One year far away from home

Today is exactly one year that I am in the Canada. Last year in the 28 June around 11 p.m. I arrived into YEG airport and the officer stamped my passport. For he, it was like other hundred stamps he make every, but for me it was something special, it was sing of my new life. I am not going to talk about the whole year here, as it is almost impossible, but I would like to present an overall conclusion about my life. Maybe you tell me “It is not none of my business”. Yes, you are right. I think I am doing this because I want to convince myself to be a better student in the next year . In the following I will say my opinion about different aspects of my life:

• Emotional: From this point of view, I can say I had a terrible year. I had lots of lonely and sad moments, especially in the first 4 months. During that time I was thinking about my family and friends almost all nights. I was always dreaming about visiting home and planning what should I do, and I finally visited home one month in the May. Besides that I did not have really good moments here, although I have very good friends here. I guess the main lack point of my life was a girl friend .
• Life style: I am quite satisfied with my life in the previous year. I started going to Gym again, I did not smoke a lot, and also I did not drink, I can say “at all”, 6-7 times drinking bear in whole year. I would say I had a healthy year.
• Financial: As you know students in all over the world have the worst financial situation, a little bit better than homeless people. Yet I save enough money to have a trip back to my home town (which was pretty expensive, something around 3500 CAD) . Moreover I save a little money which is good in comparison of my savings during the three years working. Thanks to my “cooking in the home” method which also caused me to have a healthy diet.
• Language: At the beginning I thought I would learn so much English in JUST one month that I could speak like a native person! That was a silly expectation which was completely wrong. What I think now, which also may become a crazy idea sometime later, is that things which I have learnt in one year are more than other people’s have learnt in the same time. Thanks to Jake and Lory who have helped me a lot by their “English conversation program”, the elllo with its great interviews, the books which ARM bought for me (Understanding and using English grammar and Longman dictionary of contemporary English), The Simpsons, and finally this blogg, hip hip, hooray.
• Academic: You know why I come to this at the end? Yes, because I am not satisfied with my outcomes at all. When I belive I did not learn as much as things as I expected this year, moreover I had not a good GPA. A good GPA could at least convinced me that I was not such bad, but unfortunately I don not have. My first year transcripts is shown in the following image. However considering my adaptation to the new environment and also choosing a good supervisor (Russ) I did a good job.
  

  Transcript

DSLRs vs. Compact Digital Cameras

Last week a freind of mine lent me his DSLR camera, it was a Olympus E-510 with Olympus Zuiko 12-60 mm Lens. Actually I did not test it seroiusly, I just tried to find out what is different between a DSLR and a compact digital camera. In this post I am going to talk about the main differences which I found. Hopefully this post will help beginner people like me who are wondering what kind of camera they should buy. First of all I recommend you to tke a look at this glossary and become a little bit familiar with the expressions which “professionals” use. Here are main differences:

1- DSLRs (Digital Single-len reflex camera) have a mirror which control where light should go, into the viewfinder or on the sensor(film in the older cameras). So unlike the compact digital cameras which light allways exposes (shines) on the sensor, in DSLRs light exposes the sensor just one moment, when you press the shutter button (consiquently in that moment veiwfinder will be dark). On the other hand compact digital camera’s sensor is always subject to light so the LCD on the back shows you what is going to be saved by camera. Following pictures will help you to understand it better:

This feature will result in two things:

• Professional phogrophers uses their viewfinder to take pictures (because they have to, thers is nothing on their camera’s LCD)
• DSLRs can not record movie clips (because the sensor is not exposed constantly)
• When you press the shutter button you can hear the famous taking-picture sound (bacause the mirror should move twice). The sound is not made by camera’s speaker!

Although some of DSLRs have a feature called “live preview” which shows current frame on the LCD on the back of camera (like compact ones). In this mode mirror is move away such that light goes stright and exposes the sensor and viewfinder is dark.

2- Sensor Size: Size of DSLRs’s sensors is much bigger than compact cameras, in fact the main reason of DSLR’s high price is their big sensors. A very common way to measure and compare sensor sizes is to compare them with size of 35mm films which were used in old cameras. DSLRs which have a sensor exactly equal to a 35mm film are called “full sized” and are the most expensive ones. This picture shows different sensor sizes and compare them with 35mm film:

Having bigger sensor sizes will bring you lots of good things (except the money you should pay )

• Less noise: DSLRs camera can take pictures even in the dark places, but almost all compact cameras can not. You may have heard something like this: “This camera works well outside and in the day light, but not in the night or inside home”. To explain why you should learn about ISO. When light is not enough, the camera’s sensor should be more sensitive to light (uses higher ISOs), so by using a little amount of light it will capture a good picture. Now lets compare a 12mega pixel (3000×4000) “full frame” DSLR and a 12 MP compact digital camera with sensor size of 1/1.7″ (above image). Area of each DSLR’s pixel will be 864/12^9=0.000000168 mm2 and area of each pixel of compact one will be 43/12^9=0.000000008 mm2. In the other words, each pixel of DSLR camera will be bigger more than 20 times of compact camera’s pixels. Now it is clear that DSLR can gather more light when there is not enough light. This is called “pixel density” of a sensore.
• Bigger sensors will enable you to have more control over the depth of field. So you can choose to have a picture with blurred background or sharp background. In the fllowing pictures you can see the difference:
  

  Sharper Background, bigger f-number, smaller aprture

  

  Blurred Background, smaller f-number, bigger apperture

  As you can see, when depth of field is wider then a wider range will be in focus (in the top picture all fllowers are sharp), but when depth of field is narrow then just a smaller range will be in focus (in the down picture juct the fllower in the middle of image is sharp). In my opinion this is one of the things that I really missed in my camera.

3- Exchangeable Lenses: In the DSLR cameras you can change lens in order to have ability to take different pictures. For example when you want take picture of animals, then you should have a lens which can zoom very good, for example 10 times. When you use 10x zoom to view your subject, e.g a grizzly bear, from 100 meters distance, your picture will be exactly similar to a picture taking by a 1x zoom lense in the 10 meters distance (I prefer the first option, what about you ). This kind of lenses are called tele(photo) lens. On the other hand we have wide(angle) lenses which have a wider feild of view. This type of lenses are used for landscape phography. The important thing is that some of the compact digital cameras (like mine, Nikon p5100) can be attached to other lenses to have a virtual ability like DSLRs, but this is far from the DSLRs lenses. You should not be fooled like me, I read this.

For more information take a look ar this page on wiki.

Finally I compare three DSLRs cameras to compare their features, Olympus E-510, Nikon D80, and Canon EOS 400D / Digital Rebel XTi, here is side-by-sidecomparison. I think based on their prices 522, 724, and 599, Canon is the best choise.

Useful books for Computer Scientists (2)

I start talking about my favorite books in this post, but it has not finished yet. I am going to talk about remained book now. I finished 5 of them and 4 more will be finished today.

• All of statistics [Amz, official page]: This is a great book for learning both probability theory and statistics. This book can be used as a reference guide also as a text book. If you are not familiar with the context and need more examples to practice, then you need another book. I have ordered “Schaum’s Outline of Probability and Statistics” as complementary book (I have not received it yet). The great thing about writer is that he has a good perspective of statistics, machine learning, and data mining. So his book is useful for students in all of those three majors. I suggest to read the book very carefully because if you find a type in his book he will add your name to this list, as he added mine.
• Learning Bayesian Networks [Amz]: This is one of the candidates text books for next semester’s “graphical probabilistic models” which is going to be taught by Russ.
• Pattern Recognition and Machine Learning [Amz, official page]: This book was the main text book of Machine learning course in the previous semester. Before I read the book I had thought this book should be the best patter recognition book (in comparison with Duda and Hert book). Right now I think it is better but not the best one. Some of the explanations in this book are hard to understand. I had too much problem with it, maybe because of lack of underling knowledge. But not only me, even professor claimed that some parts are hard to understand (specially Kernel chapter). After I finished All of Statistics (AoS), I will give it another chance.
• The Elements of Statistical Learning [Amz, official page]: I think, this book is the best current text for machine learning and pattern recognition. I have not read the book, yet I have heard a lot about it. If you have enough statistics background you should read this book to get a good view of machine learning. I will start reading this book after I finished AoS, certainly.
• Introduction to Statistical Pattern Recognition (2nd edition) [Amz]: Although this book is old (1990) but it is still quit useful. It uses a comprehensive mathematical formulation which makes it different from all the books. All books use examples and formulas in 1 dimension and then they say “General n-dimension case is exactly like 1 dimension case”. Fukunaga never uses this sentence because he explains every thing by the general n-dimn case! The other honor of book is its rate in the most cited list on the citeseer: 67. It is the 24th most cited book, amazing! Moreover. it is the favorit book of ARM, so it IS a good book.
  

Interesting thing about these books is that almost all of their writers are among the 10000 most cited writers of the computer science, here is their rate:

• 84. J. Friedman
• 280. S. Russell
• 419. R. Tibshirani
• 588. T. Hastie
• 965. K. Fukunaga
• 1276. P. Norvig
• 6111. M. Bishop
• 7915. W. Mackay
• 8315. L. Wasserman

Anyway, if you find time try to read these books also wish me to find time (do not waste time) and read them.

Useful books for Computer Scientists

Today I took a picture from the book shelf in my office. I did it on purpose because I want to talk about the great books which I love. Here is the picture:

From left to right:

• Molecular Biology of the cell [Amz]: The big red book is one of the main references for biology. I am not an expert in that area, so I am not a qualified person to talk about its weaknesses and strengths. However I can say my opinion, it is a big book which contains every thing about biology. It also has lots of colorful pictures. I do not recommend you to buy it because with that money you can buy a compact digital camera! But I highly recommend you to rent it from library.
• Artificial Intelligence, A modern approach (2nd edition) [Amz, official page]: This book is the Bible of AI, it is the greatest book I have ever seen. The main feature of book, also known as AIMA, is its simple and clear explanations. When ever I start to learn a new thing in AI, for example Bayesian networks, hidden Markov models, Kalman filter, neural networks, etc, I first take a look at this book. Its simple explanations and spectacular examples make it easy and even enjoyable to learn a new thing from this book. I highly recommend you to buy it, although its price is equal to a compact digital camera, it worths!
• Information theory, inference, and learning algorithms [Amz, official page]: Although I have known this book for less than 2 months and also I have read just one chapter of it (Monte Carlo sampling methods), I like it. Its writer is a physicist so he can impress you by some simple interesting things. For example some where in the book he is talking about the problem of algorithms with exponential complexity, like one that needs 2^1000 multiplication operations. Then he tries to convince you (albeit if you need, like me) that 2^1000 is a big number. “If all electrons in the universe (2^266) were powerful computers that can perform 2^40 operation per second, and all of them were working from the beginning of world, 2^58 seconds ago, then you must wait 2^634~10^190 more world ages to finish it!!!!” Amazing! He has a funny table which compares his book with Harry Potter and based on it he concludes that you must buy his book instead of Harry Potter.
• Probabilistic robotic [Amz, official page] : One of my courses’s text book.
• Scientific Computing [Amz, official page]: One of my courses’s text book.

I will talk about other book latter, stay tuned …

Smarties neat commercials

I like this commercial very much, it contains a song which has been sung by 6 different people. First take a look at it:

It seems that all of them have sung the full sing, and then a part from each one is spliced together to from it. I like the part which says “makes no different what group I’m in” the most. Last night when I was watching TV they show a new version which all of it was sung by my favorite part singer. Wow, I enjoyed it. Today I searched in youtube and I realized that each part has a “color” and full version of all of them exist there. My favorite part is “blue”, although it is not my favorite color. Here it is:

I like her voice and her gesture. After I watched all colors, I still believe this part is the best one. Here are links to the other ones. Take a look, I am sure you will agree with me:

Vote Red

Vote Purple

Vote Pink

Vote Yellow

Vote Green