 This does become a regular session at AAAS. OK, I'm going to talk for a few minutes about a pre-print that we have on the archive now titled Setting with Spatial Temporal Surveys. And as one of the speakers earlier alluded to, I think there's a great opportunity for some of the modern data sets that Strong is producing for all kinds of other reasons to do some really interesting setting work. And so I just sort of read this talk quickly with sort of four big takeaway points. The first is that I think Optimal Surveys, a couple mentioned here that are highly featured at this meeting. We have an open house tonight about tests. I think there's one tonight or tomorrow about LST as well. These surveys, I think, are primed to do new kinds of setting work that hasn't been heavily studied before. And so my soapbox point, if you only take one point away from this talk, is that we need more literature on possible types of signals or ideas of things that we can search from. Another way of phrasing that is that there's a huge opportunity here. If you are interested, if you're a student, you have a data set like every scope, if you've got sort of a wild idea, these data sets are the perfect sort of playground or sandbox to test some of these ideas out, to create algorithms that are portable to use on multiple kinds of data sets or are extensible to new data sets, or even go back through archival data. So I think there's a huge opportunity here to develop new ideas and to search over huge areas of the sky. LST is going to search over effectively half the sky, tests searching over 80% of the sky. The potential to look for new things over wide areas is already there. Point number two is that not only couldn't these surveys do set in search over wide areas of the sky for long periods of time, they might actually even be good at it if we can figure out how to do it. So as with almost all areas of astrophysics, supernovae, stellar variability, transiting exoplanets, things that are very popular these days, surveys can dominate if we frame our science for questions correctly. If we're just collecting things, if we're just going around collecting butterflies, then surveys maybe seem like kind of a waste of time. But if we frame our astrophysical questions, testing planet formation theories, testing the lifetimes of stellar activity, things like that, if we can frame our questions to make use of the surveys in the right way, then surveys end up dominating, become often the most efficient way to do science, which of course we've seen a huge growth of surveys. And so I think that's the challenge for setting. We need to start forming setting questions that can make use of the sort of cadence and the modeling that comes out of these surveys. Jason Wright and colleagues wrote a really nice paper now two years ago about exploring needles in the n-dimensional haystack. I'm gonna forget here on the stage, I think it was 11 dimensions, Jason? Nine dimensions, that's still sufficient numbers of dimensions that I can't plot it on this screen. But I can draw a nice little needle graphic here. The idea is that we have some large dimensional parameter space that we're looking for, right? We have so few ideas about what we're trying to find in any study search. We're looking in huge dimensions of sensitivity and area, integration time, cadence, polarization, things that are very specific of course to the radio. And I think that there is a very analogous argument we made about the dimensionality we're looking in the optical. Now we don't, in the optical surveys, like LST and TESS, we're not worried about polarization and modulation, but bandwidth is something the band pass we're looking through, the cadence, the distance of which we're sensitive to. So many analogous kinds of properties can be made. And so in this pre-print that we did, or that I did with some friends and I published last year, we explored trying to move this very radio-focused n-dimensional haystack into the optical and showing that making some very simple assumptions about similar kinds of transmitters and sort of brightness that we might be able to detect, optical surveys, so here is the sort of typical volume that Jason and colleagues put into his paper about radio, the optical surveys, if we make some simple assumptions about the kinds of signals you might detect, perform in this, the log haystack volume, perform one to two orders of magnitude better than the radio surveys. This is because these surveys are looking like in every scope here. Every scope actually beats LST slightly, which I love considering how big LST is and how small every scope is. These surveys really dominate in this volume, this parameterization, because they're surveying massive amounts of sky immediately over many years. So of course that's not optical, but optimal for many kinds of signals, but it is optimal for some signals that we might conceive of. To extend the analogy that they used in their paper where we might be looking at, I think optimistically they said it was like a hot tub as compared to the ocean or something in terms of the volume ratio that the study surveys were sensitive to, we might optimistically say that we're surveying a couple of swing pools as compared to the volume of the ocean. That's not a huge volume, granted, but if we're trying to suss out the life in the oceans with a swimming pool, it seems a lot more available than looking in a hot tub. It's like a much better volume to search over. Number three is that we do have, I'll call them trivial examples because they're things that we probably talked about a lot even today, but there are some trivial examples that we can point to for certain signals that we might look for. So I just want to spend a couple of minutes going through what I think are some easy examples, but these are hopefully, hopefully you will laugh at them and say, these are too easy, I have a much better idea as to incentivize you to go and run with this and create new literature. Now I want to, of course, highlight this awesome summary figure. It came out of the NASA Tech Research Workshop from about a year and a half ago. What do they talk about? What are the sort of best kinds of, what are the best kinds of signals? And the best kinds of signals are ones that a civilization would inevitably create. I don't think that the optical study literature that I'll point to today necessarily fills this inevitability criteria very well. It's a sort of convoluted or, yeah, there's sort of like complex signals that we might ask why a civilization would broadcast them, but we can search for them nonetheless. So of course, as I just mentioned, things like biology and the star, fascinating objects, astrophysically and, or setting up opportunities, big dips, long-term decays, strange modulations from the ground. These things, well, until we had better explanations, these things were somewhat unexplained and are eminently searchable from large datasets like the Kepler test datasets. One project that I'm very excited about and have seen recently some more literature on is the Vasco project, the vanishing and appearing sources over a century of observations. So here's my cartoon version of this. This is looking at sources over decades or even a century timescale, stars that would disappear or would very slowly fade with no other explanation. I think this is a very promising technique and they've got one interesting object referring back to a plate archive and hopefully with things like LSSD coming online can set new upper limits on whether or not this star, their candidates are actually disappeared or if they just had a state change, thank you. There's of course opportunities to look for things like Dyson Spheres, which are good. There was a very interesting object that's actually tweeted out a bunch a couple weeks ago, a rebroadcast or a light echo from Supernova 1987A, which I think is, again, astrophysically totally fascinating by the dust structure around 1978. But these also could be potentially a steady signal. You could imagine intelligent civilizations seeing a galactic scale event like a Supernova and rebroadcasting it as a way of saying like, oh, we saw this, we're here. This would be an interesting beacon structure. We don't think this is one of those that seem to be pretty close to what I'd be able to see here. So I don't think this is a candidate here, but something like this is something that we get right down as an algorithm, sticking to a computer and look for. There's an opportunity here. I think the laser-cooking technology idea that David Kuping and Alex peachy published a couple of years ago is a great example of something that, again, we could look for missing transits in a series of very regular transits with mission-like capital error tests. If a transits were missing one day, this would be an interesting potential source of a steady trigger that we might look at. And we might consider even more convoluted and complex measurements where we might look for spatial clusters of things, things that multiple stars that are doing things in a coordinated manner that should not know about each other. So if many transiting planets, for example, they're transiting at the same time or they've had the exact same radius and things like that. So I worked through a couple of little examples like this in the pre-print here. I won't spend any time on this, but it's the kind of thing that you can postulate and very quickly run some very simple machine learning algorithms on. You can find candidates and then you can realize that those are not significant to throw them out, but it took two seconds to run. So we can run unlimited numbers of these kinds of tests through our data. Okay, and finally, future directions. There's a ton of these surveys coming online and as they are building the infrastructure to look for supernova, to look for transiting exoplanets, we have the infrastructure to piggyback and do setting, I'll say for free, but I just mean we don't have to actually go to the telescope. We still need to pay people to do this. So it is free in the hardware sense, not in the people sense. In the future, we might imagine looking for all kinds of new types of light algorithms. We can imagine other algorithms that might be good for finding new kinds of signals. Things like I've been speculating and scratching my head about things that are repeating but not periodic. There's a ton of effort going on right now in missions like ZTF and every scope in LST and test look for periodic signals because they're very easy to look for repeating periodic signals. But what about things that repeat with a non-periodic sequence like some sort of Fibonacci sequence or a numerical sequence? We can use the same sort of approach as we used to look for periodic repetition and we can look for Fibonacci repetition. For example, this is a cartoon example that it didn't work. It's works but it's slow. There's algorithm opportunities here. Okay, and here at 10 minutes, I will just put up again these four bullet points that I think optical surveys, infrared surveys and all the other wavelengths that are going online are a prime opportunity to develop new-setting literature. They might even be good at it if you can learn how to deal with. There are some trivial examples in place and I would encourage and I would welcome, so every scope people who are in the back want to talk, I would welcome opportunities to talk about new kinds of signals in my book.