|
m****s
发帖数: 1945 | 2 被捕无疑, 检察官已经起诉。
Just googled "Criminal Charges: How Cases Get Started"
How does a criminal case get started? Usually with a police arrest report.
The prosecutor then decides what criminal charges to file, if any. The case
can then go to a grand jury for a criminal indictment or to a preliminary
hearing where a judge decides if there is enough evidence to proceed. Here's
how this all works.
Arrest Reports and Criminal Charges
After an arrest, the arrest report is sent to a prosecutor, whose job it... 阅读全帖 |
|
S****Y
发帖数: 4634 | 3 ☆─────────────────────────────────────☆
f20 (f20) 于 (Thu Mar 18 23:31:24 2010, 美东) 提到:
我觉得你可能不适合做research
实在不行就quit吧
☆─────────────────────────────────────☆
flyKiller (Nil) 于 (Thu Mar 18 23:45:15 2010, 美东) 提到:
你老板是不是那种很固执的或是年纪偏大的？或许你们性格不合？
有些教授就是这个样子，碰到问题总是觉得很简单然后脑袋一拍
一个主意。其实他们经常会犯错误的哪怕大牛也不例外。研究中
碰到的问题哪有那么简单的，有时候不仔细写下来很多地方还真
容易被疏忽。
感觉你们有交流问题，趁早换人吧。
☆─────────────────────────────────────☆
baybay2009 (bay) 于 (Thu Mar 18 23:47:21 2010, 美东) 提到:
第一年确实还没找到多少research的感觉。。。 沟通有问题，就更难进展了: |
|
|
o*****a
发帖数: 229 | 5 ☆─────────────────────────────────────☆
pascaldechin (小帕) 于 (Thu Mar 6 21:22:58 2014, 美东) 提到:
是一个同胞来interview，于是去听了，结果大跌眼镜。
我老听过这么多talk, 见过很多烂talk, 还从来没见过这么烂的，真是出离的愤怒啊。
phd毕业多年了，居然连怎么做talk都不知道。完全在泛泛地浅浅地survey his area，
而不是讲自己的工作。
尼玛你以为我们请你来是让你给我们科普的吗？
就算是科普，他这个科普水平也太差了。字体极小，图极小，slides感觉很cheap。英
语很差，专业基础知识也很差。
还很自以为是。
你说这种人怎么能混来interview? 我也奇怪，这厮简历还不错的。
现在看来都是他老板力挺他的，他自己学识实在有限，只能毕业了以后一直在老板手下
混。
我本来不理解为啥现在大家都兴phone interview先。
现在看来，确实有必要先刷一下，这一次on-site浪费了学校多少钱啊。
p.s. 如果用mac，请不要用ppt，要用keynote... 阅读全帖 |
|
m****1
发帖数: 736 | 6 http://www.nae.edu/Projects/MediaRoom/20095/130169/130172.aspx
New Members
Atwater Jr., Harry A., Howard Hughes Professor of Applied Physics and
Materials Science, California Institute of Technology, Pasadena. For
contributions to plasmonics.
Balakrishnan, Hari, professor of computer science, Massachusetts Institute
of Technology, Cambridge. For contributions to wired and wireless networks
and distributed systems.
Bardasz, Ewa A., president, ZUAL Associates Consulting, Mentor, Ohio. For
novel au... 阅读全帖 |
|
n**********o
发帖数: 1016 | 7 https://www.nae.edu/MediaRoom/178117.aspx
Abadi, Martín, principal scientist, Google Inc., Palo Alto, Calif. For
contributions to the formal theory of computer security.
Abrahamson, Norman A., chief engineering seismologist, Pacific Gas &
Electric Co., Piedmont, Calif. For contributions to seismic hazard
assessment and for leadership in engineering seismology and earthquake
engineering.
Adams, Robert W., senior fellow, Analog Devices Inc., Acton, Mass. For
contributions to digital storage and... 阅读全帖 |
|
j**l
发帖数: 2911 | 8 如何序列化一棵常规二叉树(不一定是BST)
方法一：
先序（中序或者后序）遍历这棵二叉树，对于空的左右孩子，也要输出空格或者某种特
殊字符作为delimiter，输出的序列就是序列化的结果
例如先序遍历输出
ABC@@DE@G@@F@@@
其中@表示空格字符
然后可以用递归的方式按先序（中序或者后序）序列建立二叉树（反序列化）
PROC crt_bt_pre(VAR bt:bitreptr)
read a char from the sequence, store it in ch;
IF ch = ' ' THEN bt: = NIL
ELSE [
new(bt);
bt^.data = ch;
crt_bt_pre(bt^.lchild);
crt_bt_pre(bt^.rchild)
]
ENDP
方法二：
利用如下原理
1) 已知一棵二叉树的先序序列和中序序列，可以唯一地重构这棵二叉树
2) 已知一棵二叉树的中序序列和后序序列，可以唯一的重构这棵二叉树
这样就可以输出先序和中序序列（或者中序和后序序列）作为序列化的结果
如果问起如何重构，比如知道先序序列和中序序列，则
从先序序 |
|
x******3
发帖数: 245 | 9 count = 1;
major = nil;
for a in A[1:n]
if (a != majorl)
count--;
if (count == 0)
major = a;
count++;
return major |
|
n******n
发帖数: 12088 | 10 如果把9去掉，7的right是11还是nil？ |
|
i**********e
发帖数: 1145 | 11 这样的话是nil。
原题是你可以assume是full binary tree。 |
|
V*****i
发帖数: 92 | 12 在merge sort中，需要设一个sentinel，大家都是用什么的？在现实编程中，你一般已
经知道了值的范围，所以随便设个很大的数就可以了，但总觉得这在面试中不够严谨，
大家是怎么办的？我试过numerical_limit<>::max()，但发觉很慢。
类似的是，在hash table open addressing算法中，需要设nil和delete，大家是设成
什么的？ |
|
g*****e
发帖数: 282 | 13 这个题谁讲讲？我想的，有个ptr保持list header，然后随机走若干步（如果到nil了
重新从头
走）。
第四个，A list (LinkedList) of n numbers, p1->p2->...->pn. Find k random
numbers.
1) n is known
2) n is unknown
Onsite |
|
l****p
发帖数: 397 | 14 我觉得这题主要是考怎么构建树，至于树建起来了，怎么展现比较漂亮那是另一回事了
。这年头很少人会考用ASCII字符在控制台上打图形吧。顺便贴上我的实现：
class TreeNode
attr_accessor :value
attr_accessor :left
attr_accessor :right
def initialize value
@value = value
end
end
def get_trees preorder, head=0, tail=preorder.size
return [nil] if tail <= head
trees = []
for right_root in (head+1..tail)
left_trees = get_trees preorder, head+1, right_root
right_trees = get_trees preorder, right_root, tail
left_trees.each do |lt|
right_tree... 阅读全帖 |
|
t****a
发帖数: 1212 | 15 谢谢提示。据此写clojure code如下，调试成功
(use 'incanter.core)
(defn subseqzero? [a]
(let [ac (cumulative-sum (cons 0 a))
acf (frequencies ac)]
(some #(> % 1) (vals acf))))
(def a [1 -3 4 1 -2 9])
(subseqzero? a) ; true
(def a [1 -3 4 3 -2 9])
(subseqzero? a) ; nil |
|
p*****2
发帖数: 21240 | 16 第二题练了练
def divide(a,b)
if b==0
return
end
print("#{a/b}.")
a=(a%b)*10
arr=[]
hm={}
while hm[a]==nil
hm[a]=arr.length
if(a!=0 && a
a*=10;
arr.push(0)
else
arr.push(a/b)
a=(a%b)*10
end
end
for i in 0..hm[a]-1
print(arr[i])
end
print("(")
for i in hm[a]..arr.length-1
print(arr[i])
end
print(")\n")
end |
|
t****a
发帖数: 1212 | 17 第三题很有趣，偶来给个clojure的解
基本思路是
1. 对T做预处理，构建graph，同时，记录每个字符a-z所出现的位置
2. 由于S长度<100，数量却很大(100M)，意味着有很多重复，使用memoize function会
大大减少计算量
下面的程序在1/1000大小的S数据集上运行需要1.7秒。观察了一下时间和数据集大小趋
势，在完整数据集上几分钟内也可以产生结果
(def alphabet (map char (range (int \a) (inc (int \z)))))
(def T (vec (take 10000000 (repeatedly #(rand-nth alphabet)))))
(def S (take 100000 (repeatedly (fn [] (take (inc (int (* (rand) 100))) (
repeatedly #(rand-nth alphabet)))))))
(def T-ix (group-by #(T %) (range (count T))))
(def T-map (let [pairs (dist... 阅读全帖 |
|
t****a
发帖数: 1212 | 18 第三题很有趣，偶来给个clojure的解
基本思路是
1. 对T做预处理，构建graph，同时，记录每个字符a-z所出现的位置
2. 由于S长度<100，数量却很大(100M)，意味着有很多重复，使用memoize function会
大大减少计算量
下面的程序在1/1000大小的S数据集上运行需要1.7秒。观察了一下时间和数据集大小趋
势，在完整数据集上几分钟内也可以产生结果
(def alphabet (map char (range (int \a) (inc (int \z)))))
(def T (vec (take 10000000 (repeatedly #(rand-nth alphabet)))))
(def S (take 100000 (repeatedly (fn [] (take (inc (int (* (rand) 100))) (
repeatedly #(rand-nth alphabet)))))))
(def T-ix (group-by #(T %) (range (count T))))
(def T-map (let [pairs (dist... 阅读全帖 |
|
t****a
发帖数: 1212 | 19 可以用DP来做。计算复杂度大约为n^3*len(dict)
假设输入字符串s长度为n，f为求解的函数，那么
f(s, dict) = match_dict(s, dict) 当s可以直接匹配上dict里的某字符串（通过排序
后比较）, match_dict返回dict中匹配上的字符串
f(s, dict) = (max the words length) [f(subs(s, 0, i), dict), f(subs(s, i, n)
, dict)], i=1..n-1 当s不能直接找到dict里的匹配
f(s, dict) = nil 当s为空
clojure程序如下
(defn gen-sorted-dict [dict]
(reduce merge (for [w dict]
{(sort w) w})))
(defn match-dict [s sorted-dict]
(sorted-dict (sort s)))
(def prob10-sub
(memoize
(fn [s sorted-dict]
(wh... 阅读全帖 |
|
p*****2
发帖数: 21240 | 20 (defn f [l]
(defn- dfs [xs pre next layer]
(cond
(<= layer 0) nil
(empty? xs)
(let [n (cons (inc pre) next)]
(println n)
(dfs n 1 () (dec layer)))
:default
(let [curr (first xs)]
(dfs (rest xs) curr (cons (+ pre curr) next) layer))))
(dfs () 0 () l)) |
|
s*****r
发帖数: 108 | 21 写麻烦了，只要扫两遍就好，O(n log n) + O(n) + O(n)
空间上也只需要一个输出数组 O(n)
不过没有仔细想是否正确
sort(points, cmp)
maxn, count = 0, 0
for p in points
(p.start?) ? (++count) : (--count)
maxn = max(maxn, count)
ret = []
count = 0
last_start = nil
for p in points
if p.start?
++count
last_start = p
else
if count == maxn
ret << [last_start.val, p.val]
--count
ret |
|
p*****2
发帖数: 21240 | 22 顿时感觉自己好没用。还要向800题大牛的方向努力呀。解法如下
#(clojure.pprint/cl-format nil "~@R" %) |
|
p*****2
发帖数: 21240 | 23 (defn isBST
([root]
(isBST root, Integer/MIN_VALUE, Integer/MAX_VALUE))
([root min max]
(if (nil? root) true
(let [val (:val root)]
(and
(>= val min)
(<= val max)
(isBST (:left root) min val)
(isBST (:right root) val max))))) |
|
e********r
发帖数: 24 | 24 这个可能犯规了……
import scala.collection.mutable.ArrayBuffer
def deepFlatten(list: Seq[Any]):ArrayBuffer[Int] = list match {
case Nil => ArrayBuffer()
case elem::rest => elem match {
case v: Int => v +=: deepFlatten(rest)
case l: Seq[Any] => deepFlatten(l) ++=: deepFlatten(rest)
}
}
class DeepIterator(list: Seq[Any]) {
val content = deepFlatten(list)
def hasNext = !content.isEmpty
def next = {
val item = content.head
content -= item
item
}
} |
|
e********r
发帖数: 24 | 25 这个可能犯规了……
import scala.collection.mutable.ArrayBuffer
def deepFlatten(list: Seq[Any]):ArrayBuffer[Int] = list match {
case Nil => ArrayBuffer()
case elem::rest => elem match {
case v: Int => v +=: deepFlatten(rest)
case l: Seq[Any] => deepFlatten(l) ++=: deepFlatten(rest)
}
}
class DeepIterator(list: Seq[Any]) {
val content = deepFlatten(list)
def hasNext = !content.isEmpty
def next = {
val item = content.head
content -= item
item
}
} |
|
b******p
发帖数: 49 | 26 我来贴个CPP的（注意：以下有乱七八糟的code……）
合并排序确实比较好用，我还在写第一遍leetcode，代码风格也比较乱，带了很多
debug code，还带了测试用例，试了9次才通过…
有一些多边界条件需要判断的。我的方法是加很多debug，或加很多assertion（我做别
的题里面经常用assertion……不知道是不是好习惯）
＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝
#include
#include
using namespace std;
/*
Merge sort ?
*/
// Start: 22:40
// End: 23:45 用了一个小时
//Definition for singly-linked list.
struct ListNode {
int val;
ListNode *next;
ListNode(int x) : val(x), next(NULL) {}
};
#define TOMMYDBG
class Solution {
public:
... 阅读全帖 |
|
c********l
发帖数: 8138 | 27 //MITBBS' article system will mess up the source code
// You may probably need to re-arrange the line breaks
import java.util.ArrayList;
import java.util.HashMap;
import java.util.HashSet;
import java.util.Map;
import java.util.Set;
public class Solution {
public static void main(String[] args) {
Solution sol = new Solution();
String[] dictStrArr = new String[] {
"dose","ends","dine","jars","prow","soap","guns","hops","
cray","hove","ella","hour","lens","j... 阅读全帖 |
|
c********l
发帖数: 8138 | 28 简洁不是最终极的目标，有时候需要reduce runtime complexity
见我之前写的
http://www.mitbbs.com/article/JobHunting/32658475_3.html
//MITBBS' article system will mess up the source code
// You may probably need to re-arrange the line breaks
// by coupondeal@mitbbs
import java.util.ArrayList;
import java.util.HashMap;
import java.util.HashSet;
import java.util.Map;
import java.util.Set;
public class Solution {
public static void main(String[] args) {
Solution sol = new Solution();
String[] dictStrA... 阅读全帖 |
|
s******y
发帖数: 33 | 29 As a programmer, the probability is almost nil if you want to work in the
oil company. Big oil companies hire undergrad cs student with citizenship,
think twice before you make the decision. cgg is an above-avg company in
houston. you will reach 100k income in 2 year. but the company culture is
terrible. The house price is around $110-120/sqrt in Katy/Sugarland. |
|
g********4
发帖数: 477 | 30 As a programmer, the probability is almost nil if you want to work in the
oil company.
能解释一下不？ |
|
w****6
发帖数: 796 | 31 烙印真能忽悠啊 - 下面是一个烙印的自辩，好像没有老美能
驳斥他：
hastaprontosj 1 day ago 1 3
For all those abusive commentors -
We have every kind off people in every country. If you may have encountered
few fraudster H1Bs,it does not mean every one is a fraud.Also H1B is a visa
giving work authorization to every other foreign national (non US nationals)
and not for just Indians. It has a quota limit of 60K per annum and is
shared by all non US nationals. Please DO NOT generalize people.
1) I would like to point one imp thing... 阅读全帖 |
|
w**a
发帖数: 487 | 32 这是一个店面后的assignment题，要求用java写，两天完成吧，我用了两天写出了第一
个版本，能work，但是code 很难看，没办法，只好那样交上去了。大侠们要是有时间
，可以写写么，让我学习一下？
======================
题目：
有一个小城市的所有饭店的菜单数据（见后），格式是CSV格式，要求写一个JAVA程序
，帮助客人找到最好的deal。
客人可以选一种或多种菜品，如果某个菜品是从某个combo里面来的，那么购买那个菜品
需要付整个combo的价格。
例子1：
input： burger tofu_log
expected: restaurantID = 2, deal price = 11.5
例子2：
input: chef_salad wine_spritzer
expected: nil(or any message notification)
***因为这两个东西不在一个店里，所以没有deal
例子3：
input: fancy_european_water extreme_fajita
expected: restaurantID = 6... 阅读全帖 |
|
p*****2
发帖数: 21240 | 33 def isUTF(xs: List[String]):Boolean = {
xs match {
case Nil => true
case head::tail => head.toList match {
case '0'::_ => isUTF(xs.tail)
case '1'::'0'::_ => false
case _ =>
val len = (x: String) => x.prefixLength(_ == '1')
val l = len(head)
val rest = tail.take(l-1)
rest.size == l-1 && rest.forall(len(_)==1) && isUTF(xs.drop(l))
}
}
} |
|
H****S
发帖数: 1359 | 34 kinda remember someone talked about this in his blog. Here is the idea,
import java.util.concurrent._
import scala.collection.mutable
val hq = mutable.Buffer[Condition]
val oq = mutable.Buffer[Condition]
val lock = new ReentrantLock()
def h() {
lock.lock()
try {
if (hq.size >= 1 && oq.size >= 1) {
val ch = hq.last
ch.signal()
val co = oq.last
oh.signal()
hq.trimEnd(1)
oq.trimEnd(1)
println("h")
} else {
val ch = lock.newCond... 阅读全帖 |
|
x****g
发帖数: 39 | 35 import string
class Solution(object):
def ladderLength(self, beginWord, endWord, wordDict):
"""
:type beginWord: str
:type endWord: str
:type wordDict: Set[str]
:rtype: int
"""
transformation = [beginWord]
wordDict.remove(beginWord)
length = 1
while transformation:
for i in range(len(transformation)):
fromWord = transformation.pop(0)
for i in range(len(fromWord)... 阅读全帖 |
|
b******y
发帖数: 9224 | 36 有些公司的有些工作，进去可能是个坑。虽然拿了大包裹，但经济不景气的时候被裁了
，收入一下子降到0，zero, nil.
所以，要自己分析一下利弊，不能简单的被大包裹所迷惑。
版上有些人说的也有道理，美国经济多少年就是一个周期，现在有些下滑的趋势。有人
可能会觉得IT产业是是新兴产业，哪儿都用得到的。
是，哪儿都用得到的，但好多同学没有经历过2000年dot com crash的时候。当时好多
人都不相信曾经红极一时的IT产业，一下子彻底萧条了，好多公司一夜之间就vanish了
。。。 |
|
|
e***a
发帖数: 1982 | 38 一个没有。。。NIL。。。。
希望谁能找到sync的方法。。。 |
|
d******i
发帖数: 448 | 39 对，the wonderful adventures of Nils。小时候看的好像是日本改编的动画片，很多
集的。 |
|
J*******p
发帖数: 1129 | 40 给大男孩的好书（不是教材）.有些可能是6年纪以上阅读程度的。
Hatchet by Gary Paulsen
A Separate Peace by John Knowles
The Jungle Book by Rudyard Kipling
The Phantom Tollbooth by Norton Juster
The Adventures of Huckleberry Finn by Mark Twain
The Adventures of Tom Sawyer by Mark Twain
The Last Mission by Harry Mazer
Red Badge of Courage by Stephen Crane
Watership Down by Richard Adams
The Chronicles of Narnia by C.S Lewis
The Giver by Lois Lowry
The Lord of the Flies by William Golding
Treasure Island by Robert Louis Stevenson... 阅读全帖 |
|
G**T
发帖数: 2610 | 41 欢迎继续刺激 －－ 我还是认为你其实智商比你以为的更差了。 :)
瞧不上你同学？ －－ God bless你不要知道你同学怎么想你的 :)
我跟你啥关系？ nothing。 nil。 zero。
你同学跟你啥关系？ 你要不被刺激了，会到这里来感慨（然后挨一闷棍)? |
|
b**a
发帖数: 1375 | 42 【 以下文字转载自 PDA 讨论区 】
发信人: boya (跛鸭/.1/Nil/Nirvana), 信区: PDA
标 题: 请教怎么买BB Storm 2 便宜啊
发信站: BBS 未名空间站 (Sat Nov 14 02:50:22 2009, 美东)
要给国内哥们带个过去. no contract的, 最好是verizon定制的.
我查了最低的是$539, 还有便宜的么.... 多谢多谢. |
|
|
z***e
发帖数: 5600 | 44 Each country has its own policy
The talk of these days is that Italy may puts a 8GW cap on total
solar installation, and existing installation are approaching it quickly
That means after the cap reached, demand for new solar installations
will be essentially nil.
The article below tells what happened in Spain, which provides a good
example
http://www.bloomberg.com/news/2010-10-18/spanish-solar-projects
subsidy-policies-founder.html |
|
r***l
发帖数: 9084 | 45 没人回，贴过来了。。。自己看吧，呵呵
Net revenues. Our net revenues for the three months ended March 31, 2011
were US$20.6 million, a 46.6% increase from US$14.0 million for the three
months ended March 31, 2010 and a 1.7% decrease from US$20.9 million for the
three months ended December 31, 2010. The change in net revenues as
compared to the three months ended March 31, 2010 was due to the change in
both our online advertising revenues and IVAS revenues.
Online advertising. Our online advertising revenues i... 阅读全帖 |
|
h********r
发帖数: 928 | 46 By IB Times Staff Reporter | Aug 16, 2011 3:54 PM EDT
(Photo: Reuters / Beck Diefenbach)
Apple Inc CEO Steve Jobs takes the
stage to discuss the iCloud service at the Apple Worldwide Developers
Conference in San Francisco June 6, 2011.
Enlarge
(Photo: Reuters / Beck Diefenbach)
Apple Inc CEO Steve Jobs takes the stage to discuss the iCloud service at
the Apple Worldwide Developers Conference in San Francisco June 6, 2011.
(Related Articles
Google deal shows cash-rich cos may not need to t... 阅读全帖 |
|
c**y
发帖数: 419 | 47 不是这么好测试的, 比如你需要一些过滤条件, 比如价格从止损价格线下面翻上来,要
连续2-3天站上第一天翻上来时候的高度. 还有就是w底的时候, 这第一个翻上来的尖头
是不能做多的. 这些用肉眼看是容易的, 计算机编码就不容易了.
目前2%对大盘指数是很有效,尤其是美国的3个指数, 至少2010年来是这样, 2000年的时
候就要4%左右了, 说明不同时期,市场的波动是不同的; 个股的话, 一般是3-6%,取整数
, 因为william O'nil写的那本书how to make money in stock market里面推荐止损是
7-8%,所以这是个上限.
上证指数用2或者3%,注意上面那些过滤条件,也是极其有效的.
1小时的数据你要缴费下载了. 止损率不是老换的,这样的模型才是robust稳健的. |
|
r******9
发帖数: 2632 | 48 Please forgive my poor Canadian English. You just opined on the valuation of
Weibo. All looks good, very very good, and very very good, and indeedy.
However, in my very very humble opinion, Sina.com, as a portal, should be
valued at 0, if not negative. Yes, that's right, zero, nada, zilch, zippo,
and for CS majors, nil.
com
billion. |
|
m******1
发帖数: 86 | 49 Don't quite understand why you say Sina.com should be valued as Nil. it
generate $140 mil operating cash flow annual.
If you are not happy with Sina's management, then it's a totally different
story. |
|
t**x
发帖数: 20965 | 50 so, rhere will be a qe3
Street
Wall |
|
