Simulation-based inference for AGN jet population modelling: Towards more robust comparisons of black hole jet speeds

Imagine the universe as a giant, cosmic factory where the most powerful engines imaginable are constantly firing up. These engines are black holes, and they don't just sit there; they shoot out massive, high-speed beams of particles called "jets" that stretch for thousands of light-years. Think of these jets like the exhaust from a rocket, but instead of burning fuel, they are powered by the intense gravity and spin of a black hole. Scientists have long wondered if these engines work the same way whether they are tiny (like the black holes in our own galaxy) or gigantic (the supermassive ones sitting in the centers of other galaxies). To figure this out, astronomers need to measure how fast these jets are actually moving. But here's the catch: the universe is tricky. Because these jets move so fast and point in different directions, they look different to us depending on how they are aimed. It's like watching a car zoom past: if it's coming straight at you, it looks incredibly fast and bright, but if it's driving away or across your field of view, it looks slower and dimmer. This makes it very hard to know the "true" speed of the whole fleet of jets, because we can only see the ones that are pointing right at us.

This paper is a detective story about solving that puzzle. The authors, Clara Lilje, James Matthews, and Rob Fender, wanted to get a better look at the "parent population" of these jets—the entire family of black holes, not just the lucky ones we can see. They used a sample of 174 powerful radio quasars (a type of active galaxy) from a famous survey called MOJAVE. Instead of using old-school math that struggles with complex, messy data, they used a modern, computer-smart technique called "simulation-based inference." You can think of this like a video game where the computer creates millions of fake universes with different rules for how fast the jets go. The computer then learns to recognize which set of rules makes the fake universe look exactly like the real one we observe. By doing this, they found that the speeds of these giant jets follow a specific pattern: the number of jets drops off as they get faster, following a curve where the exponent is roughly -1.32. Crucially, they found that the uncertainty in this number is much bigger than previous studies thought, and the shape of the data isn't a simple, neat bell curve. When they compared these results to the speeds of jets from smaller black holes (called X-ray binaries), they found that the two groups are surprisingly similar. The slopes of their speed distributions match up within a 2-sigma margin, suggesting that the physics of how black holes shoot their jets might be universal, working the same way whether the black hole is small or huge.

The Cosmic Speed Trap

To understand what these scientists did, you first have to understand the "Cosmic Speed Trap." Imagine you are at a racetrack, but you can only see the cars from a single seat in the grandstand. If a car drives straight toward you, it looks incredibly fast and its headlights are blindingly bright. If it drives away, it looks slow and dim. If it drives across your view, it looks somewhere in between. Now, imagine you are trying to figure out the average speed of all the cars in the race, but you can only see the ones that are zooming right at you. You might mistakenly think every car is super-fast because the slow ones are hidden from your view. This is exactly what happens with black hole jets. The universe is filled with black holes shooting out jets, but because of a phenomenon called "relativistic beaming," we mostly see the ones that are pointing directly at Earth and moving near the speed of light. The slower ones, or the ones pointing sideways, are too faint for our telescopes to catch. This creates a "flux limit," a bias where our view of the universe is skewed toward the fastest, brightest objects.

The Old Way vs. The New Way

For a long time, astronomers tried to fix this bias using standard math tools, like the Anderson-Darling test or chi-squared tests. Think of these tools like trying to fit a square peg into a round hole. They work okay if you are looking at just one thing at a time, like the speed of the jets. But when you try to look at speed, brightness, and distance all at once, these tools get confused. They often assume that the answers follow a nice, symmetrical bell curve (like a pile of sand that is highest in the middle and slopes down evenly). They also tend to be overconfident, giving very narrow ranges for their answers and saying, "We are 99% sure the answer is exactly here!"

The authors of this paper decided to try a different approach. They used a method called Simulation-Based Inference (SBI). Imagine you are trying to guess the recipe for a secret cake. Instead of tasting the cake and guessing the ingredients one by one, you have a robot chef. You tell the robot, "Make a cake with 2 eggs, 1 cup of sugar, and 3 cups of flour." The robot bakes it and shows you the result. You compare it to the real cake. If it's too dry, you tell the robot to try again with less flour. The robot does this millions of times, learning the relationship between the ingredients and the final cake. Eventually, the robot learns the "likelihood surface"—a map of all the possible ingredient combinations that could have made the cake you have.

In this paper, the "ingredients" are the properties of the black hole jets (how fast they spin, how bright they are, how they evolve over time), and the "cake" is the data from the MOJAVE survey. The robot (a machine learning model called a "normalizing flow") simulates millions of fake universes, learns which ones look like the real data, and then tells the scientists what the most likely "recipe" for the real universe is.

The Big Findings

When the authors ran their simulation, they found some interesting things that the old methods missed:

1. The Answers are Messier (and More Honest): The old methods gave very neat, symmetrical answers. The new method showed that the answers are actually "lumpy" and asymmetrical. For example, the uncertainty in the speed distribution isn't a perfect bell curve; it's skewed. This means the old studies were likely too confident. The new study says, "We think the speed exponent is -1.32, but it could be anywhere between -1.51 and -1.12." That's a much wider range, but it's a more honest one.
2. The "Beaming" Factor is Solid: One of the parameters they tested was the "beaming exponent" (how much the jet's brightness changes based on its angle). The simulation was extremely confident about this one, pinning it down very tightly. This confirmed that previous scientists were right to fix this value at 2.0 in their models.
3. The "Speed vs. Brightness" Tangle: The simulation showed that some of the variables are "degenerate," meaning they are tangled together. For instance, the way the jets change brightness over time (redshift evolution) is hard to separate from the way their speed changes. The simulation captured this tangle perfectly, showing a long, stretched-out shape in the data where the two variables mix. Old methods would have just missed this complexity.

The Cosmic Connection: Giants and Dwarfs

The most exciting part of the paper is the comparison. The authors took their new, more accurate measurements of the giant black holes (Active Galactic Nuclei, or AGN) and compared them to the measurements of the tiny black holes in our own galaxy (X-ray Binaries, or XRBs).

Think of it like comparing a Ferrari to a go-kart. You might expect them to have totally different engines. But when the authors looked at the "Lorentz factor" (a fancy way of saying "how close to the speed of light they are moving"), they found something surprising. The distribution of speeds for the giant AGN jets and the tiny XRB jets are consistent with each other within a 2-sigma margin.

What does "2-sigma" mean? In science, it's a way of saying, "There's a pretty good chance these two things are the same, but we aren't 100% sure yet." It's like flipping a coin 100 times and getting 60 heads. It's suspicious, but not impossible to be a fluke. However, it's strong enough to suggest that the physics of how black holes shoot jets might be the same, regardless of whether the black hole is the size of a mountain or the size of a solar system.

Why This Matters

This paper doesn't just give us a new number; it gives us a new way of thinking. It shows that when we deal with complex, biased data in astronomy, we need to stop using simple tools that assume everything is neat and symmetrical. By using simulation-based inference, the authors were able to see the "lumps" and "tangles" in the data that were previously hidden.

They also showed that the universe might be more uniform than we thought. If the jets from the smallest black holes and the largest black holes follow the same speed rules, it suggests that the fundamental laws of physics governing these cosmic engines are universal. It's a bit like finding out that the same blueprint is used to build both a toy car and a real race car.

Of course, the authors are careful not to say this is the final word. They point out that there are still some differences in how we observe these objects (like the fact that we see the giant jets at different distances than the small ones), and that more data is needed to be absolutely certain. But by using this powerful new method, they have taken a huge step toward understanding the true nature of these cosmic speedsters. They've proven that with the right tools, we can finally see the whole family of black hole jets, not just the ones pointing at us.