BIO: Nigel Shadbolt
TITLE: Professor of Artificial Intelligence and deputy head (Research) of the School of Electronics and Computer Society at the University of Southampton
ACADEMIC DEGREE: PhD, University of Edinburgh, Department of Artificial Intelligence
CAREER HIGHLIGHTS: He is a director of the Web Science Trust, and of the Web Foundation— both organizations are committed to advancing our understanding of the Web and promoting its positive impact on society. The Prime Minister recently appointed he and Tim Berners-Lee informational advisers to help transform public access to government information. Shadbolt was president of the British Computer Society in 2006-07, its 50th anniversary. He is a Fellow of both the Royal Academy of Engineering and the British Computer Society.
CS ACTIVITIES: EIC Emeritus, IEEE Intelligent Systems
Huge Amounts of Connectivity
An interview with Nigel Shadbolt
Nigel Shadbolt: It’s great to be talking to you, Dick; looking forward to the conversation.
NS: Well, I’m still essentially passionate about research in the field of computer science. I’ve traveled across disciplines since I started out—degrees in philosophy and psychology, PhD in artificial intelligence, a period lecturing as assistant professor, associate professor, and professor in a psychology department where I was building AI systems and intelligence systems, and now into a more thoroughgoing electronics/computer science environment here at Southampton.
I’ve been full professor in the area for longer than I care to think. But it’s still the questions that continue to enthrall me:
• How is it that this exquisite biochemical system that is the brain gives rise to feeling, thoughts, adaptive behavior?
• Will we ever get programs even slightly close to some of the capabilities we have?
• What’s happening in this new kind of environment we’re building where essentially we’ve got an augmented intelligence, not an artificial intelligence, but another kind of AI, where machines and the information processing fabric that is the Web, plus all of these millions and millions of superb human intelligence systems add up to something that’s greater than the parts?
We refer to that variously now as collective intelligence, of course. But it’s a very fascinating place to be.
NS: Not so much an award as an appointment or a role. Back in June, Prime Minister Gordon Brown asked Tim where he thought the next big thing was going to arise. You know, Tim is a bit of a national treasure in the UK—the inventor of the Worldwide Web—and he carries an awful lot of respect over here.
So when Tim said, the next thing, prime minister, is probably going to be what we might think of as moving from a Web of documents to a Web of linked data or an information Web – which is another level of abstraction of where the Web is now.
Once upon a time, we had the Internet where we kind of abolished copper cables between computers and we could think of the cloud of computers without worrying where they physically were. And then Tim came along and gave us the Web, which has abolished us having to worry about where the exact documents were in all these systems across the world and exactly what the particular formats were.
The more recent work that he and I and others have been involved with is trying to characterize these new standards, languages like the resource description framework (RDF), and other bits and pieces that let us connect information not just in documents but in spreadsheets and databases across the Web. He gave this strong vision to the prime minister that, in a sense, data—national data—was a huge and underused resource.
Somewhat like the Obama Administration when President Obama came to power and determined that a lot of federal data had to be made available—the so-called data.gov initiative. This idea’s catching flame around the world. The idea that information that’s paid for by the taxpayer and collected on behalf of running the nation, and particularly the non-personal stuff, the stuff that isn’t sensitive—you can make this stuff available and achieve really great things.
At this point, the prime minister is pretty well impressed. I was invited with Tim to 10 Downing Street to put together essentially a sort of terms of reference around making UK nonpersonal public data, well, public. One of the surprising things about this is that government departments collect lots of stuff, but in many contexts it can be hard to get at that information. I think different jurisdictions; different countries operate differently in this area.
But certainly in the UK there was a feeling we could do a lot more. So we’ve been appointed as information advisors. Interestingly, where we are with that work has been helped an awful lot by the fact that I’d had earlier research projects where we’d been demonstrating the opportunity to use these new standards. So it’s really allowed us to have a bit of a running start.
NS: It’s not actually putting the documents up there. Documents are great and fine. But that’s kind of what the Web did for an awful lot of information provision in government and in business and of course in education and everywhere. This is about data. It’s about the raw information, in a sense, behind the documents that generates many of the Web sites.
When we talk about that kind of information, what we’re talking about in a public context is where the roads are; how many vehicles a day are traveling across the roads; where the bus stops are; what the national train timetables are; where our schools are; how many kids are in those schools. You go across education; you go across transport; you go across communities and local government.
Huge amounts of information are gathered to simply help the country run or understand its state of achievement—population statistics, productivity numbers.
This kind of information, if it’s released in a much more accessible and standard form—not as Web pages, but using this new technology, this new sort of technologies—could allow for new kinds of information, I guess we’d call them mashups nowadays.
Let me give you an example. Recently, the Department for Transport in the UK released data about a particular kind of road accident—bicycle accidents, in fact. Within 24 hours of that data being made available as a particular standard form of raw data, an activist cycling voluntary group had written a black spot accident avoidance route finder. You can now plan your routes ‘round London to avoid the worst cycle accident black spots.
So the notion is that you put information out there of all sorts and stripes. Business, third-sector voluntary organizations, and individuals can begin to do remarkable things with it. It’s a little bit like what we saw with the first Web; when you put your pages up there you didn’t always anticipate the various ways those pages would get used or linked to. But now we’d never think of not putting our documents up in that form.
NS: Well, certainly we’re seeing a lot of interest from across the globe. The US CTO and CIO have been talking about similar efforts.
Already data.gov, which is the US site for data, is there. The difference between that and the UK initiative is that, because we’re a somewhat smaller country and somewhat more homogeneous in our government, we can probably mandate particular sorts of publication more quickly. We’re looking to use particular standards that will make it quite easy for machines to integrate and process this information.
There are efforts underway in Australia and Canada, interestingly enough. And individual cities are starting to jump onto this bandwagon. The notion of citywide information data assets that should be just made available has caught on in cities like San Francisco, Vancouver, London, recently, and so on. There’s momentum behind this kind of approach.
NS: Data.gov.uk, (which just launched), will be the single point of access for a reasonable amount of UK nonpersonal public data, together with recommended sites and blogs and wikis and posts and advice on how to, with respect to the UK initiative in this area.
NS: Well, it’s always nice to think that there was some kind of plan. But my experience on these things is that life is the stuff that gets in the way of your plans. Like many people I grew up with, I grew up during the technological white heat of the space race, in fact. My icons and heroes were the astronauts of Apollo and Gemini. That just seemed remarkable that with a collective national effort you had a man on the moon in 10 years.
As you look back in hindsight, you can argue what the political impetus was behind all of that. But it was a truly remarkable organizational and technical achievement. For me, this notion of, of what could be achieved by technology has always appealed.
At a deeper level, this question of what it is that makes us the kinds of objects in the universe that we are—I’m afraid it sounds grandiose—but for me, three questions that have always intrigued me are human intelligence, artificial intelligence, and whether there’s any extraterrestrial intelligence. It’s just the remarkable fact that we can exist in an apparently huge and incredibly stark universe. I mean, how has that all happened from apparently simple evolutionary beginnings?
To understand that sufficiently well that we can help improve ourselves and those around us, that’s been really the motivation for me.
I started off through the route of the human sciences and gradually realized when I’d done my first degrees in psychology and cognitive science that programming, the methods of software engineering, promised one way to make sense of these faculties we have. Of course, that is the field of artificial intelligence with which I’ve been associated since about 1978 when I went to do my PhD at Edinburgh in the AI Department.
The really interesting part of that journey was to then realize, yes, building programs teaches us much, but the more you build them, the more you realize just how extraordinarily complicated human and animal cognition is. Far from dehumanizing us, this approach to artifacts really just makes you stand in awe of what’s been achieved, of what we are.
So I’ve come into this perhaps somewhat esoterically, through the kind of bigger philosophical questions. But very early on, what appealed to me was the old Latin saying that “factum quod certum.” [It was Nielsen who used it in his lovely book on artificial intelligence.] You are certain of what you build.
That engineering discipline is something that’s very powerful. It’s something that the Computer Society and other professional bodies represent, that we now build a technological environment where very complex systems change, support, and benefit us. It’s pretty important that we continue to train and infuse people with the opportunities that approach offers.
NS: Actually there was. I could think of a number. But I have very fond memories in particular of Geoff Midgley. He taught me philosophy, believe it or not. But actually what he taught me was mathematical logic. He was married to a rather well-known ethical philosopher called Mary Midgley. The Midgleys were a pair of brilliant thinkers at the University of Newcastle upon Tyne where I studied philosophy and psychology.
He infused me with mathematical logic. I then got into the whole area of semantics. Geoff Midgley had programmed on some of the very earliest computers in the UK and had followed a route that many people did in the UK. He’d been a very brilliant student and had helped in the war effort by supporting a whole notion of what information analysis could do for us.
It was that enthusiasm of somebody who was technically gifted but also saw the bigger picture—what we could prove machines might or might not be able to do.
NS: So I’d put it down to Geoff. Mary Midgley was, as an ethical philosopher, always tweaking my tail to say, do you think there’s any prospect at all that this kind of mechanistic approach to, to understanding our nature is, is really going to buy us a deep and rich understanding of ourselves? So I was always given this slightly broader view of the questions we confront.
NS: The reason I went to study artificial intelligence at Edinburgh was that I had become very intrigued by the notion of building programs as a way of testing our understanding of human intelligence. But very few people back then really thought this was a practical way to go.
I went to Edinburgh to start reading my PhD at the very time a famous report in the U.K. was released called “The Lighthill Report” that pretty much closed down all AI research in the UK. It was some years later that the Japanese Fifth Generation effort to build intelligent machines was launched, and everybody suddenly felt they needed to get into it. So I went to do the PhD because I thought the questions would be intriguing.
As I exited the PhD and went to begin an academic research career, I then suddenly realized that one of the really interesting aspects of all this was that you could understand rather better whether you achieved anything worthwhile by seeing whether it made a difference in the world.
So with the emergence of so-called knowledge-based systems or expert systems, I found myself pretty early on trying to build software systems that could fit into real working environments to help support decision-making, to help make things a little more efficient, productive, and so on.
At that point that I began to understand the need, if you will, for a professional grounding. It’s not enough to have the best academic, theoretical ideas. When you’re in the world and there are consequences of the systems you build, you need to worry about validation; you need to worry about the properties and qualities of your systems.
So the advice may sound a little bit unfashionable at the moment. But my advice is to follow your heart and passions, because that will inevitably lead you to do the best things you can for yourself. When you find yourself in situations where your work has a significant impact on society and those around you, ask yourself whether you think you are in need of additional support to help mature that career.
With our own students, we’re trying to get them to think about professional issues around software systems rather earlier than perhaps we were made aware of them when I was a student.
NS: A lot. I think it’s the nature of human beings to form groups and communities of common interest—birds of a feather. There is something in the way that we organize ourselves into groups, into workshops, into conferences, that shows the very deep sense in which people who share common interests want to share and learn from one another.
Whether it’s the AISB in the UK or AAAI or BCS or the IEEE Computer Society, they’re about helping communities to organize, to support them, to connect elements and communities together. Very often as you begin a career you are very much, in a sense, working to a particular set of interests; you have a view of the particular things you want to do. You might think you’re a computational linguist or that you want to do a certain sort of software security career path. You need the organizational structure around you to connect you with people who have similar interests.
Over time, you realize that your particular skills often don’t just sit in a silo. They have a wider set of applications that you need to know a little bit about another area or another set of issues. It’s invariably the case as you look at complex systems. Learned societies are particularly good at fostering this sense of community. I’ve found it absolutely invaluable.
The reason I put a lot into them is that it’s a constant cycle of renewal. The communities and networks of 50 years ago are not necessarily the exact same ones we need today, but you still need the process to make them and keep them flourishing.
NS: Ah, well, decade, yes. It used to feel like decade was a very long time. And of course it is, in many respects. But in other ways, we’re almost—as William Gibson famously said—the future is already here; it’s just unevenly distributed.
You can see that some trends are extremely robust. People will talk about Moore’s Law, of course, and various developments in power and price in the hardware and electronics arena. But what’s really startling about that set of curves and projections is that 25, 35 years ago. When I started out in my earliest academic engagements with this subject, there’s been a millionfold increase in the power of the systems I have at my disposal—10 to the sixth. No other discipline has enjoyed that amount of additional capability.
It’s meant that the things we can now do in a particular sort of way that we absolutely believe we couldn’t do in a particular sort of way 25 years ago are routinely solved. Whether it’s searching the entire Web of documents and building a huge index and then working out which links to go through for the most relevant information like Google does or whatever – these are remarkable feats. Or indeed, turning over the world chess champion because you can simply build systems that can look deeper and further into the search space of chess.
Now we know where those curves are heading over the next few years. We can be pretty confident, for example, that the device ten years hence that I carry around formerly known as a mobile phone or whatever it’ll be called then will be capable of recording real-time continuous video of large amounts of my day, week, month, year, life.
So what is going to happen in a world where we’re life-logging—we’re literally logging more and more of the information about us and it’s being described for us and indexed for us and retrieved and searched for us and collectively we share it between ourselves? What does that world look like? It’s a superfluity of information, huge amounts of connectivity between mobile devices, letting us tap into huge amounts of collected information.
It’s very hard to know what the social impact of that will be, what new practices will emerge. But we know a lot about what the technological trajectory will look like. That tensioning is very interesting to me.
Also, there’s still a lot of work to do for those who are digitally disenfranchised. We talk very proudly of the fact that so much of humanity is connected by the Internet and the Web. But the truth is that billions and billions of people on the planet aren’t yet. So we have a job of work to do there as well, I think.
NS: Without the subject areas that IEEE Computer Society represents, without large numbers of our brightest and best going into these areas, we’re not going to solve the really challenging problems that we face—the problems of climate, the problems of equality and poverty, the problems of energy. All of our major human effort now is partly computationally based and requires the linking of our engineering base with our human intellectual capability to move us on to solve these very challenging areas and to create new sorts of opportunities for us.
It’s incredibly exciting. For me, engineering is such a very exciting place to be. The irony is, and the slight conundrum is, that it’s often thought to be rather prosaic and somewhat unimaginative, but nothing could be further from the truth in my experience.
- Sumi Helal: Mobile and Pervasive Computing
- Nur A Touba: Design and Test Research
- Deborah Cooper: Reaching the Under-Represented
- George Cybenko: Dorothy and Walter Gramm Professor of Engineering at Dartmouth College
- Sorel Reisman: Technology and Teaching
- Susan K. (Kathy) Land: Software Process Improvement
- Sajal Das: An interview with Sajal Das
- Don Shafer: Complex Control Systems
- Natalia Juristo: Mastering Experimental Software
- Nigel Shadbolt: Huge Amounts of Connectivity
- Shmuel Shottan: A Passion for What You Do
- Elena Ferrari: Improving Security & Privacy in Social Networks
- Harold Javid: Developing Global Understanding
- Dawn Song: MacArthur Award for Computer Security Specialist Dawn Song