Keeping your spirit high!!

Speaking from experience after 2 years into PhD program, I can surely say graduate studies may be tiring sometimes. Along the way, there are many moments where you say – “that’s it”. Sticking to what you are doing is most important. Love your work. It does not need for one to be genius to complete graduate studies. Getting focused is the key idea. One should be proactive, keeping in touch with his/her adviser. I did do some blunders during start.

One of the way to keep your spirit high is to read other grad student’s experience. I do that all the time. At least, I know there are other people out there having same situation as mine. Let me share some of the links I feel are good.

http://heibeck.freeshell.org/Grad_Advice/how2b/how.2b.research.html

http://www.cs.unc.edu/~azuma/hitch4.html

Latex UI

I have used Latex from last 1 year. Initially, I used Lyx with MikTex 2.8. Then, I switched to Tex-Maker. For beginner, it is easy to use. Finally I installed Texnic Center, which I did not like much. Now, I am using WinEdt 6.0 and it has been fun.

One more info, this tool to convert pictures to eps looks cool. Try it. I love it!

Saurav

Self Organizing Maps

Unlike feed-forward and feedback network, which does supervised learning, Self organizing maps(SOM) have unsupervised and competitive learning strategy. Supervised learning algorithms such as back propagation algorithm have input vectors and target vectors. We adjust the weights of the neural network till the output vector is same or almost equivalent to the target vector. This means the learning process is supervised from outside. But, on self organizing maps, we employ unsupervised learning strategy. No target vector is required. It employs “winner-take-all” approach. First proposed by Prof. T. Kohonen, it is widely used in wide applications such as image processing, data clustering and visualization, medical image analaysis etc.

I have written a summary report about SOM. Anyone interested can read it from here. I had done coding for clustering of colors. If you need code for illustration of SOM, please let me know.

Saurav

Clone Objects, Design Patterns in C++

I was writing some code in C++ and I needed to clone the object. But, it seems there is no default clone methodologies in C++.  I found following links helpful.

http://www.cplusplus.com/forum/articles/18757/
http://www.agiledeveloper.com/articles/cloning072002.htm
http://bytes.com/topic/c/answers/159372-using-clone-function

Also, here is a link about design patterns in C++.
http://codewrangler.home.comcast.net/~codewrangler/tech_info/patterns_code.html

One more interesting paper
C++ Patterns : Executing around sequences.

- Saurav

Facebook Mail

According to *TechCrunch*, Facebook is going to announce new mail system on Monday, 11/15/2010. This comes amid fierce competition between Google and Facebook for hiring talents. Few years back, it was Google attracting tech guys from Microsoft, Yahoo and other giants. Now, it is facebook attracting guys from Google.

I feel the web has changed its nature. Earlier, it used to be all about some static websites/system where users were expected to consume information or at the best, user’s input would have been saved in some backend database. With the rise of social networking, the playing field has changed. Now, users are the ones who produce contents of the system. And, facebook has been the best to identifiy this potential at the early phase, and has maintained its foot.

Lets talk about main topic. Google has amassed huge user activities database. They reap using it with ads. The Google empire stands on this user data. Facebook, through its millions of users, has individual user database. While google always operated by analyzing from the huge data, facebook can use individual user data to direct its ads. Facebook is already the biggest billboard for ads. It has carefully placed its foot on search by the deal with Microsoft’s Bing. Now it is going to launch email. It has avid followers in tens of millions.
I think coming generation of applications will be mostly location based. Facebook is all set to go. Google has to urgently get some good social networking feature in its domain. In future, search results also need to have the social knowledge and location. Bing is already there(with some features).

*Note: This is my personal opinion.*

Summer 2010

New blog. I hope I will get time to update it frequently.