Instant Genius - Inside the February issue with the BBC Science Focus team

Episode Date: February 21, 2021

In this episode of the Science Focus Podcast, we chat through the February 2021 issue of the magazine, which is on sale now. Managing editor Alice Lipscombe-Southwell opens the episode by telling us why it's so important artificial intelligence learns how to tell stories. Next up is editor Dan Bennett, who tells us about the world’s first airport for drones and flying cars, which is opening in Coventry, UK. Finally, commissioning editor Jason Goodyer tells about the latest developments in the study of dark matter. Let us know what you think of the episode with a review or a comment wherever you listen to your podcasts. Subscribe to the Science Focus Podcast on these services: Acast, iTunes, Stitcher, RSS, Overcast Listen to more episodes of the Science Focus Podcast: Hannah Fry: How much of our lives is secretly underpinned by maths? Robert Elliott Smith: Are algorithms inherently biased? Bergur Finnbogason: Project Discovery and its search for exoplanets Ritu Raman: Can you build with biology? Robin Ince: Inside the mind of a comedian Finding the fun in science – Dara Ó Briain Hosted on Acast. See acast.com/privacy for more information. Learn more about your ad choices. Visit podcastchoices.com/adchoices

In a place like Los Angeles, people don't stop being who they are. Writers, thinkers, creators, people with stories still unfolding. That spirit lives on at Kingsley Manor, a community shaped by individuality, creativity, and lives well-lived. So when the conversation turns to what's next, it isn't about stepping away. It's about continuing the story. Explore your options at kingsley Manor.org, a nonprofit month-to-month senior community within the Front Forge family. This podcast is sponsored by name, audio and focal.

Streaming has made music more accessible than ever, but true listening is about more than ease. It's about quality. British audio experts name audio, alongside French acoustic specialist focal, combine handcrafted tradition with cutting edge innovation and high-end materials, delivering digital precision with analog warmth. So you can experience exceptional sound at home. Music just as the artist intended.

Visit name audio.com to learn more. You're listening to the Science Focus podcast from the BBC Science Focus magazine team. With the UK's best-selling science and technology monthly, available in print and in several digital formats throughout the world. Find out more at ScienceFocus.com or look out for us in your app store. And welcome to the Science Focus podcast. I'm Sarah Rigby, online assistant at BBC Science Focus magazine. With me today, I have editor Dan Bennett. Hi.

Managing editor Alice Lipscomb Southwell. Hello. And commissioning editor Jason Goodyear. Hi. We're going to tell you all about the February issue of the magazine, which is on sale now. So first up, Alice, what are you going to tell us about? I'm going to be talking about why we should teach artificial intelligence how to tell stories. Now, this is quite an interesting topic.

We interviewed Lara Martin, who is trying to do that very thing. to see why it's important that machines and computers can tell these stories. And she said this could really help us with things like Alexa and Siri because you can think of everything's like a story. So if you were trying to do a birthday party or something for your child, you could say, hey, Alexa or hey Siri, I'm going to do a birthday party for my child. Can you help? And then it could sort of create a story about that.

So say all good parties start with cake. So go to the shop, buy yourself a cake. And while you're there, maybe get some balloons and some decorations and some wrapping paper, these are all the things you're going to need. And so that means that then, you know, Alexa and all of those other sort of personal assistants would be a lot easier then. Well, they'd work a lot better perhaps because they'd have this story they could follow to be a bit more useful in your daily life. So, yeah, that's what she's working on. And she's trying to combine a couple of different

techniques of doing it. Now, there's a modern technique where it takes a load of stories that you're and sort of run them through an algorithm, and then it starts looking at patterns and things like that. Now, this modern technique is really good, but it can just eventually turn to gobbledygook. It will start doing a story and it'll just go, bleh, and spit out nonsense towards the end. And then there's an older technique where the actual researchers

will create plot points for a story, and then the AI will sort of work its way through those plot points, so it ends up being a lot sort of better, and it can follow the story a lot better. But, you know, that is a lot more effort for the researchers. So she's trying to combine those two methods, to make these AIs tell amazing stories. And her ultimate thing that she'd really love to see,

but she doesn't know if she can do it or not, is to make Dungeons and Dragons and be played by computers. Wow. So I'd just like to go back and talk about those two different methods that you mentioned. So one of them, you said, was the more modern one, was you feed in some stories, and it tries to figure out from that how to tell the story. And from the other one, you give it a bunch of plot points. and so does it basically

decide on its own combination of those plot points in the older method? Does it basically just say, oh, I'm going to start with this one and then I think I'm going to go that way, that way, that way? Yeah, so it'll sort of follow those plot points itself and decide which way to do it to make this story. And she said it's one way of thinking of it

is almost like one of those big online multiplayer games so it's creating all these little plot points and following the story in that way. There's not as much focus on sort of the grammatical points of that as well. but it's just this creating a story that the AI is working with people to do. Right, and in the modern one, that sounds quite familiar. You see them online sometimes people saying, oh, I fed an AI a bunch of scripts from whatever TV show

and it made it come up with its own one. But it seems like the problem with that is that it struggles to form a coherent story, right? So, you know, how could you get an AI to identify each character and understand how to follow each character through the plot. Would you say that that's a benefit of the older method as opposed to the newer one? Yeah, so I think the older method, you know, that can sort of create the plot points a bit better, perhaps.

But then, yeah, the modern one, it will sort of, you know, start doing it quite grammatically well and then it'll sort of lose its track along the way. So, yeah, I think that would be good if you could combine them both, and that's what she is trying to work on at the moment. because then you'll be able to use the newer method to make the computer tell stories about anything but then you can also use the older method

to make sure that it's coherent and the plot points are all involved in there. It's really exciting. So you can kind of, there is a sort of AI Dungeons and Dragons out there right now, isn't there? Yes, there is AI dungeon already that someone has created and I think she did say that this one can end up,

you know, you take turns with it, like those old style adventure games, and that is using the new methods. So, I mean, I don't know if you've had a go on it, but I have done it. And it's quite good. So as long as you get it just right, you know, it can sort of give you an answer, but sometimes it's a bit like, oh, good try, you know, you could have done better. But you can't stretch it. You can't, you can't start going too creative and, you know, introducing pet dogs and

no, it just gets a bit confused. Yeah, it just gets confused when you do that. So, I mean, that's what she said, her ultimate aim is if you could, you can get an AI to be able to create these stories and be creative with it as well. Yeah, I was just wondering, like, how far away are we from these sort of things, making even like basic little, little stories, you know, little short children's stories or something like that rather than something more complex. I just wondered what sort of level they're at at the moment.

I mean, they are quite very basic stories they can tell at the moment. I mean, Lara did talk about if there's somewhere we could combine it, so, you know, for screenwriting for films or something, then the computer could sort of give you the ideas, and then you could run with it, and then you get the sort of issues of, oh, is that the computer that's got the copyright of that, or is that the human that's got the copyright of it?

And then you get into all these philosophical arguments. So are we saying that an AI could ever be truly imaginative? Like it could ever come up with a brand-new idea on its own? It's not just reworking and regurg. agitating ideas that we've put in? I think it could be potentially. Lara said that when she was working with these AIs, she had sort of was to do occasionally stories,

and then she said, they are, oh, you know, what's the next sentence? And a lot of the time it was just nonsense that came out. But once it said about, how about a horse that was a lawn chair entrepreneur, which is amazing. And she just thought, well, if you then decided to run with that and someone could then work on it. And, you know, it's quite an inventive thing to think of. I mean, okay, maybe 100 times you might get a nonsense idea,

but then now and again, you might get something really unusual that a human would never think about. Great, thanks, Alice. So, Dan, what was your favourite thing from this issue of the magazine? Yes, in this issue, I got quite a strange launch came into my inbox. And I didn't quite know what to make of it. Essentially, the message was that they were going to launch the world's first

they called it an urban airport but essentially the world's first airport for drones and flying cars in Coventry and they were doing it in a few days and I just couldn't it just seemed so far so far fetch because I mean we don't

see any drones really in the skies unless you you head to a park or something and then even then there's actually fairly strict guidelines about you know flying them next to people and how far you can fly them from houses and then, you know, flying cars. I mean, how long have been waiting for those. But no, it's a totally legitimate and real idea that seems to be taking place this year.

So what it is is essentially we, you know, we're seeing right now, every curb pretty much has a van on it, parked on it right now somewhere with someone doing Amazon deliveries or, you know, whatever online store you use. And, you know, not just because of lockdown, but it's been on the rise where there's just so many deliveries now. Like I live in an apartment block and the ground floor is just always piles of boxes. So there's coming a point at which it's going to start making sense to use drones to make those deliveries, not perhaps, you know, all the way from the Amazon fulfillment center all the way through to my address. But perhaps, you know, a van could take a Amazon package to, you know, a local center just

outside of Bristol and then a fleet of drones can whiz them over to their destinations. And it wouldn't maybe be my house, but, you know, perhaps it might be down the road. There's a little center where I can go and pick up my package. And so these people behind this, these airports have. kind of seen this future sort of rearing its head. And they've gone, well, for that to happen, we need infrastructure. And so I got to talk to the CEO and the man behind this idea, Ricky Sandy. And his metaphor is like, you know, if you want trains and railways, well, you need train stations and train depots. You can't have the one without the other. So we need the infrastructure

to be able to support what he sees is an inevitable thing of drones sort of taking around the squads and doing our deliveries. And indeed, he cited where this is already happening to a degree in places like Ireland and South Africa, where particularly in sort of remote areas, people are, you know, able to now get a delivery via a drone. The flying car bit, I think, is probably further down the line, but interestingly, they've secured some considerable investment from Hyundai, a Korean car manufacturer, because, and they approached them because of this, they have developed a, I think it's two-seater, maybe four-seater, effectively a flying car. I mean, it's basically a very, very big helicopter. And so, you know, if you're going to have those and

they look pretty cool, I'm sure I'd love a... ride in it. It might be a bit scary. You need infrastructure to sort of support these. And what's quite cool about these is they're these little sort of circular hubs that you can drop in somewhere. And they are technically off grid. So they have their own hydrogen power generation.

So you don't even need to sort of wire it up and, you know, get electricity to it. And on top of that, if it's off grid, using hydrogen power generation. and solar power, then it's also pretty green, pretty clean. So you can charge up your drone there, fly it over, and crucially offset that pollution, air pollution, and congestion that's been caused by all these deliveries. So it sounds like this could solve the, I think I've heard of it called the last mile problem in home deliveries.

So the bulk of the travel involved in home deliveries is apparently, well, it's quite easy to make that environmentally friendly. So, you know, you can have electric vans or hydrogen-powered trains or lorries or whatever, and that can get your product most of the way. But it's the last mile from the depot to your house, apparently, that is the bit that's hardest to make environmentally friendly because it's, you know, there's vans driving around city centres and all day long stopping everyone's houses and stuff. So it sounds like these potentially hydrogen-powered or electric drones could be the solution to that,

finally making it easy to do that very last bit of the delivery. It's tough to say right now because so for that to properly work, and it would be great because I say it is one of the biggest issues that cities are facing that last mile problem. but the thing is we don't so so the stage that we're at is that the civil aviation authority and different city authorities are in in conversations with lots of different invested parties to figure out where these drones are allowed to fly because obviously if you've got all these boxes flying over people's heads there's a definite safety concern there and there's also you know a visual pollution element like do we want us drones, littering, cluttering up the skies. So that's all in negotiation. So what we don't know is how close they can get to where we live and where we are. So it may be, and this is what Ricky,

I want to sort of put this question to him is his answer as well. Likely there's going to be little depots in places. And, you know, like we have now with like Amazon Dropboxes, we might go there to where our drone drops off our packages, you know, and they need to be close enough, I suppose, to make it all make sense. But he was kind of quite reserved to say, you know, it's not that I think this is the only way.

It's just that the more we can offset sort of problematic congestion and air pollution with ideas like this, the better chance we'll have, because the other factor is, So to make all these deliveries green, we need electric vans, which are in, you know, there's quite a few of them now. People are buying them. You know, the royal mail vans are electric. You see them around.

But they all need to plug into the grid. So there's a growing pressure on the UK's energy grid. And who's to say how we're going to deal with that in the next two years is particularly with the, there's a, you know, a supposed 2030. ban on new petrol electric cars. So we don't know how that's going to happen. So that's why it's so key that this solution that they've come up with is actually off-grid, because you can implement it without putting any extra pressure on an already sort of strained system. So how noisy will these drones or electric planes be? Because, you know, cities are quite noisy anyway. I live near a busy road and the sound of the traffic going past really gets me down. And then I think drones, when you sort of see

and they are quite noisy. You know when they're there, they're whirring over your head. So if you've got dozens of these sort of flying about all over the city, is that going to be a problem? Yeah, right? Especially if there's like a flying air taxi,

like a huge drone buzzing off. I asked Ricky this, and his answers said, so they've done a lot of testing, and actually they've worked with NASA on this one. So that's quite interesting because, you know, it says to me that the flying car element

and the airways element of it, is not so far-fetched. But they say that essentially the urban airfare take-off platform will be a couple stories high, higher than where any people are living. So if you're near an apartment block, then it would have to be higher than that. But if you're just on a level house,

then it would be a couple stories above you. And they say that in the midst of the typical urban congestion and the noise, that once you're that high, the wind noise and everything else is enough to sort of cancel out the noise of the drones. Don't know how true that is. Obviously, we can't have heard it yourself. I've definitely agree with you that drones are pretty noisy. And even when they're flying quite high up in the air, you do hear them.

But presumably these won't be too near where people live. I'm getting, I don't know. but for instance the one in Coventry is you know it's not actually near any city centres it's quite near an Amazon fulfilment centre they've got a massive one there and then it connects all these different places so I don't think I don't envision it becoming common

in you know residential areas but interestingly I was a little bit more interest actually in the idea of it as sort of instant infrastructure so Ricky was talking about how you know if you went to say somewhere that was having an epidemic and it had poor infrastructure. You know, you couldn't get medicine to people via roads or, you know, traditional means.

You could drop one of these in and you could fly medicine to people all around and very quickly react to a local epidemic. And that's quite cool. And he has said that he has had sort of interest from the UK, the UK Defence Ministry. So that's pretty interesting. I was just wondering, when you said back then it's in Coventry, why Coventry? Wouldn't it be make more sense for a city like Birmingham or London or, you know, Edinburgh or something like that? It seems like quite an obscure place. Yeah. So there's a couple things there. So one is, I think they had a partnership with the, with Coventry, the city, because it's the city of culture this year. obviously with everything that's been going on,

who knows how that's going to pan out. So it seemed like a good timing and place to kind of get eyes on the project. So they recently got government funding, so they've got a few million from the UK government. They got some money from Hyundai, and they're still looking for more investors to take the idea further.

So that seemed like a nice showcase to debut. And Coventry as well, I think, you mentioned Birmingham, but it sits kind of near the middle, not really, but, you know, Rick is described it as the middle of country. It is near Birmingham. It's near quite a lot of places. There's one of the UK's biggest Amazon fulfilment centres near there,

so it makes sense in that kind of test scenario. Also, there's a lot of engineering and motoring industry up there. J.L.R. Jagger Land Rover is in that part of the UK. So I think it was interesting because Ricky himself said, you know, I could put this anywhere in the world, but, you know, I'm a British architect. I went to university here. I grew up here. It's quite cool that we can try and help and perhaps create jobs in this area in Coventry to make it work. So he seemed, you know, he was pretty proud of the fact that he was able to secure a spot for it here, you know, the first one in England. And indeed, you know, in the next, it does seem like it's a thing that's going to happen, these drone highways. And certainly, it seems like something we're very good at is aviation and certainly small satellites and things like that. So it seems like we're kind of quite well poised to be sort of pioneers as a country in this area. All right, Dan, thank you very much.

So now we're going to finish off with Jason. Jason, what are you going to tell us about? I'm going to tell you about the latest development in the one of the biggest mysteries of modern physics which is dark matter so just by way of Congress it's this it's sort of obviously something as somebody who studied physics I've been aware where I've ever since I started studying it but it's kind of crazy to think that it was first sort of proposed almost a hundred years ago now

by Fritz Zwicki the Swiss astronomerate, obviously he found the discrepancy in the motion of the stars and galaxies and their gravitational pulse. So basically, they're moving far too fast to be held in the specific positions that they are in galaxies. And that was in the early 30s. So there's been all sorts of different candidates for what this discrepancy is. Is it, as Fritz himself thought, dark matter, something that we can't see, something that isn't luminous but has a gravitation? effect. There's been all sorts of different proposals for these. Chief among them, as far as, I mean, I can remember it, has been wimps, which are weakly interacting massive particles.

So there's loads of potential wimps out there as well, you know, they come from all sorts of different ideas, like super symmetry, like heavier particles that currently exist. We're known to exist, like electrons, heavier electrons. And there's all sorts of different. of clever maths that people have done to try and figure out exactly what this is. But so far, nobody's really got anywhere. There's been all sorts of massive detectors for these WIMPs, and nobody's really, it's basically been the case of no dice. So there's this place, this big, massive, so what they use is like huge tanks of Xenon. So these particles, Wimps, obviously, by their name, they're very shy. They don't.

interact with things. So the theory is here if we have a massive in it, they're in like super pristine clean environments underneath mountains. This particular one Zen and one T is in the upper lines in central Italy. So they try to isolate them as much impossible and it's a huge area and I've tried to detect just one interaction with one of these wimps and this enormous tank of Zenon and you'll get basically it will just whack into one of these things and it will spit out some photons, basically, and then they can detect that. So, as I say, this has been done all around the world and all sorts of Antarctica everywhere, and they haven't really,

I mean, there have been some results, but there's nothing definitive. But in this, I think it was last summer, in this particular detector, Zeran 1T in Italy, they found a really intriguing interaction. And so at first, some people were saying, oh, it's just that the system's contaminated, it's some sort of problem due to that, or it could be, like, yet another, can't relate for dark matter axioms, which also, as have yet, they're still hypothetical,

talks like, oh, it could be an axon. Well, yeah, it could be, but there's a new more intriguing option that these guys in the Netherlands have thought of, and it's an interaction between something called a dark boson. So we can get really into this now, we really shouldn't. So basically, everyone's instantly going to ask, oh, hey, is that like a Higgs boson? And it's like, it's a boson. So there's many different types of boson. So basically, they're particles that carry forces. So the most commonly known one is, of course, the photon, which carries the force of the electromagnetic force or electromagnetic fields, etc. So it's basically that. It's a part of them that carries a force. Sorry, Jason. So a photon is also a

known as a particle of light, isn't it? That's correct, yeah. So light being part of the electromagnetic spectrum, photons, it's boson. It's the electromagnetic radiation. Yeah, so we see electromagnetic radiation as humans as light. So yeah, it's a particle of light.

So we theorise that these signals that are cropping up is something called a dark boson. So what does that tell us, if that's the case, what does that tell us about the nature of dark matter? What does it tell us that we didn't already know? Okay, so dark matter is just basically, well, you know, it might not even be there. Who knows? It's what we use to, for example, there's an Israeli physicist that since the 80s made this thing called modified mutalian dynamics. Where he thinks actually there's no such thing as dark matter.

And it's just that Newton's laws of motions work differently on the scale of galaxies. But dark matter, so it's basically something that you cannot see with, you detect using light instruments, electromagnetic radiation. So it has a gravitational effect. It has a mass and it influences things around it, but we can't see it. So that's why it was originally called dark matter, because without going too much into it, basically the brighter, the heavier a star,

the more mass of stars, the brighter it is. So Zwicki's original experiments were based around sort of comparing the gravitational effects he thought, he would see by the motions compared to the luminosity. So that's why it's called dark matter. So as it's dark matter, you can't detect it using electromagnetic radiation of those sort of detectors.

So then what could it be? So, for example, some people think it could be primordial black holes. These are black holes that were created in the early universe. They're like tiny, I don't know, marble-sized black holes. And you can't see them with your usual telescopes, but you can see the effect, the gravitational effect that they have. So is it then? So that's where the WIMPs come in,

because they're very difficult to detect. but if they're, and they're very light, but if, well, they're massive, so they have a mass, but it's not significant, like a black hole for instance, for instance. But obviously, if there's enough of them, they can have an effect on the gravitational pull within a certain system. So there must be loads of dark matter about then if you're saying that it has this massive effect, and these dark photons, if they're creating dark matter, there's just millions of them zipping about all the time, surely.

Most of the eye, the candidates for a dark matter, excuse me, are hypothetical. But yeah, it's 80% of the known mass in the universe is unaccounted for. So that's, so when your dark boson comes in, it's just like a really, it was an absolutely new one of me, and it's really quite weird. But basically there are particles that they're affected by gravity, but not by other forces. So the guys researching it, say, for example, they could, like normal matter, ordinary matter,

clumped together to form massive objects, but you can still pass completely straight through them because they don't interact at all with the electromagnetic force. So you could have like a huge star made a dark boson that you could just walk straight through. Well, not walk, but your flies are in. Yeah, which is very odd, isn't it? And they say that these things can clump together and theoretically clump together in such massive quantities that they could be as big as supermassive black holes.

So the Event Horizon Telescope, if you remember, that took the photograph of the first photograph ever a black hole. So there's a new group of researchers that are saying, now, thanks to that, that technology and that ability that we now have, we might be able to detect these dark boson stars because they won't suck all the light out of the environment in the same way that a black hole does.

So if we see a certain, you know, large but dark area of mass and strong gravitational pull, but we don't detect this distinct signature of a black hole, then it potentially could be one of these. dark boson stars. Wow. So if we could get dark boson stars, does that mean we could get dark boson planets, dark boson solar systems? I mean, you see, yeah, I mean, that's, I've read about stuff like this before, right?

And it seems possible, doesn't it? But I don't know. But, like, obviously, they're going to behave in a slightly different way. So I think in a lot of cases, perhaps the analogy falls down a little bit. So they're called Darbos on Stars, but they're not going to produce nuclear fusion, et cetera, obviously, because we would be able to see them if they did.

So, yeah, I don't know. And I've heard that theory proposed before, that such a thing could exist, but this isn't particularly, due to the difficulty in detecting them, that isn't what this particular group of researchers are looking at. But, I mean, I can't speak for them, I don't know, but it seems like an exciting possibility,

doesn't it? Yeah, definitely. It sounds so exciting. I was going to say, I was wondering, when you said earlier about they used these xenon detectors to find them. Why did they use xenon? Is it something special about that element or not? Yeah, it produces when the things hit it, it produces a certain type of radiation that you

can detect, which they use for this specific purpose. It's quite an old, what, it's called Charenkoff, is that right? I think it's called Charenkoff radiation. but yeah it's quite there's been going on oh blimey I think since about the 90s this idea I can at least remember it when I was studying

maybe they hadn't I think they were setting them up or they were just setting them up but yeah that's one of you it has to be very because you know there things lots of particles can just pass

through everything without interacting with it with anything whatsoever like it happens all the time so it's you have to be very but it's like it's like a waiting game and a size game. They're trying to maximize their potentials to discover these interactions like that.

So that's why they do that. Thank you for listening to this episode of the Science Focus podcast. The February issue of BBC Science Focus magazine is out now. Also in this issue, we explore how your brain creates reality. We meet the people who are creating valuable resources from waste. and as always, our panel of experts answer your questions. Of course, there's much more inside and on sciencefocus.com. Thank you for listening to the Science Focus podcast from the BBC Science Focus magazine team.

We're the UK's best-selling science and technology monthly, available in print and in several digital formats throughout the world. Find out more at sciencefocus.com or look out for us in your app store. This podcast is sponsored by name, audio and focal. The texture and emotional depth of music can be lost through digital sources or poor signal. Name Audio believes you can have digital precision with analog warmth. Alongside French acoustic specialist focal, Name creates high-end audio systems combining innovation with craftsmanship,

so you can listen to music, just as the artist intended. Discover more at name audio.com. In a place like Los Angeles, people don't stop being. who they are. Writers, thinkers, creators, people with stories still unfolding. That spirit lives on at Kingsley Manor, a community shaped by individuality, creativity, and lives well lived. So when the conversation turns to what's next, it isn't about stepping away. It's about continuing the story. Explore your options at kingsley Manor.org, a non-profit month-to-month senior community within the Front Porch family.