We are delighted to announce details about Hahn Koo’s workshop which will take place tomorrow – Wednesday November 4th, 2.00-3.00 pm, in Clark Hall 445:
Hahn discusses how sublexical statistics can be used to characterize difference in sound patterns between native words and loanwords. To demonstrate the effectiveness of the approach, Hahn also presents a computer program that automatically identifies loanwords in unlabeled monolingual corpora. In the process, he reviews the basics of n-gram models, naive Bayes classifiers, and the expectation-maximization algorithm.
Staff and all students welcome!