Simulations vs reality

47:44strong gravitational lensing is an indirect way for us to start to better understand the structure of these small dark matter clumps. And so, that's kind of now the intellectual thought process that brings us back to my question that I'm not going to let you leave the studio without answering, which is so what is it that you do every day? >> Yeah, yeah. Now that we have all the groundwork laid, I can tell you. So, you know I uh I'm interested in finding ways to use these lenses, gravitational lenses, to tell to to distinguish between dark matter theories that make different

48:25predictions for these clumps of dark matter. So, that's what that's what me and my colleagues think about on a daily basis is how we can make this measurement statistically. And so there's a few different There's a few different angles, right? So, one is the data. So, we try to observe uh as many of these uh gravitational lenses as we can. Uh the other one is from the the theory and the modeling side. So, you know, you you maybe you have some theory of dark matter that predicts that these clumps are less numerous than our than some other theory. We'll think about how to simulate lenses with those predictions built into the simulations.

49:05And we'll compare those simulations with reality in a statistical sense and try to determine whether the physics that we put into our simulations of these lenses looks if it produces data that looks like the real thing. And if we succeed in doing that, then we think that maybe the physics that we put into our simulations for the dark matter is correct. >> That's so the the Okay, I I think I'm tracking. So, there's there's like uh there's two There's two inputs almost. One input on one end is existing theories on uh the formation of dark matter. Uh theoretical frameworks

49:47uh that provide an argument as to why this is the way it works. There's, you know, n number of these that you could potentially be using. That becomes an input into the models you create uh the simulations that would then generate examples of gravitational lensing. Right? Um because we have like and so that's one input and then the other input on the other side is we have all this data from that's been taken from our real world with real gravitational lensing and we know where all those objects are. Uh and so we take sort of a model of the physics of the world. We add the extra ingredient of the dark

50:29matter physics. And we see can we generate a lens in the simulation that maps to this a lens that we've already captured and we have well characterized. >> Yeah, exactly. So, you know, that this picture that that the lens that we were looking at the the the yes, this guy. Uh so, you know, we can simulate on a on supercomputers uh lenses that look very very similar to this. You know, almost exactly like this, but they have different clumps of dark matter around. And uh they you know, some theories of dark matter with different properties for the from these clumps will do a little bit better than

51:09others at explaining this this particular lens. So, each one gives you a little bit of information. So, you know, maybe dark matter theory A is like twice as as likely if we only had this example of of a of a gravitational lens, we would maybe think, "Okay, the clumps, you know, maybe they have these properties, so dark matter theory A is a little bit better." What makes it powerful is when you do it with 30 lenses or 100 lenses or in the future maybe a thousand lenses. Like that's when we can start to make really really uh you know, we're already making interesting statements about dark matter, but we're about to be in a new era where we'll be able to do these

51:51experiments with thousands or tens of thousands of lenses. >> This is actually a good note because um I I think one of the things so the reason I kind of talked about it as inputs earlier um is one of those the input on the real world side is dependent on our space-based and ground-based telescopes and ideally space-based because we get rid of the atmosphere and the distortion and dealing with that is no fun. Um but it's sort of a currently inherently a limited data set, but that seems to be on the path to changing uh not only with Vera Rubin, which is already online and producing an

52:32unbelievable amount of data that's available like they're doing dumps every day or every 3 days right now, as well as the I believe it's the called the Roman Space Telescope, which is slated to launch later in 2026. Both of which as I guess quote sky surveys are going to just have an unbelievable amount of data, which with that would data from either of those two now become an another an input into that real world confirmation data set.