hdu 5734 Acperience
2016-07-22 16:06
288 查看
Acperience
Time Limit: 4000/2000 MS (Java/Others) Memory Limit: 65536/65536 K (Java/Others)Total Submission(s): 617 Accepted Submission(s): 329
[align=left]Problem Description[/align]
Deep neural networks (DNN) have shown significant improvements in several application domains including computer vision and speech recognition. In computer vision, a particular type of DNN, known as Convolutional Neural Networks (CNN), have demonstrated state-of-the-art results in object recognition and detection.
Convolutional neural networks show reliable results on object recognition and detection that are useful in real world applications. Concurrent to the recent progress in recognition, interesting advancements have been happening in virtual reality (VR by Oculus), augmented reality (AR by HoloLens), and smart wearable devices. Putting these two pieces together, we argue that it is the right time to equip smart portable devices with the power of state-of-the-art recognition systems. However, CNN-based recognition systems need large amounts of memory and computational power. While they perform well on expensive, GPU-based machines, they are often unsuitable for smaller devices like cell phones and embedded electronics.
In order to simplify the networks, Professor Zhang tries to introduce simple, efficient, and accurate approximations to CNNs by binarizing the weights. Professor Zhang needs your help.
More specifically, you are given a weighted vector W=(w1,w2,...,wn). Professor Zhang would like to find a binary vector B=(b1,b2,...,bn) (bi∈{+1,−1}) and a scaling factor α≥0 in such a manner that ∥W−αB∥2 is minimum.
Note that ∥⋅∥ denotes the Euclidean norm (i.e. ∥X∥=x21+⋯+x2n−−−−−−−−−−−√, where X=(x1,x2,...,xn)).
[align=left]Input[/align]
There are multiple test cases. The first line of input contains an integer T, indicating the number of test cases. For each test case:
The first line contains an integers n (1≤n≤100000) -- the length of the vector. The next line contains n integers: w1,w2,...,wn (−10000≤wi≤10000).
[align=left]Output[/align]
For each test case, output the minimum value of ∥W−αB∥2 as an irreducible fraction "p/q" where p, q are integers, q>0.
[align=left]Sample Input[/align]
3
4
1 2 3 4
4
2 2 2 2
5
5 6 2 3 4
[align=left]Sample Output[/align]
5/1
0/1
10/1
[align=left]Author[/align]
zimpha
[align=left]Source[/align]
2016 Multi-University Training Contest 2
Acperience
展开式子, \left| W-\alpha B \right|^2=\displaystyle\alpha^2\sum_{i=1}^{n}b_i^2-2\alpha\sum_{i=1}^n{w_ib_i}+\sum_{i=1}^{n}w_i^2∥W−αB∥2=α2i=1∑nbi2−2αi=1∑nwibi+i=1∑nwi2.由于b_i\in {+1,-1}bi∈{+1,−1}, 那么\displaystyle\sum_{i=1}^{n}b_i^2=ni=1∑nbi2=n, 显然c=\displaystyle\sum_{i=1}^{n}w_i^2c=i=1∑nwi2也是常数. 转化成求\displaystyle\alpha^2n-2\alpha\sum_{i=1}^n{w_ib_i}+cα2n−2αi=1∑nwibi+c的最小值. 对于固定的\alpha>0α>0, 只要\displaystyle\sum_{i=1}^n{w_ib_i}i=1∑nwibi最大就好了. 显然b_i=sign(w_i)bi=sign(wi)的时候, \displaystyle\sum_{i=1}^n{w_ib_i}=\sum_{i=1}^{n}|w_i|i=1∑nwibi=i=1∑n∣wi∣最大. 进一步的, 上面显然是一个关于\alphaα的二次方程, 于是当\alpha=\frac{1}{n}\displaystyle\sum_{i=1}^n{w_ib_i}=\frac{1}{n}\displaystyle\sum_{i=1}^{n}{|w_i|}α=n1i=1∑nwibi=n1i=1∑n∣wi∣时, 取到最大值.
化简下, 可以得到最小值是\sum_{i=1}^n{w_i^2}-\frac{1}{n}(\displaystyle\sum_{i=1}^{n}|w_i|)^2∑i=1nwi2−n1(i=1∑n∣wi∣)2
#include<iostream> #include<stdio.h> #include<math.h> #include<cmath> using namespace std; unsigned long long gcd(unsigned long long x,unsigned long long y) { unsigned long long a; if(x<y) a=x,x=y,y=a; while(y!=0){ a=x%y; x=y; y=a; } return x; } int main(){ unsigned long long sum1=0; unsigned long long sum2=0; int t; scanf("%d",&t); while(t--){ int n; sum1=0;sum2=0; scanf("%d",&n); for(int i=0;i<n;i++){ int tmp; scanf("%d",&tmp); sum1+=(tmp*tmp); sum2+=abs(tmp); } sum2*=sum2; sum1*=n; unsigned long long div=(gcd( sum1-sum2,n)); printf("%lld/%lld\n",(sum1-sum2)/div,n/div); } return 0; }
\left| W-\alpha B \right|^2=\displaystyle\alpha^2\sum_{i=1}^{n}b_i^2-2\alpha\sum_{i=1}^n{w_ib_i}+\sum_{i=1}^{n}w_i^2
相关文章推荐
- Acitivity 带返回值的 传递 示例
- 提高sql语句执行效率及索引
- java lock 的使用
- 自定义View学习笔记(1)->绘图技术详解
- 中缀表达式变成后缀表达式
- DOM中不存在的insertAfter()方法
- win7安装到gpt分区
- 旋转动画+悬浮按钮
- hadoop启动停止命令分析
- copy与mutableCopy区别
- Thinking in Java 第19章 枚举类型
- Android天气预报app改进版
- 自定义进度条HorizontalProgressBarWithProgress
- CRC32使用实例,字符串操作,SD卡读
- 状压DP—— NKOJ 1502盒子与玩具
- JS添加标签
- Spark的RDD编程(二)
- Zabbix添加Ping外网IP监控
- datatable-提示
- iSCS协议介绍