c*****y
发帖数: 3 | 1 This is what "Formal Semantics"(形式语义) tries to solve. Formal methods
such as regular expressions, automatas can describe syntax pretty good
already. Describing semantics is much more difficult.
大家 |
|
|
|
x******t
发帖数: 6 | 4 查了篇文章Loop-Free Alternating Finite Automata，清楚了。
LTL的表达能力与star-free regular expression相同。所谓的“star-free的regexp语
言”虽然不含有*，却包含negation和intersection，因此它可以表达某些无限长的串。 |
|
z*********h
发帖数: 133 | 5 现在做的工作需要从sample language sentence induct grammar
language可能是很简单的
请问有什么这方面的现成的tool吗？（实验阶段的也可以，不一定要industry grade的）
先谢了！ |
|
t**k
发帖数: 260 | 6 search "grammar induction" or "grammar inference"
A lot has been done on regular grammar induction. |
|
w**l
发帖数: 49 | 7 计算的美丽–1976年图灵奖获得者Michael Rabin
作者 陈怀临 | 2008-03-20 05:58 | 类型 人物评述, 计算的美丽 | 没有留言 »
Michael Oser Rabin (1931–)
图灵奖获得时间：
1976年。 第十一位图灵奖(1976年)获得者。
图灵奖引用(Turing Award Citation) ：
For their joint paper “Finite Automata and Their Decision Problem,” which
introduced the idea of nondeterministic machines, which has proved to be an
enormously valuable concept. Their (Scott & Rabin) classic paper has been a
continuous source of inspiration for subsequent work in this field.
【笔者译：】
“( 授 予 Michael O. |
|
y****i
发帖数: 57 | 8 工科算法编程较多，想系统学点计算机课程。
觉得编程是工具，“善其事，利其器”。现在很多专业，都是编程实现。
觉得以前都是自己摸索编程，只要能运行就行。但是肯定不够。在运行的基础上，还要
编好。
选computer theory有用吗？
看了一下目录，觉得好像没有用。哪位科班的来讲讲。
Finite automata, regular expressions, grammars, languages, Turing machines,
computability, |
|
c*****t
发帖数: 1879 | 9 这学校没有 database / compiler 方面的人。undergrad 课程没有 automata
和 compiler 。看来师资人员太少啊。 |
|
c*****t
发帖数: 1879 | 10 这学校没有 database / compiler 方面的人。undergrad 课程没有 automata
和 compiler 。看来师资人员太少啊。 |
|
c**0
发帖数: 103 | 11 Compilers
Computer Systems Architecture 很久以前在国内上过
Digital Logic and Computer Design 很久以前在国内上过
Theory of Computation
Automata Theory
Principles of UNIX
Discrete Structures 搞定
Algorithm Analysis & Data Structures I and II 搞定
Networks 搞定
Operating Systems Concepts I and II 搞定
谢谢 本人希望加入前30名的计算机学校读博士 |
|
a*****p
发帖数: 1285 | 12 发在jobhunting有人看没人回，还是跑这里问问。。。
从跟计算机完全不相干得专业转了计算机，一切从头开始，编程不熟练，想先找个lab
volunteer一下，熟悉下计算机环
境，有人帮忙上手快，也丰富下编程经验
1。想请教下，在学校找哪些方面得lab对以后找工作更有帮助？？本系得老师目前有以
下几个方向：algorithms,
automata reasoning, informatics, AI, graphics&人机互动，network, 还有系统和
安全方向得。。
还是找工程方面得lab？？
2。看了一个老师得东西，assistant professor, 是做多点触空得，multi-touch pad
or display, 然后还帮这边得医院做
做程序。。感觉挺好玩，这个多点触空屏是不是更接近硬件了？？以后工作这块多么？
？有没有前途？？还有，他主要用
python，我现在c++和java都不熟，这2个也是市场用得更多得，也是想练c++和java经
验，不知道怎么取舍。。。还是
找找其他得lab？？ |
|
d*****i
发帖数: 194 | 13 转专业申请cs master, 拿到这三个个学校的ad，其他学校还在等消息
Upeen MCIT：这个1.5年的项目只要求10门课，学费52k
utd：学费和生活费是便宜一点，大概3门课8k，可是要我修8门的先修课，这样下来学
费也接近50k了
concordia U, 申请这个是为了申请魁省移民方便，毕竟cs绿卡看现在的情况要等个８
－１０年，不过加拿大工作机会就没美国多了
我知道upenn牌子好，可是听说读cs master不在乎牌子，在乎地点，而且这个program
只修10门课，出去工作会不会被鄙视？俺没啥cs背景，像utd要求的先修课：CS2，
DISCRETE MATH， AUTOMATA THEORY， DATA STRUCTURES， SOFTWARE ENGINEERING
， OS CONCEPTS，COMPUTER NETWORKS，PROBABILITY AND STATISTICS，除了统计，都
没学过
请问各位：就是为了毕业后好找工作，选那个呢？听说德州cs的机会挺多的，宾州会不
会差点？
关于加拿大的学校，似乎是another story,但是板上各位工作... 阅读全帖 |
|
s*******n
发帖数: 24 | 14 必修课，但拿了个B
这个课关键吗？我学了，不知道有没有用 |
|
M****s
发帖数: 429 | 15 应用最多的领域是compiler
其他方面，比方说有些text search / indexing里面，也会用到。
这门课有的学校（比较少，而且一般是牛校）是必修的，有的不是。看你将来想干什么
了。 |
|
|
d*****u
发帖数: 17243 | 17 主要是编译原理要用
如果你做计算语言学（以及理论语言学）也要用 |
|
|
x***z
发帖数: 89 | 19 MS要修10门课，下面是领域分类及其相关课程（学校提供的），请大牛帮我选十门最要
紧最实用，最容易找工作的课，谢谢！
提供课程对应的编号即可(其中中文是我自己翻译上去的）：
领域一： (PFL) Programming and Formal Languages编程和形式语言:
1）ADV PROGRAMMING LANGS 高级编程语言； 2）COMPILER CONSTRUCTION编译器构造
；
3）SOFTWARE ENGINEERING软件工程 4）SOFTWARE MEASUREMENT 软件测试
5）OBJECT-ORIENTED PROG面向对象的编程 6）PROGRAMMING FOR GIS地理信息系
统编程
7）ONLINE DOC & HELP SYSTEMS 在线文档和帮助系统
8）XML FOR MULTILINGUAL(多语言的) APPS (=applications
9）3D GAME PROGRAMMING 10）SCIENTIFIC COMPUTING科学计算
11）FORMAL L... 阅读全帖 |
|
H**r
发帖数: 10015 | 20 树这个现在还是考的吧
不过学校里教的automata theory是扯淡了点
javascript |
|
r**t
发帖数: 937 | 21 多谢各位帮忙，看来能想到的办法还是通过automata。不过俺本来是想偷懒用java
现成的东西，现在看来是没有了。只好自己吭嗤吭嗤写了，sigh～～ |
|
m******t
发帖数: 2416 | 22
to
If just for the fun of it... 2 might be easy to achieve with
a negative lookbehind, but I doubt you can do 1 in a regex --
it sounds more complex than any finite automata can handle. |
|
l******0
发帖数: 244 | 23 还用到了 finite state automata, 这玩意涉及到固定的算法，及相应的数据结构 |
|
m********t
发帖数: 13072 | 24 是的，拼错了，很久不用了，谢谢
automata theory,这门课 |
|
T********i
发帖数: 2416 | 25 语言编程也是靠悟性的。即使人悟性好，聪明绝顶，也要站得高才能看得远。如果该知
道的不知道，叫一个诺贝尔奖得主来也白扯。
其实理论对我的帮助很大。formal language, automata, computability等等。感觉真
正领会深刻的也不多。否则有些东西早就应该有人做出来，也轮不到我来做了。
wdong做过的那些东西，其实很多人也做过。但是即使都做出来了，也不代表编程水平
都强。当然wdong 的能力很强，我很欣赏。我想说的是这里面有很多不容易量化的东西
。测刷题是无论如何都判断不了的。但是当面聊20分钟，这个人多少斤两，我就会很清
楚了。 |
|
d***a
发帖数: 13752 | 26 建立体系的人，远比“做出突破性贡献“的人少得多。图灵在计算机界的地位，类似于
牛顿在物理界的地位。（但公平地说，物理学科远比计算机学科要大得多也重要得多。）
图灵的关于停机问题的证明，并不是第一个关于不可判定性问题存在的证明。对砖家来
说，这是早知道的。王垠这家伙如果上课时听讲认真一点，就不应该现在感到吃惊，还
拿这事出来唠叨半天。王垠搞的是编程语言设计吧。这个方向比较soft，编程语言设计
的好坏，怎么辩论都可以，并不需要严实的数理逻辑基础。如果是搞formal language
and automata方向的，对图灵机的理解会深入得多。
王垠对图灵机的理解，还不如一个在计算机理论课上拿A的普通学生。他这家伙当年基
础打的不牢，现在自我感觉良好，才会象个太婆一样唠叨不休。:) |
|
S*******e
发帖数: 525 | 27 架不住人海战术。。。
https://news.slashdot.org/story/17/04/20/128224/95-engineers-in-india-unfit-
for-software-development-jobs-report
Talent shortage is acute in the IT and data science ecosystem in India with
a survey claiming that 95 percent of engineers in the country are not fit to
take up software development jobs. According to a study by employability
assessment company Aspiring Minds, only 4.77 percent candidates can write
the correct logic for a programme -- a minimum requirement for any
programm... 阅读全帖 |
|
发帖数: 1 | 28 很好奇 像automata, category theory, isomorphism inference theorem corollary
lemma equivalence type theory aleph naught dichotomy iamb isomorphism 这些
都怎么翻译 |
|
f*******t
发帖数: 7549 | 29 来自主题: Programming版 - 可微分编程 DL与FP的结合，会不会成为未来的编程方式呢？
Yann LeCun：深度学习已死，可微分编程万岁！
原创 2018-01-06 文强 新智元
【新智元导读】LeCun又发表惊人言论，继昨天参与深度学习论战、喷机器人Sophia后
，今天他在Facebook主页发文，称“深度学习已死，可微分编程万岁！”深度学习真的
死了？而可微分编程又是什么呢？
LeCun又语出惊人了，这次直指深度学习——
好，深度学习作为一个流行词，现在时效已过。
深度学习已死，可微分编程万岁！
事情要回溯到前天。
1月4日，AAAI前主席Thomas Dietterich连发10条Twitter，驳斥纽约大学心理学家Gary
Marcus对深度学习的批评。其中，Dietterich提到，
“深度学习本质上是一种新的编程方式——可微分编程——而且这个领域正试图用这种
方式来制定可重用的结构。目前我们已经有：卷积，池化，LSTM，GAN，VAE，memory单
元，routing单元，等等。”
这个说法让英伟达的AI架构VP Clement Farabet深表赞同，Farabet还评价说，这是对
现今深度学习的最好总... 阅读全帖 |
|
D****g
发帖数: 2860 | 30 I have heard of so called mathtype, I actually tried once. It is not even
close to latex. Of course, if you only OCCASIONALLY typeset very SIMPLE
formulas, fine, Word is good enough. But you need to typeset some good
looking formulas and diagrams, latex is absolutely necessary. You can take
out a math textbook, almost all formulas and equations can be typeset in
latex. Many many complex diagrams (e.g. an automata, control flow, data flow)
can also be easily set in latex. Also, latex has a very g |
|
|
x****6
发帖数: 4339 | 32 【 以下文字转载自 Military 讨论区 】
发信人: xiao86 (xiao86), 信区: Military
标 题: 生物学的问题是物理千老不给力
发信站: BBS 未名空间站 (Mon May 22 21:35:45 2017, 美东)
生命的本质是不断从环境吸收能量并向环境释放热量的非平衡态动态耗散系统
蛋白质的空间构型和动态性质的本质是多体问题
多细胞合作、鸟群鱼群行为的本质是cellular automata
尼玛，从微观到宏观，哪一个都是物理里还没有解决的难题！
生物千老凭什么被黑锅?! |
|
S****8
发帖数: 401 | 33 现在不少lab用Guassian啥的simulate一些Molecular，做Quantum cellular automata
这种东西，求问大拿这算promising的东西？还是坑？ |
|
o****i
发帖数: 23 | 34 如果是二年级：push down automata, turing machine, formal language, context-
free gramma
如果是三年级：computability, complexity (P, NP, NP-hard), efficient
algorithms (greedy, dynamic programming) |
|
k***g
发帖数: 7244 | 35 元胞自动机的英文 Cellular Automata ，离散空间时间里的动力系统，71年
Schelling 有一篇paper用它来研究 Dynamic models of segregation （Schelling 用
的是最简单的一维的模型），一些搞 complex system 的人研究这个，好像在金融上也
有应用吧，不过把它放在“十六讲”里估计确实很难讲出什么比较深刻的东西，恐怕就
是在 Excel 里用 VBA 写几个 loop ，然后 run 一下，获得一个 steady state，最后
一指黑板说：这就是元胞自动机了。 |
|
K**********n
发帖数: 1197 | 36 学术界珍稀有限的原版外文资料，
Remote Mind Control Technology
Reprinted from SECRET AND SUPPRESSED: BANNED IDEAS AND HIDDEN
HISTORY, edited by Jim Keith, $12.95, available from
1-800-680-INET.
There had been an ongoing controversy over health effects of electromagnetic
fields (EMF) for years (e.g., extremely low frequency radiation and the
Navy's Project Seafarer; emissions of high power lines and video display
terminals; radar and other military and industrial sources of radio
frequencies and micr... 阅读全帖 |
|
y****i
发帖数: 57 | 37 工科算法编程较多，想系统学点计算机课程。
觉得编程是工具，“善其事，利其器”。现在很多专业，都是编程实现。
觉得以前都是自己摸索编程，只要能运行就行。但是肯定不够。在运行的基础上，还要
编好。
选computer theory有用吗？
看了一下目录，觉得好像没有用。哪位科班的来讲讲。
Finite automata, regular expressions, grammars, languages, Turing machines,
computability, |
|
s****s
发帖数: 2163 | 38 This guy is definitely a cellular automata Extremist:
Galileo proclaimed that nature is written in the language of mathematics, but
Wolfram would argue that it is written in the language of programs and,
remarkably, simple ones at that. A scientific prodigy who earned a doctorate
from Caltech at age 20, Wolfram became a Nobel-caliber researcher in the
emerging field of complexity shortly thereafter only to abscond from academe
and establish his own software company (which published this book). |
|
v********e
发帖数: 1058 | 39 why automata and logics? there're plenty of active areas in theoretical CS,
and TCS per se is viewed as a branch of applied math. applied math and
statistics are not synonyms either. |
|
|
f*******d
发帖数: 339 | 41
when I was in PKU, once heard a seminar in the math department, someone
considered memory as information stored in a cellular automata in the form
of attractors or fixed points. |
|
f******e
发帖数: 3 | 42 Although he is a genius, his arrogance makes me a little uncomfortable. As to
what he talks in that big book, I think he sure has a strong argument.
But to what extent and broadness it will impact on modern science still
needs time to see. Cellular automata is an important part of the active
complex research, but not the only one. He obviously built the whole
"intellectual structure" on it without referencing sinlge
one literature in the relating field, which provoked lots of scientists
in the f |
|
c*****t
发帖数: 1879 | 43 A bit strange how your i is defined. What did you mean by finite?
I remembered that I posted a solution in this board for L != ww by giving a
formula. Couldn't find it anymore :(
It was something like:
L = AB | BA | A | B
A = aAa | aAb | bAa | bAb | a
B = aBb | aBb | bBa | bBb | b
The trick was that A and B had different centers.
I guess that for this problem, should be nearly the same.
Or, one could argue that given L != ww is in CFG, one could always
build a stack machi |
|
A***e
发帖数: 130 | 44 The main point is the non-deterministic: we can construct
the following PDA:
for each input symbol, depending on what it is and what's the current state
of the PDA, we can do the following actions (with the help of two flags:
A: having passed symbol c or not, and B: having passed the particular symbol for
comparison or not):
1) if it's not c, and the current state has neither flag A nor B set then:
(a) push it onto the stack
-or-
(b) store the symbol into the current state a |
|
A***e
发帖数: 130 | 45
hmm, by the meaning of "non-deterministic", each sequence of
choices of a/b in case1 (i.e., one path) will only compare *one* symbol in A
and B., and there are multiple paths. In the example you gave, 1100c1100,
there are 4 paths, 1xxxc1, #1xxc#1, ##0xc##0, and ###0c###0, all of them
do not accept. 'x' means taking no action at all, # means that the symbol
will be pushed onto the stack but what it actually is does not matter, only
the it's count matters.
So, there is no need to reverse, and act |
|
c*****t
发帖数: 1879 | 46 I think that I got what you were saying and it was indeed correct.
Apparently, I didn't notice that you pop all the symbols until the
very last token (there are some technicality there, but definitely
doable) then do the comparison. Basically one comparison for each
non-deterministic path.
I think that it will be very difficult if not impossible to come up
with a CFG for this grammar. |
|
b***e
发帖数: 1419 | 47
A -> 0 | 1 -- alphabet
B -> c | ABA -- balanced string
U -> B(A+) | (A+)B -- unbalanced string
X -> 0 | AXA
X1 -> Xc | A(X1)A
X2 -> cX | A(X2)A
Y -> 1 | AYA
Y1 -> Yc | A(Y1)A
Y2 -> cY | A(Y2)A
S -> U | X(Y2) | (Y1)X | (X1)Y | Y(X2) |
|
c*****a
发帖数: 808 | 48 Theory of Automata这课 ?
what is this
"
Regular languages
Context-free languages
Recursive and Recursively Enumerable Languages
Turing machine languages
"
是不是就是教 scripting languages Powershell, Bash or Perl？ |
|
