This talk will describe a simple, general algorithm for learning to play any matrix game against an unknown adversary. The algorithm can be shown never to perform much worse than the best fixed strategy, even if selected in hindsight. Moreover, because of the algorithm's moderate resource requirements, it can be used even when working with extremely large game matrices. Taken together, these properties make the algorithm a good fit for a range of machine-learning applications, some of which will be discussed, for instance, to the problem of learning to imitate the behavior of an "expert" while attempting simultaneously to improve on the expert's performance.
Check out the slides of yesterday’s awesome colloquium on “Cool theorems proved by undergraduates” by Prof. Ken Ono from Emory University! :D
(Email from Zhaonan)
For those of you who are interested in participating in an REU this summer, there will be some students who have done REUs before at the math departments’ afternoon tea time to share their experiences. The time and location are:
Wednesday Jan.22nd 3:30 pm
Fine Hall 3rd floor Common Room
Come talk to former REU students and ask any questions you might have! There will be cookies as well.
I’ve also copied Alan’s earlier email for your information.
Here is a long list of REUs: http://www.nsf.gov/crssprgm/reu/list_result.jsp?unitid=5044
Deadlines range from early February to early March, so you should do some research of your own before our info session. Also, please note that most will ask for 1 or 2 recommendation letters.
Hi again nullset,