보스톤 10K 마라톤

일단 등록은 했는데.. 연습 언제 하지 ㅡ.ㅡ

득템

연구실 지나가다 얻은 공짜 옷

오랜만에 김밥

2011 Pan-STARRS Consortium, Boston, MA

Where is Anna Sofia

Super smart Google

How the hell does Google know that I'm the lazy guy???

P.S. 김대원 shown in the left textbox is my name in Korean.

Churches and Cathedrals in Seoul and Boston (서울과 보스톤의 교회 및 성당 분포)

I just saw one article showing the distribution of churches around area, so I test it by myself using the google map.

Let's first check around the Boston area, where I'm now.

Cathedrals

Churches. 

The number of Churches and Cathedral looks not very different.

Now, let's look at Seoul, South Korea.

Cathedrals.

Churches. WTF.

Now I can understand why we can see, after sunset, only the churches' freaking bright electronic crosses on the top of each church in the every places of Seoul.

Churches in Korea are not anymore sacred places for prayer and rest, it is just an another business model to delude people so that to legally blackmail them. I'm really sorry but it's the truth of the most of churches in Korea..

* All maps are roughly in the same scale.

P.S. Don't get me wrong. I really love to read bible and do appreciate what Jesus have done during his mortal life, just as much as I love Buddhism.

Our work on QSO (Quasar, AGN) detection using machine learning and time varibility

is introduced in the astrobite website:

http://astrobites.com/2011/01/19/computers-search-for-quasars/

http://astrodatamining.net/ also kindly asked me a short summary of our work to post it on their website.

(http://astrodatamining.net/2011/05/01/papers-qso-selection-algorithm-using-time-variability-and-machine-learning-selection-of-1620-qso-candidates-from-macho-lmc-database/)

Feels not bad that at least couple of peoples are reading our paper :)

Gaussian Mixture Model - Python

Using scikits.learn.mixture

The example is from:

http://scikit-learn.sourceforge.net/auto_examples/mixture/plot_gmm_pdf.html

Traveling salesman problem - Simulated Annealing

The optimized solution after 'lots' of iteration.

Temperature, T, was slowly decreased each iteration.

The red star is the start cities.

The blue cities are where airplanes which are faster than cars are available.

100 cities in total.

In this simulation, we assumed traveling by just a car.

Hm.. looks like not the best solution.

need to contraint T in other way?

Energy of system as a function of iteration.

E is the sum of distances between cities.

For Applied Mathematics 205b class, Harvard.

#Updated : Rather than swapping cities, reversing the traveling direction works much better. Below is the final solution using the reversing strategy.

The cIrcular cities. We all know what is the best configuration, and SA exactly found it.

One more example with 1,000 cities.

left : the initial configuration, right : the solution.

looks pretty good.