The federal government is extending its funding for their top-tier graduate scholarships.
Canada Graduate Scholarships are currently awarded to the top 2,000 masters and 2,000 doctoral students each year and are worth $17,500 and $35,000 dollars per year respectively. The addition funding promised will mean an extra 1000 scholarships awarded over the next 2 years. This is good news, especially when you appreciate that as of the 2006 budget, all post-secondary scholarships are tax exempt.
Tuesday, March 20, 2007
Friday, March 9, 2007
Flckr Photos
There are some wonderful photographs of the Unversity of Toronto campus on Flickr. To showcase these great photographs I have wrote some javascript which uses the Flickr API to grab a random photo from the list the most interesting photos tagged with uoft and display it.
Now, I took care to only grab the photo which use a creative commons license that allows there distribution, however, if you see an image of yours that you would rather not see here, please just let me know and I will take it down.
Now, I took care to only grab the photo which use a creative commons license that allows there distribution, however, if you see an image of yours that you would rather not see here, please just let me know and I will take it down.
Wednesday, March 7, 2007
Minimax Regret and MDPs
One topic I am currently thinking about is calculating policies for Markov Decions Processes when we are uncertain about the reward function. I find this to be an interesting question because we have good tools for allowing domain experts to specify the dynamics of a decision process, yet specifying the reward function is not always easy or intuitive.
Here's the idea: let a domain expert specify some kind of an incomplete reward function and use the minimax regret criterion to measure how much one could regret choosing a particular policy given the unknown reward function. We can further use the minimax regret to compute questions to ask a designer to quickly eliminate the uncertainty in the reward function.
Here's the idea: let a domain expert specify some kind of an incomplete reward function and use the minimax regret criterion to measure how much one could regret choosing a particular policy given the unknown reward function. We can further use the minimax regret to compute questions to ask a designer to quickly eliminate the uncertainty in the reward function.
Subscribe to:
Comments (Atom)