c***c
发帖数: 21374 | |
c***c
发帖数: 21374 | 2 AI严格意义上也不算是CS的一个范畴，应该是个跨学科的领域吧。
在AI的几个领域当中，我对于ML（machine learning）和NLP（natural language
process)最感兴趣。
不知道大家怎么以为？
最近economist登了一篇文章，描述了一种律师专家系统，我对这类的应用很是感兴趣，
而且觉得也是前途巨大。
在游戏业界，现在AI的应用也是如火如荼，并且因此催生了很多对于理论创新的要求。
等等，很多了。总之我觉得这个是一个前途钱景都巨大的行业。
【在 c***c 的大作中提到】
|
r*****y
发帖数: 507 | 3 AI is the current bottleneck for gaming industry.
most of the so-called AI algorithms(except A*) can't be used practically at all.
【在 c***c 的大作中提到】
: AI严格意义上也不算是CS的一个范畴，应该是个跨学科的领域吧。
: 在AI的几个领域当中，我对于ML（machine learning）和NLP（natural language
: process)最感兴趣。
: 不知道大家怎么以为？
: 最近economist登了一篇文章，描述了一种律师专家系统，我对这类的应用很是感兴趣，
: 而且觉得也是前途巨大。
: 在游戏业界，现在AI的应用也是如火如荼，并且因此催生了很多对于理论创新的要求。
: 等等，很多了。总之我觉得这个是一个前途钱景都巨大的行业。
|
l*****k
发帖数: 1059 | 4 AI is one of the most undeveloped area in CS. I have a kitten who is
only 3 months old. She learned how to use litter box, how to cry for
foods and her other requirements. She is also trained to keep away from
our furnitures, and have reactions to some of my basic signals. Have you
heared of any AI program can do the similar thing?
The first approach in AI is try to use rule-based knowledge to describe
the real world and solve complicate practical problems, and thus Expert
systems are developed
【在 c***c 的大作中提到】
: AI严格意义上也不算是CS的一个范畴，应该是个跨学科的领域吧。
: 在AI的几个领域当中，我对于ML（machine learning）和NLP（natural language
: process)最感兴趣。
: 不知道大家怎么以为？
: 最近economist登了一篇文章，描述了一种律师专家系统，我对这类的应用很是感兴趣，
: 而且觉得也是前途巨大。
: 在游戏业界，现在AI的应用也是如火如荼，并且因此催生了很多对于理论创新的要求。
: 等等，很多了。总之我觉得这个是一个前途钱景都巨大的行业。
|
b***n
发帖数: 53 | 5 undeveloped谈不上.
所有战线都有一堆天才在往前推. 推不动, 说明难推.
【在 l*****k 的大作中提到】
: AI is one of the most undeveloped area in CS. I have a kitten who is
: only 3 months old. She learned how to use litter box, how to cry for
: foods and her other requirements. She is also trained to keep away from
: our furnitures, and have reactions to some of my basic signals. Have you
: heared of any AI program can do the similar thing?
: The first approach in AI is try to use rule-based knowledge to describe
: the real world and solve complicate practical problems, and thus Expert
: systems are developed
|
l*****g
发帖数: 49 | 6 How much do you know about CS theory when you say "there is no theory
in CS"?
And successful applications of CS go far beyond programming and MS ...
For example, the RSA encryption scheme is used on the internet (for electronic
transactions, etc.) everyday. It originated within CS theory.
在学科理论上，
CS看不到突破，说透了也就无理论和规律可言，在应用上，"CS"的成功只是发生在编程上
（microsoft, orcle, GNU），掌握那些经典的算法根本不需要在学校里浪费那么多年，这 |
w*******g
发帖数: 9932 | 7 people will say that crytography is a branch of mathemtics.
CS people like to include cryto as part of CS theory but not many CS people
is really good at number theory.
【在 l*****g 的大作中提到】
: How much do you know about CS theory when you say "there is no theory
: in CS"?
: And successful applications of CS go far beyond programming and MS ...
: For example, the RSA encryption scheme is used on the internet (for electronic
: transactions, etc.) everyday. It originated within CS theory.
:
: 在学科理论上，
: CS看不到突破，说透了也就无理论和规律可言，在应用上，"CS"的成功只是发生在编程上
: （microsoft, orcle, GNU），掌握那些经典的算法根本不需要在学校里浪费那么多年，这
|
r********n
发帖数: 75 | 8
I read a kind of same thing 10 years ago, maybe in communication of acm.
【在 c***c 的大作中提到】
: AI严格意义上也不算是CS的一个范畴，应该是个跨学科的领域吧。
: 在AI的几个领域当中，我对于ML（machine learning）和NLP（natural language
: process)最感兴趣。
: 不知道大家怎么以为？
: 最近economist登了一篇文章，描述了一种律师专家系统，我对这类的应用很是感兴趣，
: 而且觉得也是前途巨大。
: 在游戏业界，现在AI的应用也是如火如荼，并且因此催生了很多对于理论创新的要求。
: 等等，很多了。总之我觉得这个是一个前途钱景都巨大的行业。
|
w*******g
发帖数: 9932 | 9 algorithm/complexity
logic/lamba calculus |
i******y
发帖数: 70 | 10 You mentioned Oracle.
As the foundation of databases, both relational model and ER model come from C
S research;
Dynamic-programming query optimization, used by many commerical DBMS such as SQL server, DB2 and Oracle, comes from CS research;
B-tree index and R-tree index, used in every DBMS, comes from CS research; Dat
abase transaction management and concurrency control, applied everyday in bank
s, come from CS research;
Data cube, a model in almost every industrial data mining product, comes f |
|
|
r*****y
发帖数: 507 | 11 not sure your so general conlcusion.
at least, Graphics achieves its Great success in reality! |
r*****y
发帖数: 507 | 12 because many fields use math and philosophy, can we say most of the research
fields are interdesciplinary? so..
interdesciplinary research is the right way to go for CS too. |
l*****g
发帖数: 49 | 13 I agree with wildThing's list on CS theory fields:
algorithms
computational complexity
logic/lamba calculus
Then, how much do you know? Can you give example-list on your so-called CS the
ories? I am glad to see them. :) |
l*****g
发帖数: 49 | 14 Cryptography is of course part of CS.
If not, then it belongs to what?
cryptography is not CS at all.
CS is used to implement crypto. that's it. |
l*****g
发帖数: 49 | 15 1. CS is not a branch of applied math. CS has its engineering part
such as Systems, Networking, etc.
2. Even we talk about programming: they more or less use data
structures and algorithms.
(though most programming in practice use only very very simple
data structures and algorithms).
i would say, CS is a samll branch of applied math.
ask, how many IT positions are doing with algorithms/lamda calculus? how many
percent in IT job market require that ?
with programming skills, lots of |
l*****g
发帖数: 49 | 16 Google itself is technology, but it wouldn't exist without
all those CS research on storing/processing/querying
huge datasets.
> Google, comes from CS research.
also technology, not science |
i******y
发帖数: 70 | 17 The stuffs in my list come from CS research, as a clear constrast to what you
have said,
"
在应用上，"CS"的成功只是发生在编程上
（microsoft, orcle, GNU），掌握那些经典的算法根本不需要在学校里浪费那么多年，这
些和学校里的CS 学究们 八杆子打不着。
"
What I know is, E.F. Codd got Turing Award for Relational Model and Jim Gray g
ot
it for transactions. They may appear to be superficial to you, they may not be
as deep as math. But they are great successes of CS research and have definite
ly
had deep impact on the life of human begins. It does not matter whether |
l*****g
发帖数: 49 | 18 But it has something to do with "computer science".
Cryptography is just Cryptography. Its theory and method has nothing to do wit
h "computer". |
r*****y
发帖数: 507 | 19 you are talking about coder (or programmer) career... it is not science of course. it is a labor worker in CS industry -- there are many similar labor worker jobs in other industry too. |
i******y
发帖数: 70 | 20 Google comes from CS research. Right?
Google is a success. Agree?
Isn't it different from what you said (shown below)?
"
在应用上，"CS"的成功只是发生在编程上
（microsoft, orcle, GNU），掌握那些经典的算法根本不需要在学校里浪费那么多年，这
些和学校里的CS 学究们 八杆子打不着。
"
So what is your point by saying Google is technology instead of science? Does
it
matter? Does it justify what you have said about the success of CS application
s? |
|
|
i******y
发帖数: 70 | 21 So what are we talking about? Computer Science or IT workers? Or if Computer
Science can secure a job for IT workers?
I don't see any connection between this and what you have said (shown below).
"
最近CS版上 开始讨论一些看似轻松 实则很深刻也很忧心重重的问题： 《计算机科学的定
位》 《AI怎么样》。
可以这样说，这反映了广大CS人的困惑和内心深处对自己前途的迷茫，不论大家本人是否
能爽快的承认对自己学科前途的没有信心和怀疑。所有的人都在外人面前装出一副自己的
学科很行，很需要常年积累的expertise的架势，其实他干的那破活儿，是个人 几个星期
就可以拿下。没办法，就是这么虚伪，教授也好，phd student也好，都要吃饭。
本人作为一个CS人，直觉上已经深感CS象伪科学，也后悔陷入这个粪坑；在学科理论上，
CS看不到突破，说透了也就无理论和规律可言，在应用上，"CS"的成功只是发 |
l*****g
发帖数: 49 | 22 1. add one more to your list: theoretical computer science
2. There are some basic principles and techniques that connect all these field
s together to form "computer science", for example (among other things):
the duality between program and data;
abstraction;
recursion;
tractability;
virtualization;
fault tolerance;
These principles/techniques wouldn't stand out without computer science.
They bear the hallmarks of CS.
actually, the entity of "computer" "science" does not exist |
w*******g
发帖数: 9932 | 23
without algorithm/lambda calculus, there won't be any IT positions.
how do you reason the correctness of an algorithm without analysis?
how do you prove the soundness of your program without lambda?
theory is invented to abstract away the details of daily work and
focuses on proving the essential properties of interests.
Theory is fun but only in the hands of talented people and I am not
one of them... |
l*****g
发帖数: 49 | 24 This is another issue then.
There are more industry positions than research positions does not mean cs is
not science. It just means CS has both its science and engineering sides.
currently, in CS industry, how much people are NOT such labor?
Except for programmer labor positions, how much research positions are there? |
i******y
发帖数: 70 | 25 My question is:
Are we talking about "if Computer Science can secure a job for IT workers?"
My question is not:
If Computer Science can secure a job for IT workers?
And again, what are we talking about? Computer Science or IT workers?
Or if Computer Science can secure a job for IT workers?
If we are talking about IT workers and their jobs, how come you end up with th
is
new topic while we were making comments on your conclusions (shown below)? Why
should we talk about it at all?
"
最近CS版上 开始讨论一些看 |
e*****n
发帖数: 39 | 26 PhD in all disciplines are risky. Phy/Chem are much worse. |
e*****n
发帖数: 39 | 27
Don't think so. Research is trying to address problems, not necessary related to math. If the idea works, it is good research result, no matter how it looks like "cheating". |
e*****n
发帖数: 39 | 28
Why is it especially for cs? |
R****r
发帖数: 227 | 29 they will can do everything... |
r*****y
发帖数: 507 | 30 agree with you.
nothing is easy and anything could be risky.
for CS Ph.D, if you are good enough, you have no problem to stay in the academia, or doing research in industry labs (not coder or programmer position). Otherwise, if you fail the competition, then IT labor workers are the way you go. For those BS/MS, IT lab workers are their targets (of course, very few ppl are just very exceptional and can run their own business, that's another story, not
covered here).
The world/life is full of comp
【在 e*****n 的大作中提到】
: PhD in all disciplines are risky. Phy/Chem are much worse.
|
|
|
r*****y
发帖数: 507 | 31 what' do you mean CS ph.D popluation?
as I remember, the total number of CS Ph.D graduate in US is less than 1K per year (including all universities). Is this a big number or counted a population?
or this number is too large already to make it a group difficult to survive in the world?
Why must we be big cows? how about small cows, or half-cows? it doesn't matter, if you can do something matched your research interests and also make a decent life, that's good enough. Don't tell me this kind of |
c***c
发帖数: 21374 | 32 所以说AI和认知科学联系密切呢
【在 l*****k 的大作中提到】
: AI is one of the most undeveloped area in CS. I have a kitten who is
: only 3 months old. She learned how to use litter box, how to cry for
: foods and her other requirements. She is also trained to keep away from
: our furnitures, and have reactions to some of my basic signals. Have you
: heared of any AI program can do the similar thing?
: The first approach in AI is try to use rule-based knowledge to describe
: the real world and solve complicate practical problems, and thus Expert
: systems are developed
|
i*f
发帖数: 85 | 33 不是说AI is dead了吗
【在 c***c 的大作中提到】
|
e*****n
发帖数: 39 | 34
If you want to be a faculty, phd is a must.
If you are not big cow, you will not be very successful in industry too.
Academia is very competitive. You must be a top phd to find a position in a research university.
If you don't like doing phd, just quit now. |
r*****y
发帖数: 507 | 35
maybe they plan to apply faculty later? or also depends directions, I know some CS directions basically no industry, so if they can't land a faculty position, then their lives could be tough.
So if you think you are not strong enough to surive in the academia, then don't pick those narrow CS fields (e.g. CS theory). Or don't need to go ph.d. for this kind of directions (it is basically like a Math major, the career is the same too, the general CS glory name doesn't help at all). |
c**********t
发帖数: 99 | 36 CS research focuses on two areas: 1) the nature of computer itself (compiler/O
S/architecture/language...) 2) the object which the computer deals with,the op
erations over the object (image processing/graphics/data compression/cryptogra
hy/networking ....).
The object which the computer deals with is data or information. How to repres
ent data in different info domains is the research in the following areas: gra
phics, image processing, speech, data compression, crypto, HCI. How to communi
cate |
c**********t
发帖数: 99 | 37 One major difference between MATH and CS lies in the data which MATH and CS ha
ndle: the data MATH dealing with is abstract, meaningless while the data CS de
aling is meaningful data --- information. So that's the reason why crypto is a
subbranch of CS instead of Math.
【在 c**********t 的大作中提到】
: CS research focuses on two areas: 1) the nature of computer itself (compiler/O
: S/architecture/language...) 2) the object which the computer deals with,the op
: erations over the object (image processing/graphics/data compression/cryptogra
: hy/networking ....).
: The object which the computer deals with is data or information. How to repres
: ent data in different info domains is the research in the following areas: gra
: phics, image processing, speech, data compression, crypto, HCI. How to communi
: cate
|
s****a
发帖数: 21 | 38
what branch of crypto do u mean here
【在 c**********t 的大作中提到】
: One major difference between MATH and CS lies in the data which MATH and CS ha
: ndle: the data MATH dealing with is abstract, meaningless while the data CS de
: aling is meaningful data --- information. So that's the reason why crypto is a
: subbranch of CS instead of Math.
|
