Emre Kaplan

update

Hi there,

I recently updated my wordpress, plugins and the themes installed. I hope you will like my new theme & photo.

cheers,

last 2

Hi there,

After a great vacation, i had problems with my life. I was alone for a while and had time to think about i’m doing. Everything seems to be good although i was alone. Beside, i was registered for the last two courses of my life. Last 2… This is important because after a month, the only two requirements to earn a PhD degree are, qualification exams (which i don’t believe in myself) and the thesis (which i can handle). Last 2… I feel like i know nothing… When i was a undergrad student, i thought i know enough. Coding was something beautiful and i was enjoying. What about today? I think i know just a little drop of an ocean and i hate coding. I have some ideas to turn into money-making projects but i don’t have energy to start and hate coding. So i have a deadlock situation. There are projects waiting to be coded and i’m looking for either passion to code or a coder to turn my ideas into running applications. I don’t know how to have passion and i don’t have money to support a coder to code. All in all, i’m totally deadlocked. My only hope is to have the funding i expect from the ministry.   What about academy? I’m done with the paper and it is quite ready to submit to a good journal. I hate to work on attack mechanisms therefore i may dive into privacy preserving techniques this time.

All other projects are in “pending” status and i hate to do something especially if it is a class project (such as i was assigned at the beginning of this term). I begin to learn coding in Android but i don’t think i will code a commercial application soon. It is for an academic project and as i stated many times, i hate coding and hate more if it is a class project.

Whoever reads this blog, please wish me luck.

on vacation

I think it is a good time to stop. I’ll go to Antalya and rest there. I hope this vacation enables me to re-energize myself and motivate for the long road ahead.

When i’m back, i’ll be writing about our new projects. I hope, we can do something that counts, at end of the day.

adios…

Plakamdanbul

www.plakamdanbul.com açıldı.

www.plakamdanbul.com is now available.

Thanks to jayjay21 and gordack for their contributions…

yet another attack on privacy preserving trajectory mining

hi folks,

We have been developing an attack algorithm to show that dissimilarity matrices is not a solution for data sharing when individual privacy is the concern. We already published a journal on this issue and this will hopefully be the second attack from another perspective. We are testing the attack algorithm with real and synthetic data sets right now and hope to start writing the paper in April.

So what is it about?
Companies track your position via GPS and cellular phones. When they publish it, they are not allowed to disclose your personal data, therefore they use privacy preserving methods before data release. We showed that current methods (dissimilarity matrix idea) is not a good privacy preserving technique because an attacker with a background knowledge can reveal individual trajectories in the data set.

route planning

Long time ago, i dreamed the cars will communicate and transfer traffic information to each other. Moreover, they can tell their current location and destination to the system and the optimal path will be returned from the server. Even if it is now our research project, it has more historical reasons why i stick to it instead of resting for a couple of days.

Today, GPS devices are so common, people can enter destination information and see the shortest path (in terms of kilometers) but is that enough? What if we can build a system which can collect the traffic information, cars’ location and their status and compute a shortest time path for the cars in the system by predicting the future traffic congestion models? This may, hopefully, solve the traffic problem in the big cities as well.

I’m currently coding the framework in Java, i hope to finish it soon and see some test results.

prime numbers

Hi friends,

I’m working on a  prime number generation and verification in parallel (of course on multi-core or multi-processor machines) project of various algorithms such as Miller Rabin, Solovay-Strassen, and Fermat’s primality test algorithms.

I’m currently evaluating their performance under real parallel computations (thanks to the multi-core technology). Results are not surprising that the speed up is proportional to the number of cores. As the number of cores doubles, speed also nearly doubles. This is true when we evaluate the prime number verification algorithms. That is, in the worst case, given a prime, concluding “prime” for a 512 bit number takes around 30 ms on .NET platform for single core, around 16ms for 2-cores and so on.

Surprisingly, when we talk about generating a random big prime number (which of course tries to verify if the randomly generated number sequence is a prime or not for each iteration until the generated number is prime) performance gain is not as expected. When we compare timings, there is a little or no difference between single core and multi core implementations. Moreover, when we discuss about 4096 bit numbers, 4-core machine makes it even worse and yields poorer performance than the 2-core machine. I guess you are asking “why?” Frankly, i also don’t know why but can guess it. Only 2 things i know up to now is that

1. the time to create worker threads, assigning some computation to them, scheduling, collecting the result back, etc definitely takes time which is around 15ms.
2. creating a random sequence is not as easy as assigning digits at random. It checks for basic divisibility rules before declaring a number as a prime candidate. This process may take long if it does not find a number which doesn’t satisfy the basic divisibility rules for small numbers in the range of [2, 10).

To sum up, prime number generation and verification is not an easy task for big numbers. Probabilistic algorithms do a great job when compared to the deterministic algorithms which takes around 6 minutes on a single core machine (compare 6 minutes to 30 ms).

on spaghetti code

Hi all,

This is the last day of 2009, actually 18 hours left for New Year. I was surfing on web and see a very nice quote about the endless discussion of the number of lines of code.

“If we wish to count lines of code, we should not regard them as lines produced but as lines spent.” ~Edsger Dijkstra

I wish a healthy, successful, and enjoyable New Year for all of us.

on Python

Hi,

My journey of learning Python started yesterday, now i can easily say that ‘ i can code in python’ thanks to the Python Programming.
This 150-page tutorial, as i mentioned yesterday discuss almost everything, not much in detail but give enough information to start coding python. I advise you to read the first 50 pages of it before diving into examples.
Good luck,


while 1 == 1:
print ("my infinite loop.")


What i’m doing right now?

Hi all,

I decided to play with Python for sometime since it seems to be good language. I can just play with console input outputs and some calculations at the moment but will read Python Programming book which can be found in here. Tutorial gives the basics in the first chapters and includes almost everything. Even if a subject is not covered in detail, there is at least a link or another tutorial is referenced.

I’m still on my Interpolation Attack project which i want tell about in the Projects section but i’m lazy to update these days. I hope i gonna complete my blog soon.