How is life going on?

Hi every body!
Long time no see!

I must admit, that it was me, who was lazy, 'coz the others have been updating even during the semester. In my case, I was really in lack of time, even for common life problems, such as shopping. (Don't worry my uni, RWTH Aachen, is well known for "killing our holidays" )There was only one friend, who was hearing from me incessantly, which was also a kind of compensation against everyday life problems.

Did you know that I can cook?!?!

What was going on during last 5 months:
I have the feeling, that it's becoming my habit, only updating during my exam-time, which is right now. Last week I wrote "successfully" my first exam, and now I am preparing my self for the next ones, which I'll write in the next weeks. Accordingly, I was feeling: "okay, it's time to write some thing!" and I am doing like this.

My second semester was really full. Without wanting to exaggerate, I learned how it is possible to work 16 hours a day, even at weekends. I wrote a 15 page paper for my seminar, which is in September and I'll participate in a research project for 6 weeks. The project is about P2P, more on that later.


Artificial Intelligent:
My seminar is about reinforcement learning in artificial intelligent. I taught to write a bit about it:

Imagine there is a robot, beside the hardware it also needs a software, which is written by a programmer. The software is responsible for how the robot will "behave". To make it easier, let us imagine the robot already disposes some important functions, such as "watching" the world or taking actions, what ever these actions are.

One of the main points is, that the programmer can never tell the robot what to do in all possible situations, hence the robot has to learn. There are different types of learning: some time there is a teacher there, which can tell the robot which actions are good and which not.

Imagine a robot playing football. It can run and shoot, it knows also the rules of the game, but that's not enough for playing football. So the robot must learn how to play. After each action the robot can be told, whether its action was correct or not, this kind of learning is called super-wised, but what if there is no teacher who tells the robot, "nice move" or "it was not fast enough"?

Here helps reinforcement learning, in this kind of learning, the robot must make a strategy out of his experience, which is called reinforcement or reward. For example, our robot now plays ping pong: After making one point, the robot will try to remember how he played and so make a good strategy of making profitable actions.

I was dealing with this subject for the last 3 months, and my first paper will be published in one month, maybe I'll put it here, so you can take a look at it. But it's more abstract than I wrote here and there are also some algorithm.


Well I hope to find more time to write here.

*************

سلام

خیلی وقت بود از من خبری نبود. که علتش بر می گرده به ترم ‍پر مشغله من. واقعا جای سر خاروندن هم نبود. صبح تا شب کلاس. شب تا صبح انجام تمرینات و اماده کردن یه سری پروژه عملی.

تو این ترم کلی یاد گرفتم که چی جوری با وجود کمبود خواب تمرینات رو با کیفیت انجام بدم. در حال حاضر هم در دوران امتحانات بسر می برم که اگه همه چیز خوب پیش بره 2 هفته دیگه قسمت اولش به پایان میرسد.

یک ماه دیگه یک سمینار می دم درباره یکی از روش های یاد گیری در هوش مصنوعی که در بالا مختصری ازش رو نوشتم.


مهمتر از همه این است که من در طول این مدت کلی آشپزی یاد گرفتم و الان داره از هر انگشتم یه هنر میباره. در عکس بالا دستپخت امروز من رو مشاهده می کنین. فکر نکنین خیلی ساده است برای خودش یکی از پیچیده ترین هنر های دنیاست. چون حتا سرخ کردن سوسیس هم روش داره...


راستی تو این چند وقته با تمام کمبود وقتی که بود یک دوست باعث شد چند بیتی هم شعر بخونم:

.............
چراغي به دست‌ام چراغي در برابرم.
من به جنگ ِ سياهي مي‌روم.

و خورشيدي از اعماق
کهکشان‌هاي ِ خاکسترشده را روشن مي‌کند

..............