Name: 計算機科學50年。Brailsford教授問答--Computerphile(計算機愛好者) (50 Years of Computer Science: Professor Brailsford Q&A - Computerphile)
Uploaded: 2021-01-14T10:55:32.000Z
Duration: 18 min 41 s
Description: 【看影片學英語】數萬部 YouTube 影片，搭配英漢字典即點即查，輕鬆掌握單字發音與用法，長久累積看電影不必再看字幕。

>> Sean: We have had some questions in from some of the fans, some of the viewers of Computerphile

過去簡單式

And we thought it would be a be nice idea to have a

chat and you tell us what you think. So, I've picked a few [questions] that maybe hopefully will

go together and and we may well do some others at another time. We're going to kick

off with this one here from Patrick Orton, who I think you may actually ...

>> DFB: Yes! Hi Patrick! He was one of my project students a year or two ago. Yeah! Hi Patrick!

>> Sean: He's asked four questions. So we'll whip through these before getting on to some others.

So: what is your favourite area of computer science ? >> DFB: my favorite area of

computer science? Well, it's  split into two really. Although it's very

very applied computer science I really do like everything to do with digital

documents - right through from the sort of XML structural stuff right down to the

nitty gritty typesetting, fonts and PostScript stuff. I love that. But if you

were to say is there an area I like doing videos on which fascinates me

- although it's not really my specialist area - the answer is, what we're doing at

the moment, which is Regular Expressions. I think they're fascinating! They're so

easy but so difficult. You get Regular Expression wars between commenters.

They're wonderfu!lThere's a way to go yet on REs before we finally say

"That's enough!" Yeah,  >> Sean: well, it's interesting because you kind of roll these first two

questions together, so we'd ask the favourite area to work in and the area you're most

interested in. What's the area of computer science that you'rer best at?

>> DFB: Do you know that's a very good question  Patrick. I can tell you are one of the

students that I supervised. How well you  know me! The answer, truthfully, is that I

don't consider myself an outstanding computer scientist in any of the areas.

I suppose you could say, insofar as I have some expertise, the whole digital

documents / electronic publishing, low-level, PostScript/PDF stuff, yes.

But actually what am I really best at? It finally dawned on me, and thanks to you

and Brady, you have given me the ability to exploit this. I'm best at explaining

things to people and, as I've often said, I, when I was younger, had to teach myself

and there's [potentially] a whole extra episode on why,  to be competent in a wide range of things.

And being forced to teach something is the best way to learn it. So I think I am

good at that. The problem I had was if I was like Brian Kernighan and had the

tenacity to write lots of books, that would have been a way to propagate my

expertise. I don't really have the patience to write books. What I wanted

was somebody to come and just talk to me and make videos from my waffle! And you

and Brady - honestly I remember that first visit, and it's been solidly downhill for

seven years ever since then, hasn't it? 2013?  something like that yeah?!

>> Sean: Patrick also had a question, which connects  to a couple of other questions, so I'm going to roll

these together if that's all right with you. What is the area of computer science

that looks to be most important in the near future?  which connects to one

Richard Denton's asked about " ... your extensive history in the new field of

computer science, what direction is it heading in and Gregor who said: " ... which

area of CS will in the near future be used the most and which do you think

there's no future to? >> DFB: I'm not very good at futurology!  I could list all of the

things which I've come across in computing and have said " ... that will *never*

catch on!"  But then, six months later, it's  the lead

topic in the whole field! Now I think, all right, let's just say recently the thing

this has quite taken my breath away has been the whole stuff about Neural Nets

Deep AI and so on. And now it's fascinating  because ... I'll tell you what is the

most fascinating thing. You had Deep Blue playing chess basically by brute force

methods [looking mny moves ahead] and back in  the 80s -- I don't play chess by the way but I  do find it fascinating --

it was Deep Blue, IBM's Deep Blue, I think, that defeated Kasparov for

the first time, Wow! you know. And I felt: "Well it's just brute force and ignorance that [is]"

But on the other hand the moment you start doing neural networks, which

learn and adapt and once you start feeding in the fact that you can play

yourself in a computer rendering of the game. You know what they've done on GO

and chess and all that, now, is to say we don't just hard-code the rules and

look thirty moves ahead. We'd get it to find out what unlikely things have

come up, in me playing myself, where it just went berserk and went off on a tangent

And I thought " ... that's a lost game" But it wasn't! It found a back-door way in and

triumphed. And I think that's what took the GO guy - was it Lee Sedol -  apart.

He just thought:" ...  that's ridiculous I'm gonna win this one!" Then, about 20 moves later

he hadn't! And I think he found that shattering, and I can well imagine that.

And I think my understanding is very limited of this - is that the Deep Mind

people have now started applying all that they did to GO to apply it back to

chess. And don't use brute force just build up millions of games and let it

learn what's a good thing to do. And I think I'm right in saying that that

piece of software, using the neural net statistical whatever, is now more

successful than the original Deep Blue was, but I'm not sure. So what you

can say is [that] taming big data via AI  is a tremendous area. It's I mean Big

Data is important and the degree to which you can tame it with so-called

AI is fine but the big question still is [this]. You go to the authors of these

"highly intelligent" programs but they're not -  and you say: "Can it do 'common sense' ? "

"No! not a hope! it's brilliant at go because it's still a narrow and

well-defined field but don't ask it to pontificate on the state of politics in

the Western Hemisphere". It hasn't a clue! So, there we are, right, something that I know

I'm interested in. I know very little about but does seem to me to be a real

marker for the future. >> Sean: Maybe onto a different subject.  We had a question from

Erik Stens and he said: "As engineers we all have our pet project. What have been

some of your hobby projects - computer  related - that you're most proud of.

And would you tell us a bit about them all? >> DFB: Right! Well, I suppose I should say at the

start that one of the advantages I've got being the ripe old age I am, is that

I've been what's called an Emeritus Professor for well over ten

years now and that this does give me the ability to pursue lunatic things that

would never ever have got me promoted to Professor but I was there already.

>> Sean: can you tell people what 'emeritus' means  For those who don't know those who aren't in academia?

>> DFB: What happens with Emeritus Professors is that it's basically a way of pension ah--you

off but in a gentle way. The idea of an Emeritus Professor - the Latin of course

means 'E' = 'out of 'merit. i.e. "because of merit" You are given this honorary position. It

varies very much as to which university are in as how active an Emeritus you

want to be, but for me, back in 2005, I came off the regular payroll

as a professor and 'cut a deal', as it were, where I will be brought back on a

renewable yearly contract to do fill-in teaching on stuff I likeds but the good

thing was: " ... with no major administrative responsibilities". So I took my pension I

got some extra top-up from the School of Computer Science and really, I think, the

last actual taught course I did, given  that I formerly 'went Emeritus' in 2005,

I was still teaching classes through to 2013. And why did I change

tack then? Well, once upon a time, two hirsute guys came into my office and said:

"Hello, we're Brady and Sean. We're going to reshape your future" you've wanted to

find somebody idiotic enough to record your waffle?

That's us!  Well, it's true isn't it ? I always love the teaching but to be able

字幕列表影片播放

計算機科學50年。Brailsford教授問答--Computerphile(計算機愛好者) (50 Years of Computer Science: Professor Brailsford Q&A - Computerphile)

stuff

expression

basically

fascinating