UAI Conference

For the last four days I was attending the Uncertainty in Artificial Intelligence conference. I got free registration by volunteering to help out with the conference. I ended up staffing the registration desk for a few shifts as well as helping move some boxes. Not a bad deal for getting to attend interesting AI talks and eating lots of free conference food.

Most of the talks were actually very hard to follow. They dealt with advanced topics and assumed that the audience was already familiar with the concepts being discussed. I still enjoyed the talks however since there were some really interesting applications for the techniques being discussed. I probably also picked up some important AI terminology which should prove useful.

CIspace

After a month and a half of hard work, I have released new versions of the following CIspace applets:

• Search - Version 4.4.0


• CSP - Version 4.5.0


• Stochastic Local Search for CSP - Version 4.5.0


• Decision Tree - Version 4.3.0


• Deduction - Version 4.2.0

These releases include many bug fixes and minor feature additions. I also made changes to the "graphToolKit" package which is used by all the applets for the basic user interface. Try out the new applets and please report any bugs encountered. Finding bugs is often far more difficult than fixing them.

I have been putting most of my effort into adding features for the neural network applet and the bayes applet. Depending on how progress goes, I should be able to release the new versions in a week or two. Some ideas we had for future CIspace projects included customizable inline applets, a new robot navigation applet, and a natural language processing applet using hidden Markov models.

I have really been enjoying working in the LCI lab. I get to attend a lot of interesting presentations about artificial intelligence. Hopefully I should be able to go to the upcoming AAAI conference (one of my supervisors is the president of AAAI!).

In other news, I have decided to spend a fifth year here at UBC and get a double major. I will be majoring in Computer Science and Cognitive Systems (Cognition and Brain stream). After I graduate I plan on attending graduate school (most likely in the U.S.).

The Joy of Tech

In Celebration of Spork

Excited about my new spork, I decided to compose my first haiku:

Hunger approaches.
Food is prepared. Spork in hand,
eating commences.

Penalty Box

As you can see, I have been enjoying my increased bandwidth limit. Quite a change from the 2GiB per *month* that I was limited to last summer.

Numbers

I am now the exclusive owner of the number 6D 37 00 22 9A 51 7D D7 5D 26 5B 75 2B 02 D2 6F. This number is mine and you are not allowed to know about it.

You too can claim a number (service may be unavailable due to the Slashdot effect).

"We own integers,
Says AACS LA.
You can own one too."
(C) 2007 Edward W. Felten

Insanity

I just finished writing my formula sheet for CS320. I decided that instead of memorizing anything, it might save time if I just try to squeeze a semester of course work onto a single A4 piece of paper. Well... I ran out of room :(. It is now 1am and there are 11 hours left until the final. I better get back to studying.

Motivational Comics

Calvin offers his views on the world of academia:




... yes, I should really be studying right now since I have three exams within the next four days. Instead I find myself reading Calvin and Hobbes and learning how to play Go. I am actually proud of my progress learning Go since I can now beat the GNU Go artificial intelligence on its hardest difficulty level. I find that procrastination actually makes me more productive (except, of course, on the task I am procrastinating).

Easter Riddle

Since there was relatively low participation for my last riddle, the first person to correctly decode my Easter egg will win a mystery prize!

Laptop Theft

I found out yesterday that Nath's laptop and passport were stolen while he was on the overnight bus from Banda Aceh to Medan. In an attempt to identify the thief, I decided that it would be helpful if I could log an IP address when the laptop connects to the Internet. The laptop is set to automatically log on to MSN, so I set up a buddy pounce to send a message as soon as the laptop connects. I don't think it is possible to trace an IP address directly through MSN because the packets travel through the MSN servers before heading towards their destination, so I set up the buddy pounce to send an irresistible URL to a webpage with a web tracker instead.

Today the thief actually logged onto MSN. The buddy pounce sent the link, and then he/she logged off a few seconds later. Unfortunately the thief did not take the bait, and didn't click on the link. I set up another buddy pounce just in case he/she decides to log on to MSN again, although I think that would be extremely unlikely. If I had succeeded in getting the IP address it probably would have been possible to track down the location of the connection... oh well :(

Cryptomni Version 1.1 Released

Changes in this release include improved source code quality, as well as the ability to encrypt/decrypt multiple files from a single key file.

http://sourceforge.net/projects/cryptomni/

Crossing the Rubicon

For the last two weeks I have been contemplating what I consider to be the most important decision of my life. I have been offered an internship position at Microsoft, as well as a summer research position here at UBC. Both of the positions are related to artificial intelligence, but the Microsoft position will probably be more focused on software development. The research position will involve working on a program called CIspace (http://www.cs.ubc.ca/labs/lci/CIspace) funded by a NSERC USRA (Natural Sciences and Engineering Research Council of Canada Undergraduate Student Research Award).

To me this decision represents much more than just what I plan on doing this summer, but what I plan on doing for the rest of my life. On one hand I can do research on fascinating topics directly related to what I am interested in, but on the other I can start a career as a software developer with a much higher salary but work on something which I don't find as fulfilling. Choosing the Microsoft position would probably open many career opportunities within the software development industry, but I fear life as a code monkey will be stressful and unsatisfying. Today I was forced to choose between the two positions, and I have decided to go with CIspace. I really hope that I don't end up regretting this choice later in life.

FoxTrot

Windows Vista Business

Since I get a free copy of Windows Vista Business thanks to MSDNAA, I decided to install it on a partition this morning. I definitely think its an improvement over XP. The Aero interface looks awesome, and a lot of the features available in XP have been tweaked and improved. My only real complaint about it is the User Account Control, which displays an annoying dialog message requesting security permission way too often. I will continue to use Ubuntu as my main desktop environment, but may occasionally boot to Vista if I need to use software which doesn't run on Linux.

Cryptomni Version 1.0 Released

Cryptomni is a program which can encrypt and decrypt files using the one-time pad cipher. A key file is created using the cryptographically strong random number generator SecureRandom. If a key is truly random, kept secret, and never reused, this encryption algorithm can be proven to be unbreakable.

This project is hosted at SourceForge.net, and can be found at http://sourceforge.net/projects/cryptomni.

PDFTron Interview

I had an interview for a summer internship position with a local software company called PDFTron this morning. I think the interview went pretty well. There were some typical questions asking about my experience and job expectations, as well as some more technical questions involving programming concepts. They also gave me four programming tasks which I had to implement using C++ (writing out code on a piece of paper):

1) Write a function which converts a string to a positive integer. Ignore the '+' and '-' characters, return 0 if the string is empty, and return any interpreted digits if a non-numerical character is encountered. Here was my solution:

int atoi (char * string)
{
  if (*string=='\0') return 0;
  int value = 0;
  while (*string!='\0')
  {
    if (*string=='0') value+=0;
    else if (*string=='1') value++;
    else if (*string=='2') value+=2;
    else if (*string=='3') value+=3;
    else if (*string=='4') value+=4;
    else if (*string=='5') value+=5;
    else if (*string=='6') value+=6;
    else if (*string=='7') value+=7;
    else if (*string=='8') value+=8;
    else if (*string=='9') value+=9;
    else if (*string=='+') value/=10;
    else if (*string=='-') value/=10;
    else return value/10;
    value*=10;
    string++;
  }
  return value/=10;
}

2) Swap positions of the nibbles in a byte:

char nibbleSwap (char byte)
{
  char nibble1 = byte&0xF0;
  char nibble2 = byte&0x0F;
  return ((nibble1/16)|(nibble2*16));
}

3) Write a function which computes x to the power of n using O(log n) multiplications.

int power (int base, int n)
{
  if (n == 0) return 1;
  if (n == 1) return base;
  int half = n/2;
  int value = power(base, half);
  if (n%2==0) return (value*value);
  else return (value*value*base);
}

This has a recurrence relation of T(n) = T(⌊n/2⌋) + Θ(1) for n>=2 and T(n) = Θ(1) for n<2. T(n) is ∈ Θ(log n).

4) Implement a function which draws a line, given coordinates for a start point, coordinates for an end point, and a two dimensional array representing grayscale pixels (8 bits per pixel). I didn't get around to attempting this question since I ran out of time.

PlayStation Emulation

I recently installed custom firmware version 3.03 OE-A on my PSP (thanks Dark_AleX!). This allows nearly flawless PlayStation emulation without having to use the official PlayStation Network service for the PS3. I have been spending way too much time replaying FFVII instead of studying for my impending midterms...

Robotics Conference

I spent the last two days attending the "Robotics for Society: New Directions in Cognitive Science" conference. It took place at Robson Square (downtown Vancouver). Invited speakers included:

James Little (UBC): Space for interaction with a robot
Gordon Cheng (ATR): Humanoid robotics perspective to neuroscience
Richard Vaughan (SFU): Assault and batteries: on the utility of robot aggression, competition and violence
Rodney Brooks (MIT): Robotics and everyday life
Alan Mackworth (UBC): You, robot, do no harm!
Hideki Kozima (NICT): A social robot in the wild world; practices in therapeutic and pedagogical applications
Stefan Schaal (USC): Computational motor control, humanoid robotics, and their societal relevance
Masaaki Honda (Waseda): Talking robot mimicking human speech production
Richard Rosenberg (UBC)

There were also various poster presentations outside of the theater. I found the conference very enjoyable because it provided exposure to a large variety of robots currently being researched and developed. My favorite presentation was by Rodney Brooks. Not only was he an excellent speaker, but he is also very well known in the field of robotics (I read a book by him in high school, as well as some of his research papers). He also founded the iRobot Corporation (well known for the Roomba robots). Unfortunately, attending the conference has put me way behind on my homework, so I am going to be spending the next 24 hours trying to catch up :(.

Dilbert

Mélange

High-definition Video:

I just acquired an amazing HD version of The Matrix. This is the first movie that I actually have to downscale in order to display on my monitor (its resolution is 1920x1080, while my monitor is only capable of 1280x1024). Here is a screenshot of the video playing in its original resolution, spanning across multiple desktops on my Beryl cube:


Mysterious Function Update:

It turns out that the graphs I made in my previous post were actually bifurcation diagrams of the logistic map. Wikipedia has a great explanation of the equation and how it is related to chaos theory.

Summer Internship:

I have been very busy looking around for summer internship positions. So far I have been applying to the big tech companies like Google, IBM, and HP. I am not very happy however, because most of these positions do not sound very interesting. I do not want to become a code monkey. What I really want is to be able to work on some interesting AI research. Unfortunately after examining the job market these sorts of positions seem to be rare. I also attended a career fair on Thursday, and applied for an internship position at Microsoft.


Das Experiment:

This afternoon I participated in a psychology study. It had the following experiment description: "This study will explore the manner in which an herbal extract, 'Salin', can affect memory." The experiment involved drinking a glass of iced tea flavored liquid, and then filling out some surveys. I soon realized that the forms were doing a terrible job at actually testing any of my memory skills, and I became suspicious. After handing in my forms the experimenter asked some questions - one of them being "did you notice anything unusual about the experimenter?" I answered no. He then let me know that he was actually a different person then the original experimenter. Half way through the study they actually switched people without me noticing. This was the real purpose of the experiment, to see how I reacted to a change that I didn't expect or consciously notice. The purpose of the forms was to determine if the switch might have had some sort of subconscious effect. The drink they gave me was a placebo. I wanted an actual memory enhancing drug!

Magic:

This evening I went to a magic show at Norm Theatre. I was disappointed since it seemed mediocre, and I knew how *every* trick performed actually worked. After the show they played the movie The Prestige. I think this was the best of the three big magic movies of 2006 (the other two being Scoop and The Illusionist). I thought the best aspect of the movie was its plot, and how it kept the viewer constantly guessing at how the magicians performed their feats.