Tuesday, November 1, 2011

Inflation?

I was wondering whether this reduction of KR to only 4 weeks also results in a worse service of the university towards the students in the sense of teaching hours. I tried to look up the old schedule (of the last years) but it seems that those links on http://rooster.uva.nl/ only lead to empty pages, although loading this empty page takes a while...

Anyway, when comparing KR with ELPL I figured that there are less teaching hours for KR, of course.
According to my calculation ELPL has 24 hours of lectures and labs, where KR only has 14 hours. Please correct me if I'm wrong.

Well of course, we should be happy about that, otherwise it might be extremely stress full. Really? What are these video lectures good for? A replacement?

Hmm, what am I paying more than 1700€ a year for??

IR Assignment 2.

Hi all.

Once more, something completly abstract and incoherent... This is way too thoretical, for even the practical part to make any sence.
So, in a few words to wrap-up:

"What the hell are we supposed to do?"

Sentences like "It is up to you to come up with a way of doing this" or "the data gives you a lot of freedom as to what you want to look at in this assignment" or "you can try very different experiments to get there" or "and much, much more".

You know, I am used to having to do something under strict specifications and explicit requirements.
e.g. Use this. Do that. Find 1, 2, 3. Build 4, 5, 6.

The endless freedom of specifications and the randomly seeming resources given for this assignment, really frustrate me... I need to work in something solid!

Sunday, October 30, 2011

MAS assignment 2 deadline

There was some confusion over the deadline for the second MAS assignment, about whether it was final or points get deducted if you are late. I sent an email to Matthijs and he just replied: 
0.5 grade point per day
So, to re-iterate, assignment 2 is just like assignment 1; you get a penalty of half a point for each day after the deadline you are late. So, if anyone feels that one more day can really make a difference (like me), then this is good news.

Saturday, October 29, 2011

Designing an unbiased reward function

Still doing part 1 of the second MAS assignment. I've been wondering about the proper way to set the reward function. If we want our system to learn, the reward function should only include the objective of the game, right? If we start letting it know that collisions are bad or that moving closer to the prey is good, then what's the point of using Q-learning? You might as well just give it the perfect algorithm from the get go.

Is my thinking right? If it is, then the correct way to do this is to only give out reward when the prey is caught. Then, the Q-values should propagate the information. For example, since collisions would delay this reward, then the Q-values of moves that are likely to lead to collisions would become low in the long term. Does that make sense?

Intelligent Multimedia Systems

Has anyone else had a look at the blackboard page of Intelligent Multimedia Systems?


First thing I noticed was what a huge mess the page is!! There's duplicate links, links that don't work, links for stuff from last year (and the year before last year) and in general just a hell of a lot of links.

It seems we will be reading loads of different articles during the next period, if last year's "required readings" is anything to go by. Its 15 articles to be precise, although for some we only need to read certain pages. There's also a weekly lab. I see that this subject was a 3-4 month course last year, so lets hope that they don't just squash it into 8 weeks without making any adjustments in the material...

And perhaps the most worrying thing I noticed: of the 24 students who were registered for the course last year, 13 failed! WTF!!

(see Tentamenuitslag_IMS_2010_2.xls)

Thursday, October 27, 2011

Copious amounts of grass-fed beef burgers

Hi all,

I should be studying at the moment, but thought about it deeply and decided to do some random websurfing instead. I came across this entertaining video about the state of the ML job market, and thought I'd share it.

While the stuff about immigration policy is mostly irrelevant to our "culture with teh scientific understanding", this is still interesting stuff, whether you want to eat grass-fed beef burgers at the google, or are thinking about creating a startup..

It is especially interesting, because the official statistics show that 50% of UvA AI masters don't find a job "at a high level". This is mostly because of epic fail on the part of the students, but it may also have to do with the fact that not everything we learn is equally relevant to real-life problems.

Moral of the story: learn your ML, and contact me when you've graduated and understand the machine learning and the optimization ;-)

Wednesday, October 26, 2011

IR Exam

So, it is open book right?

Last year it was, also open notes (you can bring any writing you want)

btw. I could have sworn that the exam was on thursday? Was it moved at some point?