# Hi-tech Trek

Under the surface activities help you take a deeper look at some of the island's featured topics.

If you missed the lectures when they were on TV, or just want to watch them again, check out our fantastic video clips.

### Breaking the speed limit

#### Lecture transcript

Download the full Breaking the speed limit lecture transcript (PDF document 299KB - new window)

#### Video clip – Ping-pong mousetraps

With the help of an audience member, Chris demonstrates the exponential growth of computer power.

PROFESSOR CHRIS BISHOP

Now we can see for ourselves what exponential growth looks like by doing a little experiment. If we can bring one of the hand-held cameras in, we can have a look at this.

So in this box I have 225 armed mousetraps and on each mousetrap we have a ping-pong ball. Now in a moment we're going to take one more ping-pong ball and we're going to drop it in through this hole in the top and it's going to set off a chain reaction. And in this chain reaction, the number of ping-pong balls flying through the air is going to grow exponentially. Okay, so, who would like to volunteer to come and set this off (laughs)? My goodness, um, let's have you there, yep. If you'd like to make your way along to the end of the row.

APPLAUSE

Now if you'd like to, if you'd like to come and stand just there and just, ah, just turn that way. Okay that's good. And what's your name?

MATTHEW

Matthew.

PROFESSOR CHRIS BISHOP

Matthew. Matthew, if you'd like to hold that. Now in a moment we're going to give you a three, two, one countdown, okay? And when we get to go, all I want you to do is to place the ball in through that little hole at the top. You manage that, yeah?

MATTHEW

Yep.

PROFESSOR CHRIS BISHOP

Okay, are we ready? Three, two, one, go!

(Laughs) Amazing, excellent. Okay, thank you very much. (Clears throat)

APPLAUSE

Okay, so, so we have a high speed camera that was looking at that and perhaps we can just do a little action replay and see that in slow motion. So there's the first ball, setting off a couple more, each of those is setting off several more. And on that curve we can see the number of balls growing in this very dramatic way. But what's really impressive about exponential growth is that the rate of growth is itself growing exponentially.

Now the exponential growth of computer power is truly staggering. It means, for example, that the computers that will be made in the next two years have as much processing power as all the computers ever made, from the very first computer up to the present day. If cars had improved at the same rate as personal computers, then a typical family car today would travel 43,000 times faster than a Formula One racing car and it would go 200 times around the world on one litre of petrol.

### Chips with everything

#### Lecture transcript

Download the full Chips with everything lecture transcript (PDF document 352KB – new window)

#### Video clip – Give me a wave

Chris uses sunglasses to explain the technology behind liquid crystal displays.

Okay, now when you all came in this evening, you should have received some little blue envelopes. If you'd like to open those blue envelopes and inside you'll find two little plastic squares. Now those little plastic squares are made of the same material as these sunglasses. So if you hold one of them in front of the other and just look at the white screen up there, when you rotate them, you'll see that they first of all block out the light and if you keep rotating, you'll be able to see the light again.

Okay, so this is an amazing effect, but what's it got to do with liquid crystal displays? Well if I take one of these lenses and I hold it up in front of this display, if I rotate the lens, you'll see again the display goes dark. So that's an amazing effect, but what's actually going on there? Well if you'd like to just come over here with me, we'll do another little experiment, if you'd like to come and stand about there.

So we have to understand why light gets blocked when we rotate those polarisers. Now light is a kind of wave, we call it an electromagnetic wave and this machine will simulate wave motion. So what I'd like you to do is just crouch down here with me. All I want you to do is to move this up and down, nice and slowly, nice, big, slow movements. You got that? Keep hold of it, okay? Keep going, nice, big, slow movements. Okay, and that's sending a wave down the machine that's a bit like a light wave. Alright.

Now what I have here are some models of polarisers. Thank you very much. These are just sheets of plastic with slots cut into them, so I'm just going to place these over the machine like this, one there and one there. And notice they have the same orientation, they're both vertical. Okay, if you'd like to send a wave down the machine then please. And that wave is passing through the first polariser and then through the second polariser. That's it, nice, keep going, nice big sweeping movements. So the waves are going all the way down to the end.

Okay, just stop for a moment. What I'm going to do now is take one of these polarisers and rotate it through 90º and then put it back again. So if you'd like to send another wave down the machine. So again, the wave is passing through the first polariser, but this time it's getting blocked, so the wave doesn't make it all the way to the end of the machine. And that's what's going on with those plastic polarisers. Okay, thank you very much.

APPLAUSE

### Ghost in the machine

#### Lecture transcript

Download the full Ghost in the machine lecture transcript (PDF document 502KB – new window)

#### Video clip – Cannonball challenge

Chris puts his life on the line in a breathtaking demonstration of the laws of physics.

Now music is often compressed with something called MP3. And MP3 goes even further, because it throws away aspects of the music which the human ear finds very hard to detect. Now taking music and compressing it, using MP3, is another example of a computational recipe or algorithm.

And there are lots of other problems that are much more complicated which computers are also very good at. So, for example, classical physics is very well understood and it's described by equations that were worked out over 300 years ago by Isaac Newton. Now I'm going to prove just how predictable and reliable these equations are by conducting an experiment, in which even a tiny variation in the laws of physics, could result in my getting killed.

I have here a solid steel ball – it weighs 14kg, it's incredibly heavy – and it's suspended from the roof of the Faraday Lecture Theatre by this steel cable. Now what I'm going to do is to take this steel ball over here and I'm going to stand with my back against this headrest and in a moment, I'm going to place it against my face and then I'm going to let go.

(Laughs)

It's going to swing out across the lecture theatre and then it's going to swing back towards my face.

Now according to the laws of physics, it should stop just before it touches me. Okay, that's the theory, let's see what happens. I think this is probably worth a countdown, isn't it? Okay, are you ready? Three, two, one, go.

APPLAUSE

I'm, I'm very pleased that the laws of physics are nice and robust inside the Royal Institution. Whatever you do, of course, please don't try that experiment at home.

So the laws of classical physics are very well defined and computers can simulate these laws very accurately. For example, it's used in flight simulators. Flight simulators today are so accurate that an airline pilot that, who's training on a new type of aircraft, can actually do all of their training on the flight simulator. The first time they ever fly a real aircraft of that type, it's with fare paying passengers.

### Untangling the web

#### Lecture transcript

Download the full Untangling the web lecture transcript (PDF document 343KB – new window)

#### Video clip – High ranking

Using colourful water pipes, Chris illustrates the concept of search engines.

Now have you ever wondered how a search engine decides which page to put at the top of the list? Well it uses lots of different kinds of information, but let's just have a look at two of the most useful things. The first of these is called page rank and it tells us how important a webpage is and it makes use of these links. We can see how it works by looking at this water model. So each of these coloured tubes corresponds to one of these webpages and the pipes between the tubes correspond to the links. So we have a link from the yellow page here to the green page and over here, we have a tube from the yellow page, going across to the green page. Now we've started out with the same amount of water in each of those tanks. I'm going to go across here and switch on the pumps and it's going to pump water from one page to another, following those links.

Now in a moment, when the water levels settle down, the height of the water in each tube will tell us how important that webpage is.

Well already we can begin to see that the yellow page and the red page are not very important pages, the water is quite low. The reason is that these pages just have one incoming link. If you look across at the green page, we'll see that it's much more important and that's because it has two incoming links. The blue page, however is also important. Although it has only one incoming link, that link is coming from another page which itself is important.

Now of course – oh it's making a gurgling noise, (laughs) not me – on the real web of course, the computer has to solve this problem for millions of interconnected webpages at the same time.

### Digital intelligence

#### Lecture transcript

Download the full Digital intelligence lecture transcript (PDF document 285KB – new window)

#### Video clip – Are you smarter than a computer?

Chris reveals the results of the Cat or Dog challenge.

So earlier we took some photographs of Jack and Shadow and we added those photographs to a data set of 10,000 images of cats and dogs. Now half of those images are cats and half are dogs. And we took all those images and we fed them through a state-of-the-art object recognition system and we asked it to classify each image as either cat or dog. Now because there are equal numbers of cat and dog, if the system just guessed at random, it would get 50% correct.

Now the computer, taking about half a second to process each image, managed to get 83% correct, so it's doing a lot better than random, but it's still well short of being perfect. Now you might think why don't we get the world's most powerful supercomputer and use that? Well of course, if we did it would run a lot faster, but it wouldn't actually be more accurate. The problem is not lack of processing power, it's that we don't yet know how to use lots of processing power to achieve good recognition.

So how well do people do? Well on the website that's associated with this year's Christmas Lectures, we conducted an experiment. Humans had to classify images against the clock, and they took an average of one second per image and they managed to get 96% correct.

And how did our state-of-the-art recognition system do with Shadow and Jack? Well, here are four images of Shadow and Jack and what we found is that the bottom two it got correct, and the top two it got wrong.

So just telling the difference between cats and dogs is a challenge for present day computers. Thank you very much.