Northern Bites Blog

Upgraded our kick-ass WordPress theme K2 to 0.9. Rockage.

A long while back I applied for a Google Analytics account and sometime during the Germany crunch it came through.  So, finally, I’ve gotten it installed via this incredibly easy plug-in called, astonishingly, Google Analytics plug-in for WordPress.

Analytics will allow us to figure who and where you robotics fiends come from and organize it in the typical Google-awesomeness way.

We aren’t the only Northern Bites around–it’s also what is I’m sure a fine cafeteria at Northern Lights College in British Columbia.

Northern Bites is also a ‘meals-for-purchase’ plan on Alaska Airlines.

We’re on a newly-built server under our complete control, and as your browser is telling you, our new address is:

http://robocup.bowdoin.edu (or just http://robocup/ if you are on campus).

Most of the significant improvements are for the elite hackers that make up the Northern Bites team, but for our fans and admirers, we offer a new, sweet dns, real permalinks to pages like Archives and Contact, and much faster and smoother page loading.  Also, check the Media page for future updates which are comfortably expected in the coming decade.

Roll, nBites!

C3PO is doing laps upon laps in our Robot Lab this afternoon in hopes of finding a walk faster than all previous ones.

The poor guy is trying really hard. We’ve experimented with policy-gradient algorithms, but now we’re running a hill-climber algorithm, which is essentially the easiest machine learning algorithm to digest and write.

Machine Learning works like this: you want to find an optimal solution to a problem, and you want the computer (or dog) to find it for you. Computers are very good at doing things over and over again, and so we can leverage this ability and have the computer try many variations of one solution until we find an optimal one.

Machine Learning algorithms guide these solutions and dictate how and when these variations occur. Flashy algorithms these days have flashy titles like Genetic Algorithms and Downhill Simplex Methods. But ours is much simpler.

Hill Climber algorithms start with a base solution and then vary this solution a certain amount of times, gauge each variation’s success, and then simply chose the best solution so far and repeat. For us, our ‘base solution’ is a walk — a set of inverse kinematic parameters — that we hope to optimize.

Right now C3PO is in lap 258 of possible 800. Come on, C3PO!!

The new server is up and running, and it works great. I’d post this on the old blog, but we’re avoiding changing anything on the old blog, since there’s a new copy running on the new server. Instead, here’s it on the new blog.

Check it out at http://robocup.bowdoin.edu. Wait, you’re here already, aren’t you?

That’ll redirect you straight to this blog again. Other important section of the site include the SVN repository, served by DAV/SVN, and the Trac Wiki and Project Manager. These new sections are available by clicking on the new links under “Bites” on the right side of the main page, or by navigating to

https://robocup.bowdoin.edu/svn
https://robocup.bowdoin.edu/trac