codesniper99's blog

By codesniper99, history, 16 months ago, In English

Hi, MikeMirzayanov

Why can't we just write a neural network which learns what obfuscation means and can read through all submissions after training the model, it will automatically flag ppl who are obfuscating?

Isn't trying to hide code or make it hard to read against policy?

The solution seems very simple to me, just make a NN. You already have so much data from past?

If you need more data,

make everyone who wants to participate in the next contest (and for each subsequent contest) , be allotted a random submission in history of codeforces. They have to judge if its obfuscation or not. Once its submitted then they can register. People with higher rating will have higher say in whether a submission is really obfuscated. Something like Amazon Mechanical Turk.

Use your loyal coders(however much remaining) to help us clean and purge this community!

(Reading through a code to understand obfuscation doesnt take more than 1-2 mins max)

Something like Overwatch in CS:GO or Dota 2.

Usually these obfuscations are something which is unpleasant to see like :

#define aCKRwrziTvHUKOb namespace
#define MPezggetwZGdgAP std;
#define ZccvuqPPyfgbiUB void
#define CrBidxvFCuXRFfi rec(int
#define zgSXTvGZhXIXbPI i,
#define lNrDOgrwmoZfdyR vector<vector<int>>&
#define OwALFRRpZsmIoMk arr,
#define PcYWVuVrBBAvHki vector<int>&
#define BiUUnCJMWNWndQS c,
#define eFEGSACCFdIjBDv dp,
#define ovMuMYANLYRNExM int
#define RKihtArgxWfmtHk n){
#define MTWPxisrdsAJYui if(dp[i]
#define IRkVmFbUteZgdXs !=
#define eZKyRSZhaiBkkpE 1e18)
#define JopfVyWEFQYPnTW return;
#define tQfHjycfjrvQDSF ans
#define xUIzdhodwOmvToe =
#define vPprlWYKQXdGFHJ c[i],
#define kgdiVvnFWODciNx sum
#define DSnARRCUUebCeAj 1e18;
#define dujRfmLoUhVpbQw for(auto
#define GBMAdcCOpAcJfVo val:arr[i]){
#define HeVKcslcUdfJfAD rec(val,arr,c,dp,n);
#define vjAxnOaoGuASLpz if(sum
#define SaXSjwuXLwcOzwh ==
#define yHlQJPkqqxGVElR =dp[val];
#define zuSudVNNnTeXLrb else
#define eNsAhgxrIasEuwT +=dp[val];
#define ThwvzPZPZZPpdld }
#define AqTIpcgDqQwpStw dp[i]
#define SCwqnTYHejUquwY min(ans,sum);
#define mdclyBiKxOAzScg signed
#define EErTqukgCvDxdBX main(){
#define eblPSYzowWROEWJ t=1;
#define KuXrrjhZXQgEjFe cin
#define AajvVdValWlkGzL >>
#define XGsRQWJdJiruOgE t;
#define PcpfoWUmFqnXNMU while(t--){
#define aMSXiKUQFESTnOo n,
#define UODNjQColDSzmZF k;
#define mJOpfluVcXHhFiA n
#define ZdaYMxfvfXISxUE vector<int>
#define eXuOZNCDQtERdfk c(n);
#define SWVdvoARmbsJOuT for(int
#define LxTydVtSGycalcP i=0;i<n;i++){
#define tYvJzxMTjlPTYVV c[i];
#define xoUPqhZeZXVnmgw p(k);
#define iBWkDhVOHchULLr i=0;i<k;i++){
#define ocVLKFsTwxyPSJG p[i];
#define nsldEPCPAEvxAKD c[p[i]-1]
#define qWVeWDPOPhnHhIx 0;

We can end this problem of obfuscation and cheating once and for all, What say?

Full text and comments »

  • Vote: I like it
  • +18
  • Vote: I do not like it

By codesniper99, history, 16 months ago, In English

Hi, TLDR: Can I solve all kinds of range query questions (the kind with N, Q<= 100,000) by just studying Segment trees?? (O(logn) query and update questions)

My knowledge till now tells me that Segment trees can be used for range query problems, I haven't found any other application yet (O(logn) query and update)

Context: I have some interviews lined up in coming days, I have studied the Segment tree from Codeforces guide and have got some good understanding for first time in my life how to implement and solve questions.

Now I see there are more advanced Data Structures/techniques (I have no idea of) like:

  1. Fenwick Tree
  2. BIT (Maybe this is same as above idk_)
  3. Order Statistic Tree
  4. Square root decomposition
  5. Tries

etc..

Now I need to work optimally and can't spend time learning more about Fenwick, BIT, and other DS too. I wanted to ask are there kinds of range query questions which can't be solved by segment trees? Will I be forced to learn about the other DS to solve some random question which may come my way (which is range based)

Thanks in advance, Me

Full text and comments »

  • Vote: I like it
  • -25
  • Vote: I do not like it

By codesniper99, history, 4 years ago, In English

for contests which are old and some even new contests, why dont the editorialists attach code to solve with the solution? Is it very hard to write code after solution or do they don't code the solution themselves. Dont tell me to look at standings for code. I want to ask why the author could'nt include his approach code with the solution. Very hard when reading the editorial need to check the code for verification but can't see it attatched. Then have to go through tens of submissions to find one which matches the editorial solution steps and makes sense and is legible.

Make it mandatory for all editorials if it isn't till now. ok bye.

Full text and comments »

  • Vote: I like it
  • -59
  • Vote: I do not like it