is easily edited for clarity.
Q: You're not an astronomer. So how did you get involved in taking a picture of a black hole?
A: I come from a computer science and electrical engineering background. I did my doctoral student in a computer vision group where you try to understand pictures. And I heard about the project, this idea of depicting black holes. At that time, I hardly even knew what a black hole was. But I tagged this meeting [where Shep Doeleman, the Harvard University astronomer who directs the Event Horizon Telescope project, was discussing black holes]. I had no idea what he was talking about, but when I left that meeting I knew that this was something I wanted to work with.
I have an interest in how we can see things or measure things that are believed to be invisible to us. And how can we come up with unique ways to merge instrumentation and algorithms to be able to measure things that you cannot measure with standard instruments.
Q: What was the role of the algorithm for this image, combining data from the telescope over the planet?
A: We have telescopes distributed around the world. For every two telescopes in the telescope matrix, we measure a single room frequency, which tells something about how quickly things change.
We get this part information. It's almost like seeing a pixel in an image (but it's in another type of domain). We have to come up with methods that take this really sparse, very noisy data and try to find the image that may have caused these measurements.
What we have to stop doing is to apply things called "regularizors" or "priors" that allow us to say "OK, of any images that might fit this data, this set of pictures is probably."
But the danger is that we do not want to inject further information into the problem, to bias our results against something we would expect to see. We have spent a tremendous amount of time making sure that what we saw was actually real and not just something that, even unconsciously, we could have been assigned to the tasks.
(To remove the possibility of bias shared by the whole team, the project shared its computer formation experts in four different groups, each working on a different type of algorithm. They were not allowed to communicate.)
F: When did you that the black hole was a hole?
A: We were convinced at least that we saw this ring-like function. Yet we did not know that other teams would get the same results.
We all gathered in a meeting in Cambridge, Massachusetts and on the second day of the meeting, we all revealed what image we reconstructed from the data. It was probably the most exciting moment I ever had with the project.
When I saw that we all had reconstructed this ring, I knew that it was an incredibly robust function.
(Many months later, computer scientists attempted to break their images. They developed new scripts or pipelines and trained these pipelines on data for discs; these astronomical structures have no holes. But when scientists fed the actual telescope data through the pipelines developed for disks, they reconstructed a ring. The function was.)
We did not get a record. We still have that hole.
This article was written by Ben Guarino, a reporter for The Washington Post.