Hello from Cambridge

This and next week I’m attending I-CAMP, an international summer school on liquid crystals. It’s taking place in Cambridge, UK. It has a focus on optics, which is very relevant to my current work for the master’s thesis. My thesis work is now mostly complete, I am presenting a poster on this topic at the summer school as well.

After only two days of lectures, I can already conclude that English food is bad. Breakfast is alright, but fried fish and sandwiches doesn’t cut it for a pasta-lover like me. Additionally, seats in the lecture hall we used today were very uncomfortable, and spending 8 hours in the made my back hurt. Apart from that, the city and its colleges are very nice, with all the grass and trees. I was surprised at first, but you are allowed to walk on most of the grass areas, which is great for any activity in sunny weather.

My poster is about numerical modeling of light propagation through a fibre filled with liquid crystals. The interesting part is that a single laser pulse splits into 8 regions. Most people won’t understand much, but the pictures are nice. You can take a look here.

Free (harvested) energy

As a Physics student, I have to prepare a seminar for my classmates regarding some topic broadly connected with Physics. I chose Energy harvesting, as suggested by a professor, without knowing much about the field or the commercial offerings. However, aften collecting many sources and writing the article, I have come to have a much better opinion of our future than before. Namely, I discovered things like this:

Model analysis report in hexameter

Unfortunately, only in slovene.

Education in America and in Slovenia

Background

I’m a 4th year (22 years old) student, studying Physics in Ljubljana, Slovenia. Without trying to sound overconfident, I am quite good at this and have passed the previous years with a very good average, despite my study being one of the hardest in the (rather small) country. So, with my country being small and all, my parents wished that I should study abroad, preferably in USA. However, having heard high praise of our university and not wanting to leave my home just yet, I decided to stay and study here.

I consider myself a good learner, especially when learning by myself, so I taught myself programming in a handful of programming languages. Recently though, I saw that Stanford university is making some computer-related classes publicly available, with video lectures and programming assignments for testing. I immediately signed up for Machine Learning. I skipped over the introductory lectures and tests, seeing that I learned all I needed for this class in previous years. But when the first programming assignment was published, I read it and dropped out of the class immediately.

The American way

It was a 15-page (yes, fifteen pages) instruction to implement a Linear Regression algorithm. For those that don’t know what this is, it’s a task of fitting a straight line between a series of points. It’s publicly available, so if you’re interested, here is the whole package with instructions and a large collection of programs to help you. It takes more time to read through the whole thing than to write a solution.

Fitting lines through data points is something we did in high school, and we learned to do it properly at the beginning of the first year at the university. It’s a very simple problem, and every single mathematical program has a function for it built-in (even Excel). It is even analytically solvable, so there is an exact formula that calculates the coefficients given a number of points.

The Slovene way

While my field is Physics and the class I’m talking about is from Computer Science, I did choose a Computational Physics program, so I have similar subjects at my university as well. I think the closest is Model Analysis, a class with more emphasis on numerical methods, mathematics and computers than on actual physics. The complete first assignment (less than one page) is here, in English.

Here, the task is to figure out how we should slow down or accelerate to arrive to the traffic light at the exact time when the green light turns on, given that we know this time in advance. Besides it involving much more advanced mathematics, the choice of tools is entirely up to us. We are always encouraged to find interesting similarities with other problems, and the assignments always give us some free space to interpret or spice up the solutions.

Conclusion

I’m definitely glad I stayed at home, where I will receive a proper education, even though it’s not a prestigious. But I don’t like being treated as a little kid in school any more. I don’t know what the atmosphere at American universities is, or how much better their access to equipment is, but as long as I’m going to school, I want to learn as much as I can in these years, not just drag along lazily.

Simple Sudoku Solver Source

Here’s the code of a brute-force implementation of a recursive sudoku solver in python. An optimised version will be released after the homework deadline. I’m posting this in english just to be evil to my classmates. The important part is highlighted.

Read more and the source will be with you.

Sudoku solver

I have programming experience mostly in C++, but we’re learning Python in school this year. The latest assignment was to write a sudoku solver, first a brute-force one and then a more intelligent implementation.

My current result takes about 0.1 seconds for a hard sudoku, which I think is fast enough, especially as I have no idea how could one improve it.