 For too long, humans have destroyed or altered insect diversity and environments. Now, insects and humans must find sustainable ways to coexist in one another's ecosystems. To help achieve this goal, Dr. Shannon Olson founded the Echo Network and serves as its global director. This organization brings together a diverse group of individuals who share their knowledge and affect meaningful, sustainable changes in their communities. She is also the special scientific envoy to India within the Danish Academy of Technical Sciences. In each of these roles, Dr. Olson brings decades of scientific knowledge and experience. She studied neurobiology and behavior in graduate school at Cornell University and has since conducted research in three continents that centers around interactions between humans, animals, and environments. Dr. Olson conducts her chemical ecology research from the perspective of a naturalist and an empath in order to better understand insects, their minds, and their behaviors from their perspective. Her studies range from controlled experiments that measure insects' neuronal responses to odors to observations of insects' responses in their natural environments. Dr. Olson is uniquely situated to give us insight to how insects operate and are impacted by changes in their environments, while being mindful of the broader goal of achieving a sustainable balance between humans and insects. Please join me in welcoming Dr. Shannon Olson. Thank you so much for such a beautiful introduction. And I am so incredibly honored to be here to speak with you today at this beautiful conference and this beautiful college in this nice Indian-like weather, so thank you so much. I think that I've probably always been interested in science. It's something that I've always been drawn to, this idea that one can think about the world and observe the world and learn more about our place in this universe around us. And I think that I've also been very passionate about nature. I've always loved going outside. I used to catch fireflies at Northern New York State where I grew up. And when I was able to go to college, I went to a small college, not so different from this one. I learned about this idea of chemical ecology. Chemical ecology is a field that studies how organisms use chemicals to interact with each other and their environments. And everything in our universe is made of molecules. We are made of molecules. This floor is made of molecules. The air is filled with molecules. And not every organism can see. Not every organism can hear. Not every organism can certainly speak. But every organism, plants, fungi, bacteria, elephants, humans, and obviously insects, can use chemicals to interact with each other and the world around them. And this concept fascinated me. So I had the privilege, after my undergrad, to study chemical ecology at Cornell University, which was at the time one of the very few places where you could get a degree in chemical ecology. And I had the honor of studying under two scientists, Wendell Rolofs and Thomas Eisner, who are considered the fathers of the field of chemical ecology. And when I got there, Wendell had been studying the system called the apple fly, which as the name implies, likes apples and feeds on them as larvae. And our big past here in North America. And I went to my other advisor, Thomas Eisner, who many of you might know, wrote a very famous book called For Love of Insects, and his quote graces the website for this conference. And I asked him, Tom, do you have any advice for me on how I can do this research on the apple fly? And Tom looked at me and he said just a single phrase. He said, Shannon, think like a fly. And I had no idea what he was talking about at the time. And so this talk is my 20-plus year course in fly psychology and the things that I've learned about flies and the things more importantly I think I've learned about myself and us as humans on this planet. So to start this course, okay, maybe we should start a little bit more simply. What actually can a fly think about? What is the information that it can process in the world around it? Well, if a fly is in this beautiful valley in the Himalayas where I do my research, it can certainly think about some of the same things we do. It can detect the horizon. It can detect landmarks. It can detect objects in the world like these flowers. We can also detect these. But insects have a number of sensory capabilities that we humans do not have. For example, they can detect polarized light. Polarization which we need special equipment to be able to detect helps give them directionality as does magnetic fields so they can tell compass directions as well which helps them to navigate in the world. Not only that but they can detect visual colors just like us and they can also detect UV cues. So you see here that actually all of the flowers, the buttercups in this picture now have black dots in them. And that's because under ultraviolet light, buttercups have black targets so that pollinators like bees and flies can find where the nectar and pollinate are. They also can detect humidity and temperature cues which give them lots of information of where important objects are including flowers that have lots of nice wet nectar inside of them. And they can even detect electrical fields as well. So they have an enormous number of sensory capabilities that we humans do not have without special instrumentation. And they can in different ways utilize these different capabilities to think about their world. So that's what they can think about. But now they also need to detect things from a distance. Fly's eyes are compound eyes. They actually have many eyes, right? They have many tiny lenses in their eyes. And they have a physical limitation because of the size of those lenses to have high acuity. Unlike our eyes, which are much, much bigger, of course, so we can see things from a distance. They cannot see things clearly from a distance. So a lot of these visual cues are not very clear to them. So chemistry and odors that travel in the wind give them a lot of information to find things from a distance. And that is the world of chemical ecology that I've been working in. So all of these stimuli are important. But now, what does a fly think about? Well, I'm sorry I'm not going to tell you this because I don't know. They haven't told me yet in 20-some years. But we can think about the things that they do in the world around them. And most insects have to, of course, reproduce at some time in their life, and many of them do not reproduce right away, which means they have to sustain themselves long enough in order to be able to find a mate and reproduce. So they also, in the process of finding a mate and finding their sustenance or food, have to avoid being eaten or eating something poisonous by accident. So this concept of love, hunger, and fear as it will really is something that most insects have to think about at different points in their lives. They have to find mates, locate food and nutrition, and avoid danger. And in order to do this, one of the most fundamental things that they have to do is identify what is their mate, what is food, what is danger. They have to identify the objects that represent those qualities, and you don't want to mix them up. Obviously, you don't want to eat something that's going to kill you, and you certainly don't want to run away from your mate. Even after several years of marriage, you don't want to do that, OK? So this is very important for them to identify objects. So fundamentally object identification is probably one of the basic things. Now, how do they identify objects? So let's listen to my daughter, when she was one and a half years old, to hear how we identify objects. That's how we learn about objects, right? We learn from those around us, first from the people that we're living with early in our life stages, maybe our parents or relatives, then from our friends and teachers, and also by just experiencing the world on our own. But many insects, most in fact, and definitely flies are solitary, which means that when they are born, they have no society that is able to teach them. Their mommies and daddies and relatives are generally long gone from this planet. And they have to at least be able to survive long enough to find that sustenance and avoid being eaten. And you can't just go around just trying to eat everything because likely you'll find something poisonous or you'll exhaust yourself in the process. So how does a tiny fly with such a tiny brain, right? They have only about 500,000 to a million neurons versus our 80 billion. How are they able to identify objects without anyone teaching them, which means they're born with this ability? This is something that I set out to try to understand when I moved to India. And since I had been doing my research in the United States and in Europe, I really tried to think about what the differences were in this brand new world I was in, in the Indian subcontinent. And so we started by looking at systems that existed in multiple places. And one of these systems of flies is called the hover fly. They're so called because they hover around, but they're also fantastic pollinators and quite prolific pollinators. They're not as efficient as honeybees, but they are perhaps more prolific in that they can exist on every continent on this planet except for Antarctica. And as you can see from this picture, in many cases the same species actually exists in very diverse habitats. So these are pictures that I took myself in my home base in Bangalore, India, in southern India, at 2,500 meters up in the Himalayas where this picture, the backdrop behind this image, is taken, and even up in the Linnaeus Garden in Uppsala University in Sweden. So myself and my colleague, Karin Nordstrom, set out to try to understand how can an insect identify flowers? Flowers are their food. How can they identify something like a flower when flowers themselves are so different? Right? A rose and a lily and a daisy and a lilac and all of these flowers are so different from each other and definitely the flowers that grow in Sweden and the Himalayas in Bangalore are very different. So how can this singular species be able to identify what a flower is when the flowers are so different? So to do this, we first of all thought about the cues that they could identify. Some of those cues I mentioned to you before, UV, landmarks, odor, temperature, humidity, and we measured them from flowers that these hoverflies were visiting and how we did that was literally by following them around. We went to each of these environments, we followed them around with our eyes and our noses and our ears and tried to locate when they were landing on flowers and as soon as they landed on a flower, we immediately measured, as you can see here, myself and Karin and Pragedesh, our postdoc, we're preparing our instrumentation to record all of those signals from the flower. It's color, it's smell, it's humidity, it's carbon dioxide and we did this across all three climates, about 146 different flowers and we had over a million data points and then we used something called multivariate analysis which is a statistical technique that allowed us to predict what should a flower look like to be attractive to a hoverfly in Sweden or in the Himalayas or in Bangalore. What kind of cues seem to be the most attractive? Is it a certain color or a certain smell or a certain level of humidity? And since the statistics gave us predictions for physical parameters like cues and chemical cues and shapes, we were able to turn our statistical predictions into actual physical objects that we 3D printed. So these are the 3D printed statistical predictions of what hoverflies should like as flower objects in different parts of the world and they look a little weird and that's because we weren't trying to replicate actual flowers, none of these are supposed to be a daisy or a rose. What we are replicating is what the statistics told us should be attractive which is why we have this weird electric blue flower and a green flower which doesn't exist. And then we went out and we planted them, these flowers that had colors and shapes and smells in each of our environments again and we waited and we watched and lo and behold, our hoverflies came to them in each of our different environments and we found some really interesting things about how hoverflies identify things across the world. For example, one of these electric blue flowers was only attractive to hoverflies in Bangalore. The white flower object was only attractive in India but this yellow flower object was attractive everywhere we tested it. In fact, we've later tested it in many, many different places across the world and we have found that it is attractive to hoverflies pretty much everywhere we tested and that gave us a thought that this might be the object that hoverflies are born with the ability to detect as a potential food source. This is what is innately coated in their brain from birth since they don't have anybody to teach them what a flower is and they have to at least have some concept of a flower when they're just born. So in order to test this, we took our 3D models and we broke them up. We took out the pieces of them. We actually worked with artists for this and designers to help us parse these flower objects. We took away the color, we changed the shape, we changed the smell and through several different experiments, we actually exposed our hoverflies to these objects and we gave them a choice and you can see this beautiful little hoverfly. She is just born. She's never seen a flower before. She was born in our lab and she was taken from the soil up in the places where we studied them outside and you can see her making choices between them and by doing this experiment, we found what is the basic idea of a flower if you're a fly? And it turns out that it's a pretty simple thing. It has to have color. It has to have hue in the visible range, particularly has to have some yellow sort of spectrum, at least yellow to our eyes, maybe not theirs. It has to be very bright against the background and it has to have a high contrast against the background, so it has to be very discernible from the background. It has to be symmetrical and it has to have a smell, not just any smell but plant smell, either grass or flowers or many different smells. So this kind of scented hexagon is probably the simplest thing that can identify a flower to a naive hoverfly and then, of course, they learn, they experience their world, they get new information as they go along and then they perfect this template and then they go to other types of flowers, blue flowers and pink flowers and lots of different flowers over time, but their first instinct for a flower is this. Then we thought, well, how good is this very simple algorithm at identifying flowers? So we decided to pit it against the machines, right? We are in the world of artificial intelligence and as you probably know, one of the branches of artificial intelligence is machine learning and particularly a topic in machine learning is deep learning. These are neural networks that are based broadly on the way the visual cortex processes information through several layers of neural connectivity. So these artificial neural networks, we then used to compare how they identify images against how our flies identify images and since these neural networks are very commonly used nowadays for facial recognition, recognition of all sorts of things in your phone, for example, in your photo program. So what we did was we compared our little simple fly algorithm of just color and brightness and contrast and smell and we gave these two algorithms and the neural networks objects that were not flowers and objects that were flowers and we did give them smell. Now, the computer couldn't smell, okay? So what we did was we coded smell through into the data itself. So that was a parameter in the data and then we said, okay, how well does our fly do against the most advanced neural networks that we're using today? And it turns out pretty darn good. Our fly algorithm actually was about 90 some percent, 95 percent almost accurate, no matter how many images that you gave it whereas of course the neural networks needed to learn and they needed to experience images. So they did become as accurate but they needed quite a number of images to do so and of course we know if you do any work in computer science you're usually giving these programs tens or millions of images to learn from and not a thousand like we did. But what was more fascinating to us was the time that it took, right? Because our system didn't learn, it was processing in a very, very fast rate every single time, whereas of course the number of images increasing meant that these neural networks were taking longer and longer. So our tiny little fly brain can do something that currently artificial intelligence is not yet able to do and that's because they have figured out how to do this through the learning of evolution in order to help them identify flowers very, very quickly as they're flying through space. Now, then comes the big question, how can we start thinking like a fly? This is the question that Tom asked me to think like a fly and this is really difficult, right? They are so different from us and we really can't put ourselves in their world. We can do our best to follow them around but we don't know what it's like to experience the world at their size. We don't know how to see the world as they see it. We don't have the sensory capabilities. We don't have sensors small enough to put on the flies back and follow it through this Himalayan meadow. So because we couldn't put ourselves into their world we decided to put them into our world using a technique called virtual reality. So in my lab we created a VR for flies, okay? So much like we would put on VR goggles and we would look at the Sistine Chapel we put a VR arena around our flies so we couldn't put goggles on them that was a little too small. So instead of goggles we had a panoramic display that wrapped around them in 360 degrees because a fly can see in the back of its head, right? And we also had the fly there in the center and it looks like it's stuck to a needle. It's actually glued with a tiny bit of super glue just to its back to hold it in place, harness it long enough to do the experiment. We could remove it afterwards and that's because we didn't want the fly to fly out of the world, of course. We gave it wind and odor. I mentioned odor cues are really important for insects to identify things from a distance and we needed a camera to film it. So now I'm gonna show you how this VR actually works in practice and the flies that we used for this are not hover flies that I've been talking to up to now they're actually Apple flies. The Apple flies that I used way back when at Cornell with Tom and Wendell and that's because they really love apples and they love apples and trees. So we knew that this was a great experiment to test well if our VR is working they should be going to Apple trees just like the real flies do. So here you can see the image that the fly would see of course in reality it's wrapped in 360 degrees which is why it looks a bit distorted. On the right of the image you'll see the trajectory that the fly would be flying if it were actually in the real world which of course it's not but this is in meters and the little fly is down at the bottom and it's a silhouette of it and how we work the world is by observing how the fly beats its wings. So it decides where it wants to go if it wants to go left the fly will beat its wing harder on its right if it wants to go right it will beat its wing harder on its left much like you would if you're swimming and by measuring those wing beat amplitudes we could then turn the world in response much like it would be if it was playing a video game, okay? So this is what it looks like in action. So here you can see this fly flying in the world and it's flying towards this apple tree and the cool thing is that it's actually weaving in and around this virtual apple tree and if you look very, very closely at the fly and you'll see this again as the system resets itself and it goes again to another tree actually when it approaches the object at close range it throws out its legs in front of them and a fly will do this for two reasons one, it's about to crash into something or it wants to land on it and this was the surest information we had that these flies actually felt that they were in some sort of physical world and they thought they were going to land or crash into something and this allowed us to do some really cool experiments and as best we could from their point of view one of the things we found was the reactive distance of flies, okay? Now reactive distance is pretty simple it's not how far away you can see it's how far away you can identify just like in a doctor's office you can maybe see those tiny letters but you don't know what they are we could tell that a fly could perceive a four meter average size apple tree from an apple orchard from about 24 meters away which is the first time we've been able to figure that out because it's very hard to measure this in the real world more than this we were able to figure out some really cool things about how they perceive the world like this image the image on the left and the image on the right are slightly different the image on the left is actually a cardboard cutout of a tree in the virtual world and the image on the right is 3D they could tell the difference they could tell the difference in subtle lighting so we actually had to make the world into Tatooine with two suns on either side just to make sure that the lighting was always symmetrical so they weren't always going to the left we also found that they could use depth perception in flight which hadn't been known up until that point because doing depth perception while you're flying is a pretty complicated thing to do and for me, one of the most fascinating things was what happened when we gave them nothing at all when we just put them in a blank world of grass and sky we couldn't put them in a completely blank white world because that was so foreign they would just spin in circles so we had to give them something called optic flow which is a pattern that allows them to orient themselves and we did this by making a grass pattern and a cloud pattern in the sky and we put our flies into this world they actually often went to some point in the horizon where these clouds had a specific pattern and we have no idea what they saw in these cloud animals if it was a fly cloud they saw or an apple cloud or something else but it was fascinating to us because even in the absence of any objects at all they were trying as hard as they could to make sense of the world around them and they also were super quick at adapting to their world when we first started our experiments Pavan, who is the graduate student who helped build this arena he actually made a mistake in the code okay, so he actually mixed up the left and the right so actually when the flies thought they were supposed to turn right they would actually go left and when they would go left they would go right I don't know if you've ever tried to drive a car or a bicycle that had reversed directions but if any of you have it's really hard for our brains to switch our left and the right it takes quite a while to practice and you often fall over if you're on a bicycle these flies did it in two wing beats which is then they fly about 200 hertz a second 200 times a second that's so fast that we didn't even notice so for the entire month we were flying these flies in reverse and they didn't even tell the difference because they're that quick at adapting so to the best of my ability right now here's what I think to think like a fly is to observe the world as it is to try to make sense of it and to quickly adapt to changes in your environment that's a fly's way of thinking and it's very different than the way we humans think okay, we humans are prediction machines our 80 billion neurons are not just thinking about what's happening now we're thinking of what's gonna happen in the future and we're constantly using that information to process what is going on around us and this is very very important it helps us for example to find things in a cluttered environment very quickly which is a big advantage we have as humans and I'm gonna show you how awesome this actually is okay, so I'm gonna play a little game with you okay, I'm gonna ask you to find the toothbrush in this image and I'm going to only give you a second to do this all right so because that's all that the scientists at the University of Santa Barbara gave their subjects so you get the same amount of time you're gonna see a bathroom, a normal bathroom cluttered and you're gonna have to find the toothbrush and tell me what color it is now keep it to yourself because it's a little bit of a quiz okay so don't tell the neighbor next to you are you ready, you ready here we go okay, okay now here's the quiz how many of you saw a white and green toothbrush raise your hand, very good how many of you saw a blue toothbrush raise your hand, all right, not too many how many saw a yellow toothbrush raise your hand, congratulations to the people who found the white and green toothbrush indeed there is a white and green toothbrush there also congratulations to the people that saw the giant blue toothbrush Right behind it. This is how we perceive our world. We don't expect to find a giant-sized toothbrush on the kitchen counter, because nobody's going to be able to brush their teeth with that. So we make predictions about our world constantly. And as it's been said by Christopher Chabris, who actually wrote a book called The Invisible Gorilla, what we pay attention to is largely determined by our expectations of what should be present and not what is really there. And that is fundamentally different than how a fly detects its world around us. And I think that while this ability that we have is one of our greatest strengths, like many things, it also has become our greatest weakness because we expect to see the world like this. This is Goa, where my daughter is standing, watching the sunset. But much of our world looks like this, which is one of our field sites, just on the other side of India, where we study plastic pollution. And I expected to see the world like this, where I was doing that research on hoverflies early on in my talk, up at 4,000 meters above sea level. But when I returned the next year, this is what I found. And this is what we do. We don't see the world as it is currently. We see the world as we expect it to be. And we don't adapt to the world around us. We expect our world to adapt to us. And this is causing a lot of problems. And when I thought about this, and I thought about what Tom Eisner said to me so many years before, I realized, I don't think he was talking about my research at all. I actually think, truly, I believe he was talking about me and you, that we need to think a little bit more like a fly. We need to be able to better observe what is actually going on in this planet right now and what we can do to adapt to it more quickly than we are right now. And that realization was why I ended up founding the Echo Network. What the Echo Network is, it's a community of businessmen, farmers, nurses, doctors, NGO leaders, academics, all people from all walks of life who communicate with each other about what is actually happening on the ground in different places in this planet. What is happening? What are people trying to do about it and get a sense of this reality? Not what we expect it to be, but what it actually is. And then find ways of adapting to this and how we can adapt better to our planet. And we're now in over 45 countries across the world, and we have over 2,000 members. And we do this through students, much like the ones in this room. We actually work with young students, many of them from India, but also from other places in the world. And they become what we call our ambassadors. And they work between these different individuals in these different organizations to share information and to learn about the world and how we can adapt to it. And we do the adaptation in four different ways. We increase the capacity of people for taking science and technology information and using it in their businesses, in their governments, in their communities. We work on scaling innovations at scale across India and across the world. We do research and knowledge gaps where we don't have enough information currently. And we put everything in a digital space so that it's really available to anybody who wants to access it. And this is the way that we really try to listen to the world around us, listen to nature, and also try to be a bit more like a fly. So I'll leave you with this. I hope that sometime today after my talk that you're able to go out wherever you are. And I hope you just look around you and I hope you see the world as it is right now. See the insects around you and what they're doing and try to think how can I do a better job? How can I adapt a little bit better? And how can I think like a fly? Thank you so much.