Going Dark: The Mystery of Vanishing Stars

okay you're live when we look up in night sky the stars seem eternal in some senses they are to a human lifetime but stars like people are born they live and eventually die of all areas of modern astrophysics and knowledge of how stars work is arguably the most complete and well understood of any physical phenomena in the cosmos so when a star just disappears astronomers are left scratching their heads as to why and what natural or artificial event has just occurred now you could say that losing one star is just careless but when comparing sky surveys of 70 years ago to recent studies it turns out that as many as a hundred stars may have vanished so the question is how well do we really know stars and what or who is responsible for these extraordinary findings good morning good evening uh good afternoon to everybody and welcome to the april 2021 edition of seti talks my name is simon steele i'm the senior director of education outreach at the seti institute the city institute has been holding these monthly talks on cutting-edge topics of astronomy in the search for life in the universe for over 10 years for most of these most of this time the talks are in person in downtown silicon valley but a silver lining to this current pandemic means that as we're now virtual we're able to reach audiences across the globe so please do type in the chat or on facebook let us know where you're tuning in from we also welcome your questions type them into the q a function on zoom or comment on facebook and youtube uh we have behind the scenes uh frank and rebecca who are gonna pass these questions onto the panels at the end but please hold off on your questions until the moderator gives you the nod uh we won't get to them all but we'll do our best to get to the good ones and the exciting ones and the controversial ones without further ado i'm very pleased to welcome two astronomers who are going to dig deep into a case worthy of the great sherlock holmes the mystery of vanishing stars and i'm going to hand you over now to our moderator of the evening and the conversation and a star that shows no signs of dimming himself uh the seti institute senior astronomer seth shostak take it away seth thank you very much simon and in fact i'll turn up the light here just to avoid any dimming here uh this is going to be an interesting talk tonight i mean they tell you that all the time but it really will be because we're not going to be doing the usual colloquium where somebody might discuss the the effects of inverse competence scattering on the line with forbidden transitions and white dwarfs i know you're interested in that kind of stuff but that's not what we're going to talk about we're going to talk about what seems like a very simple experiment of course it's not really simple but it's one you can understand right away a simple experiment with an extremely profound potential consequence so this would be really something different before introducing the speakers and getting right into it then let me acknowledge the behind the scenes steady institute employees who make these talks possible rebecca mcdonald lee lee jasmine head juice and very especially frank marshes i'm seth shostak and i'm a senior astronomer at the seti institute and uh let me uh be so brazen as to point out that the seti institute has been doing these talks for a long time you just heard about that but the seti institute is a non-profit so if you would like to to encourage these efforts to uh shamelessly uh sell the kind of science we're doing and the science that others are doing if you would like to help out on that just go to the city institute's website seti.org and uh you know weigh in on that okay let's now meet our two speakers i'll give a brief introduction for each and after i've done that they will each have a couple of minutes to tell you something about what they've done the first is beatrice villarroyal i will mangle her name consistently but not always the same way she's the principal investigator of the vanishing and appearing sources during a century of observations projects that that makes an ask acronym that closely fasco i can't get it right either vasco and uh that's the winner of the milpitas prize for convoluted acronyms pantries got a phd from uppsala university and did a post-doc at eth that's the swiss federal institute of technology in zurich since 2018 she's been a postdoctoral researcher at nordica in stockholm and iac at tenerife uh spanish island she's been a postdoctoral researcher at nordica well sorry she's been a post doctoral researcher at naughty to in stockholm from well i don't know how many years but this year they all were beatrice received the l'oreal unesco prize for women in science in sweden for her vasco work james davenport is a research assistant professor of astronomy at the university of washington and the associate director of the deer act now that's another convoluted acronym for the data intensive research and cosmology institute he specializes in survey astronomy using large volumes of data from telescopes worldwide to study nearby stars in the milky way he's an expert in the magnetically regulated activity of stars including hot flares and cool spots and how these events may impact life he's also become an advocate for using big data techniques to search for intelligent life in the universe and i'm sure we're going to hear about that okay those are the abbreviated bios but what about the work each of the speakers will now spend four minutes each talking about their research after which we'll have a knockdown drag out discussion in which we hope to elucidate for you what's really important in this effort okay beatrice the uh zoom room is yours okay so i'm going to share the screen um how does undo um i think i did it in the opposite direction so let's try share their screens yes so i will stars i have seen them fall but when they drop and die no star is lost at all from all the star stone sky now this is a poem and by a british poet that shows his view of the night sky today we know that if a star dies iris transforms over billions of years into a white dwarf perhaps it explodes in a bright supernova but could a star actually just vanish or could any object on the sky just vanish this is something i've been thinking about since i was an undergraduate student i was really um wondering about this for a long time until i as a phd student found a way actually how one could check it and this is how the vasco project was born in the vasco projects we are actually searching for vanishing stars by comparing data from the 1950s to data from today and we are hoping to find our ideal candidate which is a star that has been there and you see it in many images and and later like now has completely vanished because uh our ideal candidate could basically be a techno signature a signature of stellar engineering because we don't know of any processes that would make stars vanishing in the milky way today except for the hypothetical idea of possible failed supernova however in the vascular project we also have our initial candidate selection where we select anything that could have been seen in the catalogs from the 1950s and are not seen today anymore this includes a lot of natural phenomena like optical after glows of gamma ray bursts variable stars maybe a flash of m dwarfs variable quasars and acting geolactic nuclei actually also other type of techno signatures can enter like a communication laser might give rise to a short transient event etc so this is what goes into the initial vascular candidate selection so have we found anything yet we have 100 initial candidates indeed here you can see a image from the 1955 this is from the post one data you see something here you do not see it in the 80s you do not see it in 2000 plus you do not see it in the most recent catalogues we have here another candidate it is seen in 1953 actually it's still seen in the 80s but it's not seen later in the sdss nor in pan stars which is the most recent and so on we can get around hundreds of candidates where most um most are only seen in one image from the 1950s here's another example i think this one is quite pretty it's seen there it's not seen there and here we see another one and it's not in there either so what do we think they are we think that these are short-lived transient events that were only lasting for a few minutes not really a star that has vanished most of them and um in these cases we are trying to understand what are these short-lived transient events uh we have excluded asteroids we have excluded variable stars we have excluded variable agn for this and we have excluded a lot of different things and what we are left with are optical after glows of gamma ray bursts of fast radio bursts and we also left with still the possibility the possibility that this might be end or flares and we're working on it submitting proposals to try to understand what we have seen and we are still searching for the ideal candidate by improving our methods and yeah and the citizen science project please very much okay thank you very much and the floor is now yours james i will stop sharing great i will switch on my sharing here good morning from a very sunny seattle there we go and i'll just spend a couple minutes talking about um sort of compliment what bh has talked about um what i'd say seti with big data astronomy what i think is the future or i hope is the future um so i like beatrice i got interested in this early on in my career and it's something that's been in the back of my mind as i've developed as an astronomer and it's this idea of using these large catalogs these large surveys and these giant telescopes and giant data sets to search for uh the truly spectacular and the unknown now a survey that we work a lot with and are helping develop at the university of washington is the vera rubin observatory which is going to run the legacy survey of space and time or we just call it lsst now here is the computer rendering of what this will look like here's a person for scale so this is roughly a house size telescope that's going to survey the night sky every three nights in chile and we'll do so for about 10 years now this is great pr imagery yep flying through outer space all fantastic but what makes me really excited um is here is actual footage a nice little drone shot of them installing the top end or so the top section the top mirror of the telescope this is an 8.4 meter telescope they had to get the largest crane available in chile up onto the mountain to do this it's a massive engineering feat and it's a really spectacular data set which will revolutionize effectively all of modern astronomy this will be the great sort of data tied that lifts all the science boats uh from the stars from the galaxies uh looking for planets it will do everything but most excited for me is that this is a massive data set that will help us explore the sky regularly every night so something like 17 billion stars will be observed over a course of 10 years that's 20 terabytes of data give or take a night something like 10 to 15 petabytes of data over the survey this is a giant imaging survey that taken so you can see this map of the sky that it'll photograph over and over and over and what i think is so special about this is not just this enormous catalog of images but this becomes this sort of portal for astronomers to not have to go to the telescope but instead to run our experiments through our computers and so the game is not to travel to the telescope and to take one picture of one object anymore it's instead to write algorithms to write software to search through 17 billion or more of these objects to look for the things that are strange and i think this is the true uh revolution in seti is it'll be able to do seti projects over effectively the entire sky all the time what makes me so excited about the vasco project and i'm so excited to talk uh here today with beatrice is that i think it's one of the best and simplest seti projects that have ever been developed here i've just stolen some of the images from her fantastic paper and what we're possibly looking for here in the most optimistic case is something like a dyson sphere or some kind of mega structure that's been created and we would say as an astronomer we would look at some chart like this of brightness over time and we're looking for something that drops off so instead of analyzing images from these big surveys we're looking for something that's got lots of measurements and then it falls off a cliff the star appears to disappear and what we can do with lsst or big surveys like it is look for new candidates we can use this kind of shape of the of the event that we're looking for we can write software that's going to look for these kinds of events and we can check on these old candidates this is an eight meter telescope this is one of the largest telescopes in the world and it's going to allow us to go back to all these candidates that vitra's and others are generating and get new data at even fainter limits to see what's going on with these stars are they still there are they still changing or have they truly disappeared and finally what i just say is that this is one of many possible experiments that we can run with surveys like the lsst we're searching for 17 billion stars to do something that we don't expect and on top of that there'll be transient or sort of flashes of things that are unexpected the night sky this will be a movie of the night sky that will truly be revolutionary and so i'm really excited to try to develop these new ideas all right james thank you very much patrick's thank you and now you're no longer going to be in control of the conversation because we're going to grill you with a couple of questions and by the way to those watching you will have your chance to ask questions later on but let's begin with you beatrice uh okay this work where you study the behavior or the appearance of millions i think stars you didn't say how many objects but it's someone reminiscent of clinical trials of for example the vaccine for for covin right you have tens of thousands of people that get the thing and the reason you do that is to find effects that only might be at the 0.1 level right one in a thousand people have this kind of reaction so you're looking at a lot of candidates and to see which ones you know have done something and you found 100 that did something so my question to you is that all right we're talking about stars that have disappeared but surely it could have been that you know the stars were also disappeared as it were before the palomar sky survey photos were made and in fact what you found was an appearing star we know about appearing stars novi that sort of thing uh have you you know sort of considered that and trying to figure out well how many appearing stars would i expect to find anyway um we are um so let me just reiterate if i understood your question correctly i you mean in case we do this surveys if we can find also a number of appearing stars well in the cross match we have done right now we have only specifically looked for vanishing things we have required all the objects that were there in the 50s to have vanished in the new ones in order to do the experiment the other way we will have to turn around the process but i guess my point is only that if you had made photos if you had a survey from the 1940s example you might not have found these hundred objects either yes exactly so actually the hundred objects we have might have been something not actually stars it's we think that most of these are something that just appeared very shortly and we're taking on an image in the 50s and then vanished so it's actually not so it's an event more than a star okay and and have you estimated how many for example novi or supernova you would expect to find in an experiment like this and not really we have we we're actually planning to do that um we have a theoretician who is going to look into that okay we estimated the number of failed supernova we expected and that was basically nothing yeah yeah but there are no you know players and stuff like that those one can expect but supernova in the milky way we're only looking at milky way data i mean it's a 100 one the last one was 100 years ago so we don't expect anything to be found in the supernova direction in the milky way or in this latest list okay all right let me ask you something james vasco shows that the sort of discovery that demonstrates a kind of discovery that can be made with very large samples heck the most important astronomy discovery i guess of the 20th century the hertz from russell diagram you know which explained you know the nature of stars in many ways and it was only done once we had well in the the late 19th century when there was just this massive amount of photographic spectra of stars right that led because was cecilia payne gopashkin went through these things categorized and suddenly we knew a lot more about the behavior stars and it was just because all these glass plates were stacked up okay so uh we're going to have a lot of data here from the vera rubin telescope you just told us about that and we can see things that change over the course of days not necessarily decades it's totally unexplored territory do you have any expectation that it will be fruitful other than the usual expectation it's a new instrument we'll find something new i mean that's right at first blush whenever we go into a new realm of data a new wavelength regime if we go to the infrared or to the ultraviolet or the radio whenever we go anywhere new we find new things we have been consistently surprised by the unexpected in astronomy and so that's that's good job security for us which is which is great um i i do think there's a lot of opportunity uh with the ribbon observatory and other big telescopes big surveys um to ex to to discover what we i guess what we don't expect from normal stars and i think a great example of this is something that came from the kepler mission which i know the seti institute has talked about which is like boyajian's star here is a apparently very normal looking uh f-type star slightly slightly more massive star than our sun that one day decided to do something very bizarre and we're still arguing about it almost 10 years later about was it comets or was it dust or was it a bunch of moons we do think this is not a dyson sphere or an alien megastructure but it is totally new behavior for this kind of star so we can almost guarantee that we're going to see things like this in these data sets and i like your your use of the the hurts upon russell diagram because um it's it's something that pops out that tells the story of how stars live and die and evolve and it comes both as you said from a large set of data but also it comes when we learn how to sort our data how to make that graph how to which axes on the graph to look at those kinds of innovations pop out when you start taking new kinds of data in that case it was taking these as you said big surveys of spectra and colors and brightnesses we're taking all new kinds of measurements this is a precision instrument which takes measurements in time and so what will be the new hertzberg muscle diagram where we sort a billion dots to look for the strange outliers in time or in color or in variations i don't know we won't know what those revelations will be until we take that giant catalog and just cut it every possible way we can think of all right so so clearly i mean you seem to be convinced and i think justifiably that bringing on this new telescope not that it's the biggest telescope in the world not that it operates at particular wavelengths that haven't been looked at before but simply by photographing the night sky as much as you can see there every three days they're going to be big discoveries yeah i mean i'm how about this i'm betting my career on it right we're all we're all we're all very invested in this is you know this is one of the top funded priorities for astronomy worldwide because we do think it is going to as i said lift all scientific boats we're going to discover new kinds of supernova and transients we're going to make the best estimate of planet nine's um uh existence we're going to look for things that go bump in the night that we've never even thought to look because we never looked for that long because nobody had one telescope doing this for a decade we're starting to look at things that change over a human lifetime and i think that's going to be really unique yeah i think i think it's also the case that you couldn't have done this 30 years ago because you had photographic plates for imaging and they were much less sensitive of course but you also couldn't make you know thousands of photos every night that's right and and we couldn't have processed them i mean one of the unique things that the lsst will produce not only is this image of the night sky every three nights but anything that it detects that changes between um between visits it will publish within in real time within one minute so if you are another telescope looking for a particular kind of supernova or asteroid and the lsst project observes one within a minute they will publish an alert in real time on the internet that anybody who is you know any other telescope who's interested in these kinds of events can go and chase so it really is kind of this democratization of the access to the night sky with this massive telescope those kinds of facilities are only available because we have the internet because we have large computers because 10 to 15 petabytes of data while it's too big for my laptop is not too big for the giant cloud servers to handle so it really is a great moment in time where we can actually crunch these numbers for the first time all right back to you beatrice someone could argue that the vasco project is the most important seti experiment going you know seti has traditionally looked for signals radio signals flashing lasers stuff like that uh and there could be societies out there that are millions or billions of years more advanced than we are maybe for whatever reason they don't flash lights into the sky or maybe they don't send signals that by chance happen to arrive at our antennas just in the couple of minutes that we're actually looking for them so uh what you're doing is going to i mean it seems that it's the best way to look for those societies that are way more advanced than we are but you know might be a thousand light years away that sort of thing they could make a change to the comic cosmos after all i mean these these punitive societies that are that are very big we've talked a little bit here about dyson spheres but i mean you know that's that's a 20th century idea probably 23rd century ideas that will be just as impressive so do you think that uh you know i mean you sort of implied this that such activities have already taken place that the aliens some aliens the best aliens if you will i've already done something that you could pick up this way um if hm well i i'm sure i'm an optimist i wouldn't be looking for et unless i believe that they have already created visible thickness signatures so i i guess that there might be quite a lot of techno signatures out there let's say that let's make an example i'm going to just my mind experiment let's say e.t has been um very interested in our earth for whatever reason sent here a lot of probes some i don't know hundred thousand years ago or ten thousand years ago just to put it make it slightly more contemporary and um possibly these probes could be floating around in uh in a solar system and one of the things that one would be able to pick up it would be something that looks like satellite cleans in in the 1950s for example and these kind of things we could be able to pick up with vasco today someone will try to like use modern data most of these things will be um confused with human-made things so this is one of the things that one could actually pick up in a radical time scale if the if it would exist as an example okay can you give us some idea of the sensitivity of the experiment obviously the dimmest thing you can recognize depends on how far away it is but you know within 200 light years of the vasco experiment within 200 light years distance you have a million star systems so you know how faint could a star be uh and still show up in this kind of an experiment or is it you know severely limited that way well and you snow catalog has a limit of 20 mag in in the r band um yeah that may not mean too much to to people do you know what what distance would the sun appear to be 20th magnitude oh now you with this quick stress i will give you a random number okay but i mean you know that's that's maybe the first reaction that people say okay this is a nice idea but clearly if you're not at least as bright as a star will not see you e.t yes you don't agree with that do you sorry what did you say did you do you agree with that that radical statement i mean that we might not see it or yeah yeah i don't i think i don't quite follow could repeat okay i know i i won't torture you torture you anymore with this question no i mean you know people can very easily tell you for example if you're doing an experiment with this antenna over here trying to find a signal yes and it's coming from a couple of hundred light years away you can sort of you know reverse that argument and say well how much power do they need for us to be able to see that and and this experiment seems to be working on the the scale of you know stars in other words things that people don't build presumably and uh so i just wanted if you have any education we already have the technology like on earth in order to um produce something visible for those who are far away from us i mean the laces we have today if we really would want to i'm sure we could just direct one laser in one direction and hope that someone picks it up but why would we do that why would we waste energy if we don't even know where we would beam it so when it's a lot of luck it's uh i mean it's a big space uh james once you get going on this uh vasco like experiment using the vera ruben telescope how often will you do it will you just continually compare you know the latest last night's data with the data from three nights ago and just do that endlessly or will it be a finite uh experiment the way it has been so far i think i think the uh the former is the right answer we do it endlessly and hopefully i don't have to do it hopefully we teach the computers to do it right that's uh that's the ultimate goal here is i'm not going to be able to search through any fraction of the of the candidates so so beatrice has turned up a hundred candidates um using millions of stars from from the existing sort of grant we would say ground-based archival data stuff from the 1950s every night lsst that is online is going to produce you know five plus billion stars it's gonna it's gonna observe and it's gonna try to compare those automatically to its archive that it's already generated that i mean we can only guess how many of these sorts of vanishing candidates are gonna be popping in and out every night but it's probably hundreds a night so very quickly we'll run out of available astronomers and available eyeballs to chase these things and that's why we have to start getting really picky about which ones are the most interesting which ones are the most likely so being able to put those constraints on how far away could this thing be or or how dim could it have gotten as you said those are going to be really important constraints to understand knowing everything we can about these objects over the decade will be really important to say all right this one is probably some exotic flare star interesting but we're going to throw it out and we'll be left with hopefully some number that is actually chaseable that we could go after with even bigger telescopes or even more long-term dedicated observations to see if they pop out or maybe they turned infrared sources or you know something that uh explains what's happened to these stars or have they truly just disappeared baitries you know the way disappearing stars might have been found in the old days the old days are not so far in the past actually uh is that you would you know have these photographs and you'd have here's a photograph from 50 70 years ago and here's one from last week and you put them both into what's called a blink comparator you put them into a machine and some astronomer would sit there you know ruining their vision by comparing these two things and seeing any that blink because they've changed are they computers or you say anything this is the difference of my glasses before i did the vasco uh study and this is after three months one day of trip one diopter in three months that sounds pretty rapid to me okay well well well you see where i'm going with this i mean today we have machine learning right artificial intelligence and so forth will these i mean how long will it take you to do the vasco experiment again assuming you know you have the very reuben data and machine learning or new programming that allows you to save your vision um well that depends on how long time it takes for us to set up a very reliable code because it's not the easiest because we have this hundred objects it's not a um ideally when we want to have something like maybe 5 000 transients that you can train your ai data set on with hundreds we are it's a quite small sample in comparison to the uh 24 000 of objects that we also look through visually so i guess that once we set it up it might be a question of maybe two three months to run through all the data but we're also actually setting up some automatized methods where it can probably go in a few days but then we also lose a lot of candidates yeah but i mean you just heard that james is going to give you a billion candidates you know at least a billion observations not necessarily candidates you know every couple of days so clearly you've got to automate this to do something faster [Music] does that put you out of a job probably um that's why no no way it doesn't your your job is safe like we need the training data you're absolutely right i think i think veterans hit the nail on the head is that we only have 100 candidates of these things we're going to have a billion 17 billion likers we're not going to know what we're looking at we need more candidates i think she's exactly right we need more training data but you say training data james right that that's examples you know from existing data sets so that the machine learning algorithms have learned something and then can find things in the new data but you know once he's done the training data that's like showing somebody how to do your job then they usher you out the door that's that's my question beatrice doesn't seem worried okay i never think about the future well of course yeah who does but but i i guess my bigger point here is that this sounds like uh you've opened up a door that is very very promising and the first thing you try to do is what people who are looking for signals from aliens would look for and that is to automate as much of it as you can and now you can sit back and wait for uh the result you're hoping for and book your flight to stockholm so you can collect your prize actually we are automatizing it together with a spanish virtual observatory um we have a collaborator there who is like he's not usually machine learning but he's using like more classical image analysis and uh he's doing fantastic job with like getting new transients now and he's redoing basically the vasco search that we were doing but um without the catalogues and directly with the images and there is also really promising just let me follow up just a little bit uh there have been a lot of ideas coming across the screen here from people who have you know their their favorite explanations for these disappearing stars um if somebody were to get you you know across the table at a dinner and ask you well look i mean you've had plenty of time to think about what these things might be i would guess an optical afterglow of gamma-ray bursts or some type of flares from stars and that would be my favorite explanations is the natural phenomena um interesting enough for us to kind of look more into it is because of the time scale that it's really a few minutes and you only have time to take they only had time to sit on one plate and because this plate had the red plates had about exposure times of 50 minutes so that would be my guess some natural phenomena i might be wrong of course but that's um the first things the first reasonable explanations i would find how would you decide well we are going to do follow-up observations if we find some infrared counterpart to most of these transients that would be the m dwarf flares uh hypothesis but some of these because i've been looking at the amplitudes and so some of these are like it is like really big changes now if we can exclude that then we're only left with gamma-ray bursts unfortunately burst optical counterparts and that i'm not sure we're gonna be able to kind of sort out very easily however um if we it's possible that we will find in some cases some type of counterparts exactly on that spot and the question is if it looks like a m dwarf or not if it doesn't look like an dwarf then it gets much more interesting and if we find nothing it's still very interesting for us as well um and if it's late lsst can follow up on these points if lsst just keeps watching on them maybe one day it will be seen again i can also imagine that that it's simply things that needs to be monitored during a longer time period these 100 points okay so it sounds like what you're saying is look here are the various classes of astronomical phenomena that i know about that could produce a rapid change in the brightness of something so let's one by one eliminate those but surely you know you're going to have a a whole if you will cascade of these tests and at some point you have to allow for the possibility that the the only explanation you can think of is some sort of artificial origin and that you really have found the aliens yeah it's for instance a lasers uh could maybe give this red transients also on the scene or satellite gleams in the 1950s before the first satellite was sent i mean that then you would have a you for flying there or something like that we can't exclude those things either can you find it i'll ask you something james i mean is this vasco uh approach able to find things that are very nearby you know people have been suggesting for decades there could be objects in our solar system that you know bear on these questions yeah i think i think the answer is definitely maybe uh which is which is the great science answer um there's a lot of people going to who are planning to use lsst and the very observatory to study comets to study asteroids um so things that we can visibly see moving in the night sky that will be whizzing across the telescope's field of view within each exposure so being able to track those objects as they trace themselves across the sky is really good um bread and butter science for a lot of people and you know really important for us here on earth if we want to stop killer asteroids and uh understand the origins of life in the solar system there's lots of good like i said bread and butter science that means we want to study comets we want to study asteroids and things that move um but things that move in unexpected directions things like omumu which might be interstellar visitors this is i think a really exciting new avenue um and you know without wading into the debate about what the origin of umuamua was lsst is the perfect observatory to be discovering one of these things like every month if they are just you know chunks of rock or ice floating through interstellar space we expect to see something like one every month with this facility and so tracing them as they come into the solar system watching as they go out of the solar system it's it's a huge opportunity for science and then of course it's a huge opportunity for uh et work for techno signature work if these things decide to take a 90 degree turn one day we'll be watching will you be able to you know significantly improve our catalogue of uh near earth objects things that might you know ruin everybody's old day by slamming into the earth or i think you know i think the estimate is that um within the first month of lsst we'll discover more asteroids than uh are pre are currently known in all catalogs and so there's like fun cataloging issues here where we do keep these real-time catalogs of asteroids these names you know they have these long telephone number like id numbers uh and we've actually broken that catalog already thinking about how many this this facility is going to discover we need to add more digits to the phone numbers as it were because we're going to discover so many little chunks of rock floating through the solar system and we have to keep track of them because you don't want the same rock to show up a year later and to think you've discovered a new one you need to keep track of it between exposures and be able to map its orbit so there's a lot of great computational challenges there when if you only get a couple of quick snapshots of something and being able to predict where it is a year later so it's a great um science and math problem uh which i there are some really bright minds here in seattle and elsewhere working on you know that that's such an interesting result that's almost guaranteed i guess you could say is guaranteed to come out of this instrument i'm just sort of curious you know the proposal that went into develop and build the veer ruben scope what did it have is its number one science motivation what what was clearly going to be beneficial about this kind of an instrument um i think thinking back to the 90s when this was first planned out so it's sort of a successor to the sloan digital sky survey which we operated in new mexico here in the u.s but just on a much more massive scale the big motivator there in the late 90s and early 2000s was dark energy and the origin of the universe so trying to understand what the distribution of galaxies are very very far away so at the sort of edge of the observable universe and being able to map very deep images of galaxies meanwhile they they sort of realized throughout the 90s and 2000s as we did these kinds of smaller scale surveys that they're all all these things that people like me get our phds on flare stars things that go bump things that move there the amazing thing is you build a telescope to study dark energy and you end up writing papers on comets and like the origins and the families of comets that are in the solar system and so it really is an amazing opportunity for everybody to do science it's not perfect for everything we're not going to do radio astronomy with it we still need facilities like the transiting exoplanet survey satellite to discover a little transiting planets around other stars you know it's not perfect for everything but it really can impact every area of astrophysics and i think that's you know so i i believe i mean this is me i don't know not quoting the website but i think dark energy is still the number one science priority but there are working groups in stars and in the galaxy and in planets and then comets and there's people working worldwide on preparing for the science from this facility in every area of modern astronomy uh how would you describe your own motivation for undertaking this project this this is a big project i've i've known people who've done these sorts of surveys and it's just you know and until you can automate it it's an enormous amount of work what was what what did you hope to find when you started this well uh a star that has vanished or or anything that has vanished of course if i see a big big beautiful spiral galaxy that mentions i would be even happier wouldn't be so happy for them well it's just curiosity um until i get my answer i will go on i guess you surely have uh in fact considered how you're going to indeed and i think we've touched on this briefly how you're going to decide you know every time you do this it's going to be every three days you're going to find vanishing stars and you know you you're going to need some way of recognizing what things are that are going to have to automate that and you can only do it on the basis of what we already know astronomically speaking so uh it sounds like in the end it comes down to the same sort of problems that other people who are doing seti have which is oh here's a signal that i didn't expect how do i know whether it's really e.t yeah the first thing one wonders is of course could this be some type of instrumental error if you find if you find a very interesting candidate you wonder if you wanted like if there's something wrong with these plates that's the first thing we we wonder and that's the first things we want to test for if there are some obvious uh artifacts or some obvious signs of problems and the next thing one test for is that if there is some natural explanation for it and then once you start running out of it and you wonder if you have missed something and then you might be stuck for another one or two years just wondering yeah it's um so we would actually need to develop like complete um scripts to follow on how to um handle each of these candidates we haven't had time to do it yet but we are learning but it's certainly the case there's a long history of this now that every time something unusual or unexpected is found in the sky whether it's quasars or pulsars or you know uh stars that get 22 percent dimmer suddenly one night that kind of aliens usually get to blame often they get to blame because there's no other explanation but of course they're almost well there always is another explanation eventually but it takes a theoretician's a year to come up with it so uh you're happy you're okay with that right because you're going to get a lot of funny phenomena very quickly that's true that is true i can't argue against it all right well we have a couple of more minutes here um i i would just like before we turn this over to the audience to ask their questions and that is james the vera rubin uh telescope is not the only if you will rapid fire telescope i mean um wasn't found with vera rubin of course it was found with another another telescope on uh i think it wasn't mauna kea but it was haleakala i believe the pants telescope [Music] anything you want to say about why this telescope is really the only one or the best one or something about i mean if you're going to find new phenomena this is the way to go build it it's like a large hadron collider build it and you'll find something interesting the you know the one thing that the lsst project is going to have that nobody else has is what we would say is depth or aperture this is an eight meter telescope it's a you know it's it is unlike any other facility in the world in terms of its ability to look at very faint things so that puts it in its own class of objects that it's going to observe that's that's how it gets to 17 billion stars to study over a decade but you know let me take off my lsst hat for a minute and just say that we have a dozen other so-called you know survey telescopes looking at the most bright stars think telescopes that are you know the size of my coffee cup here um you know essentially just camera lenses that are studying the sky looking at the brightest stars and they're also doing phenomenal work and i think if we do our jobs right if we develop our algorithms to study these phenomena if we develop the follow-up capabilities that beatrice is talking about where we can take all these candidates and classify them like that's interesting that's interesting the rest of these are garbage we'll we'll write nature papers about them but they're not aliens if we do our job right these same algorithms are going to work on the transiting exoplanet surveys and they're going to work on the bright star surveys like assassin and every scope and there's a bunch of these other acronyms that we can go into but there's there are these surveys already operating around the world so we're developing our techniques here to watch the sort of real-time transients the sort of 60-second releases that come out so we've got a test data set right now out of california called the zwiki transient facility or ztf that has nightly alerts something like a hundred thousand to a half million of these alerts every night so about an about factor of 10 less than we expect from lsst but it's the perfect data set to learn what the uh the garbage is which just also happen to be good astrophysics but there are lots of these surveys and i think if we do our jobs right the same code that we write to find the aliens analysis he will be operating on hopefully every one of these surveys across the world and in outer space i like the fact that you said that everything you find that isn't et is garbage that's sort of an interesting point of view i mean i mean every every one of these candidates is going to be a boyajian star or it's going to be a rare r-corona borealis star there's you know there's a bunch of other things that have right now sample sizes of 1 or 10 or a few and if we discover 100 of those that's a revolution in that area of astrophysics okay we didn't discover e.t but it's still its own scientific revolution just very briefly because we want to go to questions now from the uh the viewers if you will our audience here well you mentioned that you can also do surveys of the sky with very small instruments and one of the advantages usually of a small instrument is you have a wider field of view what's the field of view of the you know the vera rubin telescope uh this is the number i should know off top of my head it's something like 100 square degrees per image so it's it's wide in a telescope from a telescope standpoint for something so large uh it's something like if you hold your hand out in front of you it's something like the size of your hand or your fist at arm length it's roughly that size and so it what it's doing is taking this huge mosaic what's so impressive about this facility from an engineering standpoint is you've got a camera that's the size of a person attached to a telescope the size of a house moving in a essentially a small football stadium sized building that is rotating and taking rapid fire snapshots it's these 30 second pictures and so you'll this whole facility will just move or the top part of it will just move every few minutes and take pictures of the sky it's a massive amount of glass and metal to be whipping around um and so it's a really amazing thing but it is the whole goal the reason it's so the reason it's shaped the way it is is it gives it that wide angle um yeah something like 100 square degrees all right i'm just trying to do the math in my head you know compared to something like the old palomar telescope i shouldn't say old i mean still use it the palomar telescope the 200 inch uh that sounds like almost a thousand times more area of sky than you would get with that telescope all right look we're going to uh open this up to people's questions be sure you put the questions in the q a when uh icon whatever it is at the the menu item down there at the bottom of the screen not in the chat so please do that the other thing i think we're going to do before you start asking your questions or before we start answering your questions is to mention our poll we always have polls here uh particularly with our connection to warsaw and here here's the poll question and it may come up on the screen so you can choose from maybe you can put that up on the screen here you are the best way to search for e.t either look at the entire universe right such as we talked about here or search nearby star systems or radios or flashing lasers radio signals fly from data or just look at our own solar system because after all there could be some probier from some distant civilization just checking us out so decide what you think are the best uh is the best choice now they were i have to tell you there's no guarantee that that will have any influence on future setting but who knows it might so if you do that while you're conjugating on that uh here's a chance that you know well i've already said you've you you ask your questions we couldn't get to everybody's questions of course when we do this but we will try and have the best ones those will be selected by rebecca and frank all right um can we go to the first question here rebecca yeah okay everybody thank you says so we have uh we have what we decided to do with rebecca is to put together questions so we have a we maximize the chance of having the good answers for all of them so we have a first question for beatrice specifically about electromagnetic information on those disappearing star you mentioned follow-up survey infrared spectroscopy can you please elaborate a bit and tell us what exactly you will be able to uh to see by having a follow-up of a disappearing star assuming that he happened just today what will you do um okay so let's say i have my ideal candidates i have this beautiful candidate that was seen in the 50s and the 80s and three days ago just finished so the first thing i will do is i uh apply for this like emergency time at the grand account telescope that is a 10 meter telescope and i would say let's let's have a look if it's still there's something still there or not that would be the first thing and i would want to say it for instance um maybe we would look there and there would be absolutely nothing you look in the optical you look in the infrared and there's absolutely nothing then i would start getting slightly stressed um of of course maybe you look there and you and you actually find something you find something beautiful glowing in the infrared because i don't mean maybe to build that isosphere maybe they built an isotherm it could be in it could be nice or maybe even the case that you find absolutely nothing just some a little bit of radio but you then maybe you have a look at the star from beginning as you know this could have been a yellow super giant maybe it's actually this was a natural process because if it says yellow supergiant it left a little bit of radio there some radio remnants that could be a failed supernova maybe this yellow supergiant star against all probability odds for the milky way would still kind of uh collapse directly into a black hole and and then uses and then in in two weeks you there will be absolutely nothing left there or in two years there would be absolutely nothing left so these are the ways how we could approach it i could also take like follow up with guyancy has it been stable all the time was it like perfectly would the light curve be completely flat before it vanished or is there a strong variability and these kind of things could help us to try to figure out what is happening because if you see that it's kind of is smoothly i don't know it gets fainter fainter and stronger stronger and later see this beautiful infrared thing maybe you can think about if it's would suit with a dyson sphere build up for example okay i was going to ask james to follow up but i think it's maybe better to go to the next question otherwise we won't get to too many yeah so the next question is for james we have a lot of questions when you talk about the vera rubin and sst people are very curious to know the details so what i meant by details is when will be the first light when the telescope will start operate scientific operation when the data will be processed when the alerts will be available on everybody telephone so they know when there is a new supernova in the sky do you have that i have some of those numbers off top my head you can of course go to i think it's lsst.org and they have big fact sheets of everything you could possibly want about what telescope weighs and what the color they've sprayed on the outside is and you can find everything about it there but it's something like i think we're about one year delayed due to the pandemic just like everybody else we the mountain was closed for a long time construction was halted so the dome is on the telescope is inside things are moving forward again so we're looking at first light somewhere in the next year and science operations about a year after that so there'll be about one year of commissioning where we're taking limited amounts of data we're trying to figure out all the ghosts that are hiding in our machines and anything that's in the wiring that we didn't expect and how we forgot to program something in the software the data will come out as i said the alerts will come out within 60 seconds they funnel through what we call brokers you know certain websites though i think based kind of like space twitter they'll basically just be streaming out in real time and you'll be able to if you're a scientist or you're an enthusiast you'll be able to write some simple programs that will watch those alert streams for things you're interested in that will send you so there's people thinking of ideas about cab sending me a text to wake me up in the middle of the night because my very magical thing happened making sure that that whole pipeline goes from telescope to my pocket or to my nightstand i think that's going to be really important and the data comes out the data is essentially free and publicly available worldwide with some asterisks and limitations on how we move data around the world but that will come out i think there's some regular schedule for publishing big catalogs something like monthly time scales those catalogs coming out so it's a huge opportunity for the enthusiasts it's a huge opportunity for astronomers worldwide at big universities like mine or at small colleges people can access this data almost every night thank you thank you for providing the details so i'm assuming it's not a question for them it's a comment on my side i'm assuming their choice will have probably the hundred candidates uh already ready into the database so in case one of them reappear batteries will know about that in the next 10 years good so i have a question another question for you beatrice this time um yeah a lot of people ask about those hundred candidates that you mentioned do they have anything in common do they disappear yes did some disappear at the same time do they move what kind of analysis you have done within those papers and can you tell us answer all questions uh and directly i can answer some of the questions and for instance many of them and these are 100 independent candidates that have been found from different times of the survey and they are what they do have in common is that most of them not all but most of them are seen on the red plates uh red photographic plates and um and most of them are only seen on that red play on the red plates but or not on any other plate which is uh speaking in favor for that they are something very short-lived something that appeared shortly and later vanished during the exposure time of the observation so that is what they do have in common and then they vary quite a lot in uh they are also very strong i would say like um in comparison they choose very quickly i don't want to don't don't give a long answer but if you see something on a red plate right you immediately look at the blue plate for that same exactly same era okay so you do that so what you're saying is that something disappearing tends to be red um this there are two two different ways of seeing it either it's actually is very red or we happened to catch them on a red plate because they have longer exposure times because of the times of the red places around 50 minutes of the blue pla blue plates it's around 10 minutes so you have a bigger probability to see something it's not color it's time it might i think so okay sorry frank uh no i have a now a question for james uh first but bear truth you really don't answer into all the questions to my question yeah you were done right you mentioned the did you discuss the proper motion of those targets but ask if they're close or far away and do we have it do you have any proper motion estimates and no well we know that they can't from the images we know that we haven't even blinked the images and tried to check and this and there is no chance that they just moved out of the view because it's because they're also point point like and that's something one needs to remember how did they be moving that fast they would be elongated so it's something very short and not something that moves fast okay so we have a question there's something that i didn't realize but uh it's true we talk about the dyson sphere but none of us clearly say what is it doesn't sphere so you have ton of questions about that what exactly is a dyson sphere what will be the what the dyson sphere will look like why a star disappear invisible and you mentioned appear in the near infrared because it doesn't sphere so james can you give us kind of a dyson sphere 101 class i i can try to give you a short dyson sphere 101 yeah so the general idea as it was proposed um is that a civilization would essentially build a giant ball or structure around its star you might think of pasting on the inside of the structure solar panels or you could plant plants or whatever you want to do something to harness that sunlight so we only get one earth's amount of sunlight but we have this whole orbit around the the sun that we could build a giant ball or a big bubble around our star and capture all that energy capture all those photons of light that are streaming off in all wavelengths so there's a huge energy budget available from the sun if you had one earth's rate orbital radius worth of solar panels that's the general idea and so the the vasco project very broadly speaking is looking for stars that disappear because they've been encased in this giant opaque bubble um now there's lots of like physical problems with like where would you get that much solar panel technology where would you get enough like rocks or metal to build something this massive in scale and so there's lots of like practical issues and this is something that's been researched a lot it's like would it even be stable could you even build a giant sphere with that sort of scale and or would it just rip itself apart from sort of gravitational pull of the star and torque from uh spinning in the orbit so there's a lot of concern that it's that a solid sphere is probably not likely and so a lot of work now is thinking about dyson swarms or sorts of elongated pieces of structure so we might not see it fully disappear we might see it become fainter and fainter as more these sorts of swarms of chunks of objects get launched to sort of islands of solar panels get launched into orbit but you know the short answer is i haven't built one so i don't exactly know so i saw that while you answer into this question we get more and more questions about atmosphere so if you want to know more about dyson sphere now we know who to contact james davenport don't give your phone number though just give you an email address that would be enough you can tweet me your questions okay so another question and this is for both of you and it's a question that um is um it's a bit about the ai algorithm machine learning algorithm that you mentioned so the first question is batteries mentioned that you have now a set of 100 objects and people ask is that enough to train an algorithm if not is there any other alternative so you can start building smart algorithm based on the simulation for instance and then i have the third second part of this question which is interesting is that who is in charge of doing this is that is this work done by astronomers and how astronomers qualify for that so multiple questions uh maybe i actually should answer the first one and they jams after if you want go ahead uh we have our i.t team that are that have been working on the a algorithm of the vasco project in the citizen science projects so they are in charge of developing the ai algorithms however we still have that ai is still in training and is going to be in training for for a while while we are running the citizen science projects what was the first question i forgot so the question is is that enough to have android objects for the training data and if you can run some simulation play with a simulated uh images for instance i prefer to wait and collect more objects for the citizen science problem they are not going to disappear the candidates can only disappear once what about you james what was the the state of uh development at lsst this is a big this is a big area because it not just for et work but it's of course for everything um we need to know how many binary stars we'll see and how many white dwarfs we see and so doing really careful simulations really massive simulations of what our we think our galaxy is made of our best guess and then on top of that trying to add this sort of time components into it since we don't have a decade survey to attach to these simulations we're having to make a lot of guesswork so trying to estimate how many of these flare stars we should see is for example is a big is a big effort um so we do have um i mean within the lsst project we have science working groups that work on this um there are whole phd theses being written on these kinds of simulations both for asteroids and for flare stars and for everything so there's a lot of work going on before the telescope even turns on because we know that the algorithms are going to be the key to our success for our for developing the science uh are there enough training samples uh no and i think that's that's the issue when you have a sample of one or a sample of nothing when we're looking for something that's new there's no training data and so on the one hand we have the ability to search for everything we know and we can develop training samples for everything we know but if we're looking for the unknown unknowns as we call them we have no training data and so there's different machine learning approaches to look for those kinds of outliers there's a whole field in computer science that looks for what is rare and totally unique in a data set um and so there has been the beginning of partnerships between computer scientists and statisticians and astronomers so i think beatrice and i are both like proficient programmers but we're not computer scientists um so um our our job is to understand what the cosmos are telling us we we do need computer scientists and statisticians and data scientists to come tell us what are our data telling us and and we'll help figure out what the story is of the physics that's going on that's why i think job security is good even though we're going to automate the telescope and automate the software we need the humans to tell the story so yeah just know one more question says and then i think yes so that's a it's a related question so the algorithm are not fully ready especially you mentioned that they are trees as well so aren't you afraid of being overwhelmed with data or events in the future and people say why don't you ask for help do you have a citizen science program by chance well yes we do have a citizen science project i can't even show if i if i can find my slides um so i can show how it works can i share something yeah go ahead okay um [Music] um did this work no yeah no yes so let's see you can yeah we can see so we have a web page where we actually have both the ai is implemented here and a classical citizen science project where you can actually where users can help um to classify and search for the vanishing stars of course we can i hope that we are going to be overwhelmed with candidates we actually are already getting a lot of candidates um from the citizen science project i checked today we have 75 000 classifications and now we will need help of uh probably students master students to look through all the candidates we're getting because um it's a lot of work so yes we are at risk of being overwhelmed with candidates and i but i hope that this is something that one can solve right in some practical way so if people are interested they are welcome to go to this webpage i will stop sharing the screen okay well uh lamentably we can't take all your questions but you've gotten some information from uh our presenters about where you might go to get more you just saw the link for beatrice's uh citizen science project and uh james said earlier you should go to lsst dot o-r-g and check out that uh that instrument there uh we have time well we don't but we will take it anyhow we have time for one more uh if you have a lightning question and lightning doesn't refer either to the voltage of the question or uh to the speed of it but sort of the latter this is a question i hope that you can answer in just a couple of sentences each of you and that is something to return to something that we have talked a little bit about but you know just a broader picture view of this some people think that you know these these capabilities to peruse big data sets as demonstrated by what you all are doing here oh yes simon reminds me we will get to the poll results we'll get to the poll results but i didn't want to do that first because as we is our normal behavior here if you have to not give the result that most people did you know you will be excised from the viewing capabilities here let me just finish this quick oh all right no let me finish this question i think it's there all right so we see the application of big data to a very very interesting problem but some people have said that in the future the future being maybe 10 years from now right it will be the case that not just astronomy will be done this way but that essentially all research that involves observations will be done this way for example finding the cure for cancer or why some people are allergic to peanuts or whatever you have these huge data sets many of them are available right now and people who have a 5g connection and a cell phone will be able to with very little effort analyze these data in mind maybe find the cure for cancer without knowing the first thing about medicine or cancer so my question is if somebody in the audience wants to study astronomy right are they choosing the wrong thing or should they just study computer science james um it's an intriguing question uh to keep it short i think there's always going to be a market for astronomers i think to go back to your analogy of the cure for cancer i'm not i'm not that kind of doctor so i won't i won't project what the cure cancer will be or how it will be discovered but i would wager that um somebody on their cell phone the citizen scientists might find the data that reveals the cure of cancer but we need doctors to tell us that this indeed cured it and so to extend that back to astronomy we have many of these citizen science projects like beaches like the zuniverse project and the lsst program will develop more i'm sure this is going to be hugely influential and if people are interested there's no better time to get involved as a citizen scientist as an enthusiast the data is overwhelming already which is great but we still need the physicists and the astronomers and the specialists who are going to spend their careers hopefully like us going and studying these and explaining what we see because we're going to find the unexplainable until we get that magical piece of follow-up day that tells us what the answer is okay so the astronomy undergraduates shouldn't shift their majors yet hey what's your answer to this um i i basically agree completely with james on everything he said i mean there's always going to be some it's our work as astronomers to try to understand what we have seen and i kind of more believe into citizen science than i actually believe into the ai if i may say like that this might sound quite heretic but my feeling is feeling is that all these automatical things manners we will have to analyze data will be used to do things faster but not better necessarily you still believe in human expertise okay all right frank can you show the poll results i think lee will will do that thank you okay here you go searches for billions of possible 43 48 looking in our own solar system is not terribly popular which surprises me by the way because you could you read them because oh not everybody can see it i'm sorry okay what's the best way to search for et possibility one search the observable universe of billions of possible et candidates such as we've been talking about uh here today 43 answer two search nearby star systems for a radio signal or flashing lasers 48 that's the the leader here but only by a little bit that's if you will conventional study and the last possibility search for technologies in our own solar system maybe some probes that the klingons have sent long ago nine percent so uh well okay that this that's that's okay it sounds like people are don't think that we're barking up the wrong arboreal fixture with the kind of efforts we're taking now to find the aliens okay i'm going to turn this back to simon steele and thank you all for your attention well thank you thank you seth thank you so much uh beatriz and and james um uh seth master moderator uh rebecca frank and lee in the background from which uh without whom this would never have happened um i i of all the fascinating things have been discussed tonight i i really like the idea that even though we've got new technology coming into play with amazing new telescopes we're still turning back to look at 1950s photographic plates and even the glass plates from the harvard collection that shows that there's there's there's still data there's still science to come out of these amazing um uh history of observations and i'd even take it back further the fact that this is just a continuation of the oldest human tradition of looking up into the night sky to search for meaning and to search for stories in the stars it's it's just fabulous so um thank you everyone also for joining us from around the world where i got just a few of the countries india pakistan portugal hungary canada england ireland and scotland colombia sweden the philippines and turkey and from the u.s we've got virginia montana new york and even somebody from san jose so so the locals are tuning in as well um much more important than winning the nobel prize you all get or you both get a seti talks mugs and patches so congratulations on on on the most prestigious award frank is showing up an example there on the screen um city institute as seth mentioned already is a non-profit research and education institute these events these seti talks are brought to you for no cost we do very much appreciate donations and we would like to thank those who have donated during the talk tonight thank you so much these donations really help this program uh this free presentation um happen um please do go and visit our website uh you can sign up for monthly newsletters and keep up to date with all the work that's going on at the seti institute all the research and also city talks and other presentations we're going to be doing throughout the year um and with that i will thank the speakers again and thank the crew again and say see you next month on next week next month seti talks thank you very much

Channel: SETI Institute

Views: 50,123

Rating: 4.7048345 out of 5

Keywords: planetary science, dps, AAS

Id: 7qLnrzcqvpE

Length: 75min 40sec (4540 seconds)

Published: Wed Apr 21 2021