News
Achieving the Impossible: A Celebration of Research
By: Becky French
Last updated: Friday, 16 August 2024
![Speakers Alice Eldridge (Professor of Sonic Systems), Thomas Nowotny (Professor of Informatics), Mari Martiskainen (Professor of Energy and Society), and Paul Statham (Professor of Migration).](/broadcast/images/uploads/2024/08/20277.item.jpg)
In a celebration of Sussex research, alumni and supporters came together, to explore and celebrate some of the remarkable contributions academics are making, in addressing some of the greatest research challenges of our time.
Speakers Mari Martiskainen (Professor of Energy and Society), Paul Statham (Professor of Migration), Thomas Nowotny (Professor of Informatics), and Alice Eldridge (Professor of Sonic Systems) shared insights on: the unfairness of fuel poverty, the realities of multiculturalism and migrant integration, the urgency of balancing the development of AI with sustainability, and the importance of eco-justice.
Watch the videos to learn about the challenges our researchers are taking on. Their work is tackling some of the most urgent problems that people and planet face – challenges that seem #ImpossibleUntilDone.
Hear from our researchers
- Video 1: Mari Martiskainen transcript
TITLE CARD: ‘IMPOSSIBLE UNTIL IT’S DONE’ / ‘A CELEBRATION OF SUSSEX RESEARCH’ / ‘Fairer more sustainable energy’ / ‘Professor Mari Martiskainen’ ‘University of Sussex’
[Applause]
Professor Mari Martiskainen: I want to start my talk by asking you, the audience, which one of you had a hot drink this morning—either a cup of tea or a coffee, or used another kind of means of making a hot drink? Please put your hand up now. Ah, quite a few of you, thank you so much. I'm also going to ask you another question, and this is a little bit more personal, so don't feel obliged that you have to put your hand up. Did any of you have to think about how much making that cup of coffee or tea might cost you in your energy bill? Have a think about that, because in the last few years we have seen an energy crisis in the UK, and our bills have gone up. Everyone has noticed it, whether you can afford your bill or not, you've certainly heard about it in the news. It's been front-page news in all of our newspapers in the last few years, and our energy bills on average are still about £400 more than they were about three years ago. So it's a real issue.
I want to tell you the story of Joe, who is one of our research participants. I interviewed him a couple of years ago for one of our research projects where, as part of the Sussex Energy Group at the University of Sussex, we wanted to find out how people actually use energy. So we went around and spoke to people like Joe, and I interviewed him. He told me that for six years he didn't use the heating at home because he was so worried about his energy bill. He couldn't afford to do a hot meal; he wasn't cooking anything at home, he wasn't making any hot drinks, and he told me that he was literally kept alive by a charity that brought him sandwiches.
Joe is not alone in the UK. We have about six million households—that’s six million—that’s the same amount of households that a small country, for example, might have. I come from a country, Finland, and the population is 5.5 million. That’s the whole of the population, but in the UK there are six million households who cannot afford to heat their homes, to cook meals, to basically pay for their energy bills. They live in fuel poverty, and that means they have quite a lot of dire experiences in their daily lives.
If you live in fuel poverty like Joe, it affects your mental health; it affects your physical health. Particularly, it affects children who might not have a warm space at home to do their homework in. They might go to school, for example, smelling of damp because the home is so cold and damp, and for that reason, children might get picked on. So while we have a lot of people who cannot afford their energy bills, we also have a huge number of people who can. People like myself—you know, I can make a cup of tea; I can make multiple cups of tea a day and not really think about whether I can afford my energy bill. But I do think about those cups of tea because it's consuming energy, and we consume too much energy in many cases.
We have this excess consumption, and it’s partly because of our lifestyles. They have changed a lot in the last few years. For example, in 1970, people had an average of 17 appliances in their home. If you think now, for example, how many electrical appliances you have in your living room, in your kitchen, in your bedroom, even in your bathroom, even though it's dangerous to have electricity and water mixed up—and particularly in your teenagers’ bedrooms. Anybody with a teenager will find out, you know, there will be more than just a smartphone in there. So our lifestyles have really, really changed, and we are consuming a lot more energy in the way that we heat our homes, the way we travel, the kind of diets that we have, what we basically eat, and what we consume in our daily lives.
I think it's really important that we make the link between energy, emissions, and climate change. I'm sure everyone in the audience has heard about climate change. It’s happening; it's here; we are feeling it. Some other countries are being far more affected, but already in the UK, we are also feeling the effects of climate change, and it’s not going to go away just by ignoring climate change. It's not going to go away—we need to do something about it. We have actually legal targets that by 2050 we have to have a net-zero society. Presently, about 40% of our emissions come from households. So that's me, you, everybody in the audience—we all consume energy that creates emissions, particularly if it comes from fossil fuels and other polluting sources. We need to do something about this.
In the Energy Demand Research Centre that I direct for the University of Sussex, we are building on previous research which has shown that we actually have a real opportunity here in the UK. We could reduce energy demand by 52%—that's by half, you know, that's a big figure—50% is, you know, it’s a big figure for somebody who does research. But we need to figure out how we can get there, and that will require a mixture of measures. It’ll require a societal shift so that we can actually start talking about energy in the way that, you know, we can't just keep consuming the way we do going forward. We need to think about our diets; we need to think about our home energy consumption; we need to think about transport—so the way we eat, the way we travel, the way we heat our homes.
But also, this needs to be going across the whole of society, so we need a whole-systems change in this. It’s really, really important that we do this fairly. So when we are talking about people like Joe, for example, it's really important that when we are looking at our net-zero transition, we do it in a fair and sustainable and just way so that people like Joe can actually have a sufficient amount of energy services in the home, so that they can have a hot meal, so that we don't have children who can't do their homework because the home is too cold, for example. So we have to make sure that people who are in energy poverty or fuel poverty can actually have a sufficient level of energy services. But at the same time, we really need to also encourage those that might be consuming too much to actually start thinking about how they can bring their consumption down, how they can make sure that the energy they use is actually sustainable, and if there is some way that they can reduce what they’re about to do in their daily life, for example.
We do think, at the University of Sussex, particularly in the Energy Demand Research Centre, that it is possible to reduce energy demand. I mentioned this requires a whole-systems change, so it means that we have to consider industry, buildings, transport, AI—you will hear about AI later. But by doing this, we can also take opportunities in the huge number of benefits that we will have from a sustainable net-zero energy system. This is issues like cleaner air when we don't have polluting fossil fuels anymore, better diets—so thinking about low-carbon diets. And I’m not here talking about low carbs in the kind of form of biscuits—I’m here talking about plant-based diets, for example, new jobs, new skills, and also our energy security to make sure that the energy we use is being produced here in the UK, for example, from renewable energy sources.
So with that, I will just leave by saying that the UK really has a real opportunity here to lead a change that will make our society and the way that we use energy better for all of us, and particularly for our future generations to come. Thank you.
[Applause]
END CARD: ‘University of Sussex’
- Video 2: Paul Statham transcript
TITLE CARD: ‘IMPOSSIBLE UNTIL IT’S DONE’ / ‘A CELEBRATION OF SUSSEX RESEARCH’ / ‘Culture Clashes in Cross Border Connections’ / ‘Professor Paul Statham ‘University of Sussex’
[Applause]
Professor Paul Statham: So when you had your breakfast this morning, did you feel British values under threat? As you made your way to this building, this wonderful building here around us, did you feel pressurized by Sharia law, or did you see a city bustling with diversity?
I find it quite concerning that three out of 10 people see the world like this. I'm slightly less surprised that those who gave us Liz Truss do, but that's a story for another day. How can it be that this is normal for so many of us?
There are two stories at play here. The first is the only story in town in media and politics. This is "The Clash of Civilizations" as it's called. It's about societies cleaved into two: on one side, white Christian majorities, and on the other, an alien culture—Islam—brought by immigration that becomes a threat.
But there is another story that you don't read in the newspapers. It's a story of incremental social and demographic change, transition through diversity brought by ethnic mixing. It's a place where kids of different colours play in the schoolyard, their parents work together, and more and more of us choose partners who are from different backgrounds than us. It's a place where girls wearing the niqab in Bradford speak in broad Yorkshire accents. It's also a place where the Muslim son of a bus driver can be democratically elected the Lord Mayor of this city—London.
So it's also a story of aspiration, of opportunity, of change through diversity. It's a story where British Muslims have become so much a part of British society that, for all intents and purposes, it doesn't really matter whether they're Muslim or not.
Now I've been working on these topics for 25 years, from the days when we were just a few oddballs studying the political accommodation of minority religions until it came to the forefront of European consciousness: 9/11, the global war on terror, as it was called, and endless multicultural conflicts.
Now, what's concerned me looking back over all these 25 years is the degree to which politicization has really impacted on how we do research on these topics. I've called this, not without a slight hint of sarcasm, the "Muslimification of Muslims."
Now at Sussex, we did a massive survey research of Muslims across Europe, across six European countries—first and second generations. It was one of the biggest that's ever been done. And of course, we asked all the questions about culture, values, identities—all the hot topics. But we didn't just look at culture as a source of cleavage; we also looked at and asked the questions that quite often don't get asked about Muslims these days, about how culture can be something that brings people together, that works across boundaries. So, you know, "Do you have a member of your family from a different background?" or something like that—all these kinds of questions that for other types of minority research are quite present and quite normal but have rather been left out of the equation when we're talking about Muslims.
So what's the problem then with this "Muslimification of Muslims"? How does it work and why is it problematic? Well, the first issue is to do with the category of "Muslim" itself, because Muslims in Europe are a highly heterogeneous group by color, by type of Islamic faith, and that's even before you start talking about all the other things we need to talk about like gender, age, class, status—all those kinds of things. But increasingly, research lumps all these people together as Muslims because it's simply mimicking what goes on in the political debate. Another reason they do it is because the quality of their data is quite often not good enough to do anything else, but they don't often say that.
Now our research was able to take this umbrella of "Muslim" and look at differences within, and we saw very different trajectories of acculturation within that, so one size does not fit all.
The second problem of lumping all Muslims together because they're Muslims is because it turns the focus on religion and religiosity. Religiosity has become a sort of über-explanatory variable when we talk about Muslims. Against this, our research saw that, in fact, religion wasn't really that important. Those cultural aspects that cross boundaries were much stronger and had much more explanatory power. So, feeling British, feeling German, feeling Dutch, reading British, Dutch, or German newspapers, and having a member of your family who's intermarried—all those kinds of things worked much more and pushed in an acculturative direction.
Now, the third point is about perceptions of democracy that we have. These are the so-called conflicts over liberal democratic values. Back 25 years ago, when I was setting out, liberal democratic values were seen as a good thing. This was going to be the civic stuff that held us all together facing the challenges of globalization, of which immigration and super-diversity are, of course, part. But today, liberal democratic values are brought in as a stick to beat Muslims. It's basically an ethnocultural test of Britishness that's set as a standard to which Muslims and other cultures can never reach.
Interestingly, our research showed that when we asked the type of questions that are used in studies on liberal democratic values, Muslims themselves saw themselves as being relatively close to the rest of society in a way that wasn't really problematic. But when we turned the question around and asked majorities, it was them who saw big problems. This shows us, then, the barrier is constructed on the side of the majorities. It underlines what was demonstrated in the opinion poll that I showed you at the start.
So what's the role of the social sciences in all this? Well, today, the social sciences are quite often polemical. It's critique versus counter-critique. For me, however, sociology isn't just about being a social justice warrior; it's about making visible the way the world actually works. And once we can understand that, then maybe we can start to think about how we can try to make it better. In a sense, I think sociology needs to reflect on something that George Orwell said 75 years ago: that it's a struggle to see what's in front of your own nose. This is a challenge that the social sciences need to take up, and it's an important one. For a better world, we need better research.
[Applause]
END CARD: ‘University of Sussex’
- Video 3: Thomas Nowotny transcript
TITLE CARD: ‘IMPOSSIBLE UNTIL IT’S DONE’ / ‘A CELEBRATION OF SUSSEX RESEARCH’ / ‘Climate Conscious Intelligence’ / ‘Professor Thomas Nowotny ‘University of Sussex’
[Applause]
Professor Thomas Nowotny: Artificial intelligence—AI—everybody is talking about it. Some people are very excited about the technology and what it can do. Large language models such as ChatGPT can write text that looks like it's written by a human. Some people think that they can even pass the Turing Test, a test that was designed by Alan Turing long ago to find out whether a conversation partner is human or not just based on questions and answers. AI can also do other things. It can make lifelike images and videos, and you will notice during this little talk I will use images that I have created with AI. So watch out for the prompts that I put on that I used to create those images with Bing Image Creator. AI will also soon be able to do many of the repetitive things we do in our jobs and free us up to do more generative things, more creative things.
Other people are really, really worried about AI. Some people worry about their jobs because AI can do all these repetitive things that they do in their jobs and feel they might soon be out of a job. Other people worry about those lifelike images and videos and what you can do with them, especially in an election period like right now. You can create a different reality with AI that some people might believe and this will endanger our democracy. Yet other people are worried about the singularity. This is the scenario where AI becomes sentient and conscious and takes over the world a little bit like the Terminator movies.
But I think there's a much more immediate problem that we need to talk about: energy. Training AI systems and building them costs a lot of energy. These systems are trained on millions and millions of examples. It has been estimated that building GPT-3, the engine behind the original ChatGPT, cost as much energy as driving a car the distance to the Moon and all the way back again—that's 760,000 kilometers or roughly 19 times around the circumference of the Earth. This sits on the background of the fact that we're already using about 4% of the world's energy for IT even without AI. The whole of aviation only takes 2%, so this is a massive amount of energy already used. And if we continue with the current trend, where Transformer models—that's the technology behind these large language models—increase in size about 750 times every two years, if you continue this trend, you can make the calculation on the back of an envelope that by 2030 at the latest we would use all the energy produced on Earth just to make chatbots. So this clearly can't continue; this trend is impossible to sustain.
So is there no other way of doing intelligence? Of course there is. All of you, as you're sitting here in this room, have an amazing intelligence device right here in your heads. The human brain is estimated to use only about 20 watts of power—that's the same as an energy-saving light bulb. So how is the brain different that it can be so much more efficient than current computer systems in current AI? Well, some differences are obvious. You're not plugged into the wall and working on AC power supplies of 240 volts, and your head's not filled with silicon chips that are stacked on each other and get really hot when you're thinking. But there are also other differences that are important. Current AI is based on neural networks where the neurons in the neural network exchange their activation states every time step of the computation with every connected neuron. This creates enormous data flows and it's the data flows actually that make computers consume a lot of energy.
In contrast, in the brain, neurons work more in isolation for longer times and only exchange information occasionally with so-called spikes and these only transmit about one bit of information at a time. This idea of spike communication and sparse communication has inspired engineers and scientists since the 1980s to try to build computers that are more like brains and that are more energy efficient. And many of these systems have been built. But do you have a neuromorphic phone in your pocket? Why do I not have a neuromorphic watch? What is the problem here? The problem is that we don't actually know how to train these spiking neural networks that have this different paradigm of sparse communication to a level where they can compete with the artificial neuron networks that work on the traditional computers.
And this is where we come in with our research at Sussex. For more than a decade now we have developed software and systems to simulate spiking neural networks very efficiently on standard hardware to find out how we can train them better, how we can make this work so that we can move to less energy-intensive systems.
So let me unpack the problem for you a little bit more. The way artificial neural networks are trained is that you have the neural network and you have a task and millions of examples. As I already said, these examples are fed into the network one by one, we record their output, their answer, and then measure with a so-called loss function how good it was. When the loss is high, the answer was bad; when the loss is low, the answer was good. And then using mathematics, we can determine how to change the neural network to decrease the loss function, to decrease the badness of the answer and therefore make it better. You can picture this as going down the loss landscape from a high point where you start to the lowest point where the loss is lowest and therefore the network performs well and gives the right answers.
And for artificial neuron networks with their dense communications, this works very well. We have kind of a loss landscape that's rolling hills and where it's easy to find the way down the steepest way down the hill into the valley. However, for the spiking neural networks, this landscape looks terrible. You have spikes and troughs and you don't really know where to go to make the network better. We have worked on this and most recently we have found a way to change the way we measure loss and that has transformed this landscape ever so subtly so that we now can go down the error landscape and train spiking neuron networks as efficiently as the artificial neuron networks we know today. And we have achieved a state-of-the-art performance in a speech recognition task just last year.
We are now working on a project with Intel to take these networks and put them on their Loihi 2 neuromorphic hardware. This is a computer that they have built with the same aim of making more brain-like, more energy-efficient computers. And we hope that we can demonstrate that we can actually harvest these savings of sparse communication in your networks in a concrete example of speech recognition on hardware. So we have achieved a lot and there's a lot more to do, but I do think we are on the way to a less energy-intensive, sustainable AI for the future.
[Applause]
END CARD: ‘University of Sussex’
- Video 4: Alice Eldridge Transcript
TITLE CARD: ‘IMPOSSIBLE UNTIL IT’S DONE’ / ‘A CELEBRATION OF SUSSEX RESEARCH’ / ‘Learning to Listen’ / ‘Professor Alice Eldridge ‘University of Sussex’
[Applause]
Professor Alice Eldridge: Following on from energy, migration and AI, I'd like to zoom out to planetary health—and of course, they're all related. By many measures, human well-being is at an all-time high: health, wealth and longevity—never have so many had so much. But these gains are not evenly distributed, and they come at great cost to our planet.
Reports that our Earth is now outside the safe operating space for humanity in terms of climate change, pollution, land use change, and biodiversity loss are terrifying and can be crippling. How on Earth can we address these issues? The framework of planetary health reminds us that human health and planetary health are intertwined. If we first consider the impact of breaching planetary boundaries on humanity, the need to act directly to return each to safe operating space is obvious. But we also need to consider the impact of humanity on the planet and various thinkers—from cyberneticians and philosophers to activists, ecologists, feminists and spiritual leaders—all suggest that the root of our current ecological crisis is a fundamental error in our thinking. The core tenets of Western philosophy, religion, and even the rationalism that drove human progress also underpin the extractivism and colonialism that are threatening life as we know it. At root, we seem to have forgotten that humans are not distinct from the rest of life. I'm not separate from you, and neither of us is separate from the environment.
Of our various ecological crises, threats to biosphere integrity are closest to my heart and mind, and it is also in the worst state. I don't think this can be solved by any one discipline. We need to act fast to restore degraded ecosystems and protect the last remaining biodiversity hotspots. But we also need to preserve and regenerate ideologies and cultures of planetary care at the community level. And individually, each one of us needs to find ways to ameliorate eco-anxiety and shift our felt relationship with the Earth.
Now this was a bit of a doomy start, but I'm going to bring some hope. My name is Alice Eldridge. I'm a professor of Sonic Systems, and I made that up because I wasn't sure what discipline I belong to. I wanted to create a space to consider how systems across scales—ecological, cultural, and personal—are cohered by sound and how we can understand, protect and regenerate these systems by listening to them. So in the next few minutes, I'm going to attempt to persuade you that learning to listen across levels and in a range of ways can help save the biosphere.
Our subject of study is the acoustic environment, or soundscape. This is familiar to you—you hear the soundscape all day, every day. The soundscape emerges from biological, geophysical and human processes and is simply the set of sounds perceived in a given place. Soundscapes physically manifest as vibrations, and we can measure these, which is good for science, but soundscapes are really all about meaning. Soundscapes are flows of information that have biological, cultural and social value. Organisms across the tree of life rely on sound to stay alive. Birds, bees, bats and beetles, whales, fish and frogs all vocalize, and that sound we heard was the sound of a reef. Even coral polyps are sensitive to sound and navigate to the reef through its sound.
So let's first turn to the planetary-scale issue of biodiversity loss. Many fear that we are now entering the sixth great extinction, but again there's some good news: lots and lots of nature recovery projects, locally, regionally, and globally. But in order to make informed decisions, we need to be able to measure the impact of our interventions, and we are operating at scale—this is happening all over the world. So we can't just stand in the field and count birds anymore. We need some kind of remote method. Satellite methods are very attractive, but we can't see through the canopy or through the ocean or under the Earth—but we can hear.
Just as doctors have listened to our lungs, intestines and hearts for millennia, we are now learning to make digital stethoscopes for ecosystems—technologies that can help us hear ecosystem health. How does this work? As I've said, for many beings, sound is central to staying alive: finding a mate, a home and some food. Ecology has long recognized that life actively alters its environment by living. Trees make oxygen, beavers make dams, humans make piles and piles of rubbish, and creatures alter their acoustic environments by vocalizing. So if you think about that, the corollary is that patterns in the environment left by things being alive tell us something about the organisms that are there and their interactions—in other words, the ecosystem.
These digital stethoscopes are built from very simple audio recording devices left out in landscapes to record over weeks, months, days and years, or even live stream—think kind of an ecosystem baby monitor. Audio recordings are then analyzed and this is the difficult research—to gain insight into the health of the environment. This is now turning from proof of concept to application and it's being used in terrestrial, marine, freshwater and even soil ecosystems. There are lots and lots of scientific research questions left open which we are working on with colleagues across the University of Sussex. But I want for the last few minutes to shift from that to think about the cultural and personal power of listening, but I am happy to talk to you in drinks afterward if you'd like to.
We first tested this approach about 10 years ago when I was working with a really brilliant conservation biologist, Mappack—some of you may know—in the Sussex countryside as well as the awesome forest of Ecuador. So aside from the basic thrill of science, something struck me: evolutionary theory and our early data all pointed to this novel idea that ecosystem health could be listened to. But if we really could hear ecosystem health, surely the people who live and work in forests and oceans would be the experts here. Over the next few years, I did find traces of this. In North Bali, I learned that night fishers listen through their oars to find the reef—that crackling and popping we heard earlier is where the fish live, so it's a fantastic way to find life in the dark.
Jump forward a few more years, and a past PhD student, Paloma Escaso, was back in her home country of Ecuador. She was working with a group of indigenous communities in the western Amazon who wanted to gain autonomy for their territory—the Pueblo Ancestral Kichwa Kawsak Sacha, the people of the living forest. In order to do so, they needed to persuade the government that their biocultural heritage, the space around them, was worthy to be registered in the national register. So rather than the surrounding biodiversity, which is amongst the highest in the world, or their material culture, it was the Takina, the Sacha Supai—the songs of the forest—which the Kawsak Sacha communities wished to have recognized.
So during lockdown, we co-developed very simple methods to help them share what Sacha Supai meant to them with the authorities. What surfaced was an exquisite articulation of the importance of the voices of the forest—not simply to speak of the health of the forest, but to guide them in every aspect of living. These Takina carry mundane signals: the call of a Sarasapo toad signals the time to plant corn, the Potoo bird signals that the water level is about to rise, and the Cotomono monkey alerts that there's a jaguar nearby. But it's also through these songs that they communicate with the forest guardians, the forest spirits, who invoke soil fertility, cure diseases, create pacts of mutual care, and remember and transmit ancestral teachings.
So more than simply reinforcing the value of listening to ecosystems, Sacha Supai carries the Kawsak Sacha cosmovision—some core concepts by which they live: Sumak Allpa, a land without evil, i.e. without extractivism; Sumaq Kawsay, a life in harmony—the idea that no one can be well unless all beings are well. So perhaps their territory as a global biodiversity hotspot is not a coincidence. For me, then, beyond bolstering faith in the new science of ecoacoustics, this partnership provided insight into worldviews that might help us globally to find paths to planetary health with wider groups. Our next step is to build a kind of a network of what I'm calling multiscale listening to weave the emerging science of ecoacoustics with traditional ecological knowledge in part to align future conservation technologies with local needs—and ideally avoid a kind of new imperialism—and in part to create bridges to relearn regenerative thinking in the West.
But there's a simpler personal message here too. We don't need to be Manicapanas to feel the connection to the rest of life through sound. Much closer to home, in Sussex, I've been experimenting with ways that simple acoustic technologies can bring urban populations back into contact with the wider songs of nature. By creating live audio streams from rewilding projects in Knepp in West Sussex and recordings from our local biosphere, sharing with audiences from all walks of life—online, in festivals, in classrooms and even in the hospice—I’ve been witnessing the power of sound to reconnect people with the rest of the natural world. This simple visceral experience of listening to the rest of life seems to provide a powerful experiential shortcut.
Again, there’s a lifetime of research to really understand how this works and its long-term impact, but it seems that this simple act of listening helps people to relax the tensions that cut us off from the world. We can engage with the felt dimensions of experience and dissolve these centuries of Western thought that told us that we were separate from the rest of life.
The future really needs us to act now, but we must also pause and listen. I believe that listening across scales—to ecosystems, cultures, and our own Sacha Taki here at home—will help. Thank you.
[Applause]
END CARD: ‘University of Sussex’