a*****g
发帖数: 19398 | 1 Deep Learning Isn’t a Dangerous Magic Genie. It’s Just Math
Deep learning is rapidly ‘eating’ artificial intelligence. But let’s not
mistake this ascendant form of artificial intelligence for anything more
than it really is. The famous author Arthur C. Clarke wrote, “Any
sufficiently advanced technology is indistinguishable from magic.” And deep
learning is certainly an advanced technology—it can identify objects and
faces in photos, recognize spoken words, translate from one language to
another... 阅读全帖 |
|
b*s
发帖数: 82482 | 2 a French mathematician born in Bourg-la-Reine. While still in his teens, he
was able to determine a necessary and sufficient condition for a polynomial
to be solvable by radicals, thereby solving a long-standing problem. His
work laid the foundations for Galois theory and group theory, two major
branches of abstract algebra, and the subfield of Galois connections. He was
the first to use the word "group" (French: groupe) as a technical term in
mathematics to represent a group of permutations. A ... 阅读全帖 |
|
o****y
发帖数: 790 | 3 【 以下文字转载自 Faculty 讨论区 】
发信人: outcry (outcry), 信区: Faculty
标 题: 学术界之风花雪月（5）食色，科研之性也！
发信站: BBS 未名空间站 (Sun Oct 24 02:24:48 2010, 美东)
disclaimer：学术八卦，文责不负：）
－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－
我认为，要真正理解学术界的精髓，只需读懂“食“和”色”两个字。您可能不解，怎么
学术科研这么“阳春白雪”的东东和“食色”这等俗务扯上啦？
为了不使我的讨论太突兀，first, allow me to formally introduce myself:)
我就是传说中的“三俗”哥。第一，我好吃。基本到了“食不厌精”的地步。曾经
我一红颜知己评价我“吃嘛嘛香，干嘛嘛行”，虽然我知道是忽悠，但仍然深以
为许：）我还特好酒。不管洋酒，国酒，我都喜欢。对酒，我不但舍得花钱，
而且舍得花时间去研究和鉴赏。In order not to digress, I'll save this
topic for another day.... 阅读全帖 |
|
V******9
发帖数: 1281 | 4 【 以下文字转载自 Faculty 讨论区 】
发信人: outcry (outcry), 信区: Faculty
标 题: 学术界之风花雪月（5）食色，科研之性也！
发信站: BBS 未名空间站 (Sun Oct 24 02:24:48 2010, 美东)
disclaimer：学术八卦，文责不负：）
－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－
我认为，要真正理解学术界的精髓，只需读懂“食“和”色”两个字。您可能不解，怎么
学术科研这么“阳春白雪”的东东和“食色”这等俗务扯上啦？
为了不使我的讨论太突兀，first, allow me to formally introduce myself:)
我就是传说中的“三俗”哥。第一，我好吃。基本到了“食不厌精”的地步。曾经
我一红颜知己评价我“吃嘛嘛香，干嘛嘛行”，虽然我知道是忽悠，但仍然深以
为许：）我还特好酒。不管洋酒，国酒，我都喜欢。对酒，我不但舍得花钱，
而且舍得花时间去研究和鉴赏。In order not to diverse, I'll save this
topic for another day.... 阅读全帖 |
|
S*******t
发帖数: 3956 | 5 【 以下文字转载自 Faculty 讨论区 】
发信人: SlowRabit (慢吞吞的小白兔), 信区: Faculty
标 题: 翻译：高妹对尹希教授的专访
发信站: BBS 未名空间站 (Sat Sep 26 11:16:17 2015, 美东)
（作为年仅31岁就晋升哈佛正教授的青年才俊，尹希教授最近得到了很多关注。其中一
篇关于他的专访进入了一名同是美国大学华裔教授的女科学家的视线。因为觉得这样的
采访不过瘾，她突发奇想，想亲自采访一下尹希。于是就有了这个我们今天看到的两名
科学家之间的对话。这位女教授使用了化名，这是她出版自己几部科幻小说的笔名。对
，没错，这位女教授同时也是一个女文青，已经写了几本书，还都是英文，Amazon 就
能买到。感兴趣的不妨找来看看。很可能有一天，她也会成为传奇式的人物。那今天这
篇采访就会成为另一段佳话。）
Interview with Dr. Xi Yin
--- By Fiona Rawsontile, Sept 2015
Translated by Slow Rabbit
Thi... 阅读全帖 |
|
a******g
发帖数: 938 | 6 我们8系本来就是当年的2系衍生出来的 本科教育还是很不错的吧 我自己觉得如果好好学
是能学到很多东西的 即使是偶这样混日子出来的 偶尔还是能在南大北大的同学面前吹吹
牛说当年欧们学了这个数学那个物理还有神经 Blabla... (不过细胞免疫什么的我们倒是
学的没他们多啊... )
UCarea说的不错 结构生物是我们系的highlight (偶自己就是结构生物的啦...可惜当年
申请的时候昏了头 非要申请分子啊神经啊心理学啊... 结果拿到手的offer都不好
好的offer又都是结构的... 晕啊...) 其实现在看来 我们USTC的数理背景那么好 做生物
物理本来就非常合适 像偶这种非要转subfield却没有什么基础的比起南大北大的同学就
差了一些... 现在倒有点嫁错郎的感觉了...
当然大牛无所谓 到哪里都会牛的 不像偶 当年就没好好学molecular and cellular...
现在学的好吃力啊...惭愧...
目前还不打算改行... 其实是偶现在虾米都没专长啊...不过偶知道以前欧们系的3个科大
的2个改行了... 偶要是计算机好偶也改行了...
有些fie |
|
c**h
发帖数: 368 | 7 ah? 有那么多学地理的么? 夸张了吧
当然找工作难易要看subfield. 地理中的GIS 是好找工作呀 |
|
r*****y
发帖数: 507 | 8 Interesting, CV has too many "so called top-tier" papers? >1000 top-tier
papers PER YEAR in a subfield??
BMVC四大会议，再来两篇PAMI/IJCV，就算大功告成了。
>1000 papers per year on those places.
total less than 150 papers per year,
IEEE T Med Image+Elsevier Med Image Analy。 |
|
s******g
发帖数: 3841 | 9 网络、OS这么大的领域一年30篇top conf papers（Sigcomm, OSDI/SOSP）
CV一个subfield一年1000篇top conf papers?! |
|
r*****y
发帖数: 507 | 10
not several, I would say, in each subfield, its top 5 candidates take
most of the interviews from good CS programs. |
|
d*****t
发帖数: 7903 | 11 想学习一下，有电子版最好
还有，想请教一下computer vision 和photogrammetry 的关系。我觉得
photogrammetry 是不是个subfield啊？谢谢 |
|
j*****n
发帖数: 1545 | 12 一句话很难概括 Wiki 上 说的很好了
Semidefinite programming (SDP) is a subfield of convex optimization
concerned with the optimization of a linear objective function over the
intersection of the cone of positive semidefinite matrices with an affine
space。
Boyd那本书有讲 |
|
z****e
发帖数: 54598 | 13 这种东西，wikipedia上随便找就有
1 Subfields
1.1 Mathematical foundations
1.2 Algorithms and data structures
1.3 Artificial intelligence
1.4 Communications and Security
1.5 Computer architecture
1.6 Computer graphics
1.7 Concurrent, parallel, and distributed systems
1.8 Databases
1.9 Programming languages and compilers
1.10 Scientific computing
1.11 Software engineering
1.12 Theory of computation |
|
a*****g
发帖数: 19398 | 14 Deep Learning Isn’t a Dangerous Magic Genie. It’s Just Math
Deep learning is rapidly ‘eating’ artificial intelligence. But let’s not
mistake this ascendant form of artificial intelligence for anything more
than it really is. The famous author Arthur C. Clarke wrote, “Any
sufficiently advanced technology is indistinguishable from magic.” And deep
learning is certainly an advanced technology—it can identify objects and
faces in photos, recognize spoken words, translate from one language to
another... 阅读全帖 |
|
p*****y
发帖数: 133 | 15 most schools do not have these specific classification of majors for
accounting. just master of accountancy or master of taxation. For
accountancy, you can do either audit or tax, or corporate accountant.
Information system audit and control sounds like IT auditing, another
subfield of audit. |
|
w*******e
发帖数: 63 | 16 入门就是第一本吧，写得比较有意思。第二本比较深，而且有些过于琐碎。后面两本都
是进化的
subfield了。对molecular evolution感兴趣可以看李文雄的那本。 |
|
d*****r
发帖数: 2583 | 17 This is only cell biology field?
you can first sort all the journal results, don't choose any subfield, and
then export to text, each time you can export 500 journals. |
|
x********u
发帖数: 430 | 18 Can you export the subfield for Biotechnology? Thanks. |
|
h******e
发帖数: 44 | 19 关于1. 我可以给你举一个反例，不代表我支持或反对这种做法。实际上CNS的editor，
尤其是cell，有对reviewer的veto right。在极端情况下，即使三个reviewer都反对，
但editor支持，他会直接发。Cell在八九十年代有非常多这种例子。现在NS的subfield
editor可能只是postdoc出身，没有这种魄力和判断力，但以前的资深editor有。这是
Cell在八九十年代迅速崛起的一个重要原因，它能发表当时具有争议、判断困难、但有
潜在重要性的文章。
其实elife不需要二轮评审的主要原因是大家都觉得revision过程太繁琐了。加了一大
堆supplementary figure上去，各种画蛇添足，一篇Nature逼成JCB，其实大家看的还
不就是那四五个图，后面那些东西根本没什么必要在一篇文章里面搞完。这样搞下去，
可能不利于科研进步。
也没说所有杂志都这样，只是elife用这种形式而已，给一个出口渠道，让一些在一般
情况下发不出或是要拖很久的文章尽快出来。别的杂志和经典的revision流程又不会消
失。
的* |
|
o*******8
发帖数: 114 | 20 哈哈哈，
and you are absolutely sure you mean what you are talking about?
"其实elife不需要二轮评审的主要原因是大家都觉得revision过程太繁琐了"
-太精屁了，我的倾佩之情，正如滚滚长江东逝水，又如无边落木萧萧下， 一发不可收
拾啊。
subfield |
|
p**t
发帖数: 160 | 21 让一些在一般情况下发不出或是要拖很久的文章尽快出来
──────────
这个有没有选择性？
subfield |
|
h******e
发帖数: 44 | 22 关于1. 我可以给你举一个反例，不代表我支持或反对这种做法。实际上CNS的editor，
尤其是cell，有对reviewer的veto right。在极端情况下，即使三个reviewer都反对，
但editor支持，他会直接发。Cell在八九十年代有非常多这种例子。现在NS的subfield
editor可能只是postdoc出身，没有这种魄力和判断力，但以前的资深editor有。这是
Cell在八九十年代迅速崛起的一个重要原因，它能发表当时具有争议、判断困难、但有
潜在重要性的文章。
其实elife不需要二轮评审的主要原因是大家都觉得revision过程太繁琐了。加了一大
堆supplementary figure上去，各种画蛇添足，一篇Nature逼成JCB，其实大家看的还
不就是那四五个图，后面那些东西根本没什么必要在一篇文章里面搞完。这样搞下去，
可能不利于科研进步。
也没说所有杂志都这样，只是elife用这种形式而已，给一个出口渠道，让一些在一般
情况下发不出或是要拖很久的文章尽快出来。别的杂志和经典的revision流程又不会消
失。
的* |
|
p**t
发帖数: 160 | 23 让一些在一般情况下发不出或是要拖很久的文章尽快出来
──────────
这个有没有选择性？
subfield |
|
O******e
发帖数: 4845 | 24 我推荐大家都读一下这篇文章：
http://www.pnas.org/content/111/16/5773.full
全文：
By Bruce Alberts, Marc W. Kirschner, Shirley Tilghman, and Harold Varmus
By many measures, the biological and medical sciences are in a golden age.
That fact, which we celebrate, makes it all the more difficult to
acknowledge that the current system contains systemic flaws that are
threatening its future. A central flaw is the long-held assumption that the
enterprise will constantly expand. As a result, there is now a severe
imbalan... 阅读全帖 |
|
i*****k
发帖数: 25 | 25 I think that most of the professors here are doing research in operation
research or operation management, or Econometrics. Of course they should be
good at maths. There are other research areas in business which don't need too
much maths background, like OB, Accounting,even finance and marketing.
Although there are a lot of people in marketing and finance who are doing a
lot of maths work, there are many research subfields in these areas which
don't need so much maths background. It really depe |
|
e*******r
发帖数: 27 | 26 as far as i know, risk management is a subfield of
finance, particularly financial engineering.
even you decide to join the finance program, some
prof. are indeed doing research in risk management.
if the risk management offer you got is from stat/math
dept, i suggest you accept the finance offer. because
there are much more areas which may attract you. |
|
h*****0
发帖数: 145 | 27 New AEA Journals
Following a three-year study and discussion of the subject, the Ad Hoc
Committee on Journals (Robert Hall [chair], Judith Chevalier, David Colander
, Peter Diamond, Alan Krueger, and Daniel Rubinfeld) recommended to the
Executive Committee in April 2006 that the Association start four aggregated
field journals, each published in four issues per year. Each new journal
will publish articles in a group of subfields with loose boundaries. The new
journals are not subsidiary journals |
|
f******h
发帖数: 37 | 28 You just compete with more econometricians than in China, all subfields of
economics are similarly competitive right now in the USA. I feel it is about
10 candidates for 1 opening, which won't change much in 2 or 3 three years |
|
g****n
发帖数: 1 | 29 比如好找工作，好发文章。最好的杂志是什么，多谢了 |
|
d****e
发帖数: 251 | 30 IP里的litigation对语言要求是否和其他subfields一样？ |
|
A****s
发帖数: 129 | 31 For A is a element of A
just avoids it? or there is some subfield about it? |
|
B****g
发帖数: 43 | 32 What subfield? There is no such book summarizing the recent developments or
future directions for today's math.
If there is one, I refer you the Proceedings of the ICM, it has a lot of
nice
reviews from the best experts in today's math. |
|
j*****h
发帖数: 2577 | 33 多谢，建议很中肯
我不是太懂数学，托付的人也没说什么subfield，这点也让我很为难
有什么比较好的textbook么？由浅入深，等等？
or |
|
B****g
发帖数: 43 | 34 He or she must specify the subfield, otherwise we can't help. |
|
H*********r
发帖数: 659 | 35 Lawrence Shepp is Larry Shepp, as William Clinton is Bill Clinton.
Shepp was in Columbia, and he doesn't need to be the expert doing Phong does.
Yau isn't expert in that subfield either, he was later invited to involve. Shepp questioned how Phong behaved inappropriately as a professor/advisor. |
|
a*****g
发帖数: 19398 | 36 Deep Learning Isn’t a Dangerous Magic Genie. It’s Just Math
Deep learning is rapidly ‘eating’ artificial intelligence. But let’s not
mistake this ascendant form of artificial intelligence for anything more
than it really is. The famous author Arthur C. Clarke wrote, “Any
sufficiently advanced technology is indistinguishable from magic.” And deep
learning is certainly an advanced technology—it can identify objects and
faces in photos, recognize spoken words, translate from one language to
another... 阅读全帖 |
|
|
|
J**a
发帖数: 215 | 39 From Wikipedia, the free encyclopedia
Soft matter or Soft condensed matter is a subfield of condensed matter
comprising a variety of physical states that are easily deformed by thermal
stresses or thermal fluctuations. They include liquids, colloids, polymers,
foams, gels, granular materials, and a number of biological materials. These
materials share an important common feature in that predominant physical
behaviors occur at an energy scale comparable with room temperature thermal
energy. At th |
|
|
k**i
发帖数: 6 | 41 happle那个ID说，tenure的教授应该H > 18。但是物理不同领域，以及同一领域的理论
和实验的情况差别都太大，实在很难理解会用这么个一刀切的标准的来衡量。
说实话我从来就不理解为什么那么多人喜欢提这个H-index。Hirsch这个人已经多年不
正经作research，搞这个index出来玩也就罢了，竟然还给出一个用这个index衡量各种
教授的标准。也不知道他是物理的什么subfield里面，根据什么提出来的这个标准，样
本空间有多大。 |
|
d*********p
发帖数: 193 | 42 啊，原来这个版上还有学cosmology的人啊？原本以为这里的老中们没有人学cosmology
这种没有钱途的subfield呢~ |
|
P*****s
发帖数: 375 | 43 A nice interview I just read is being shared with you guys here. :)
[ref]
http://fionarawsontile.blogspot.com/2015/09/interview-with-dr-x
Interview with Dr. Xi Yin
By Fiona Rawsontile, Sept 2015
This interview was inspired by an earlier interview of Dr. Yin I saw on the
Internet, which made me think that we can’t expect someone who normally
writes for entertainment to understand a physicist. To “provoke” a
scientist, we need another scientist. So I volun... 阅读全帖 |
|
s**********n
发帖数: 868 | 44 Based on the papers I've read from Nature Communications in our subfield, I
think their average quality is better than the average for PRL. NC benefits
a great deal from taking very good papers beaten down from Nature and
Nature Physics. It has grown in size but is still much smaller than PRL (~
1000 physics papers a year vs ~3000 for PRL). The $5000+ fee is outrageous
though, and I bet it is hurting the journal quite a bit.
I know people like the legacy and "purity" of PRL, and it is sad th... 阅读全帖 |
|
s******1
发帖数: 239 | 45 The debate between Rational Choice vs Other Approaches is also a "frequently
revisited and unfalsifiable topic." In political science research, there is
always a tension between rigor and richness (or formality vs. substance).
As long as Political Science Department is not renamed as Mathematics
Department, such a tension will go on and no final winning empiror.In
American politics subfield, RCT is a dominant approach,but in IR maybe not
the case. The key is to find an approach that best fits y |
|
k***g
发帖数: 7244 | 46 不过感觉 IR 文章的比例严重低于其他各个subfield，不知道何故？还是 IR 的高质量
的paper都去投 IO 和 IS了？ |
|
s******1
发帖数: 239 | 47 Weather forecast is also complicated, and natural scientists would not be
surprised to predict it with high accuracy all the time. The question is: to
what extent should Pser take natural science as the model? What other
things can PSer predict other than voting behaviors in American Politics
subfield? Or can PSer predict at all?As an overall trend, political science
can describe, can explain, but it is almost impossible to predict?
Politics is sure to be more complicated than clock, maybe to so |
|
s******1
发帖数: 239 | 48 Previously PS relies on cognitive and computer sciences to simulate human
thinking (this approach is called "cold cognition" and Herbert A. Simon is
the founder).Now some scholars rethink about human nature from the latest
development of neuroscience and emphasize the emotional aspect of human
decision making (this approach is called "hot cognition"). This "hot
cognition" approach is trying to revolutionalize American Politics and
International relations subfields.For international relations, y |
|
