I’ve been interested in ML since my Bachelor, when I attended an homonymous class taught by Roberto Navigli. My project was to solve a task taken from the just-finished SemEval 2010, the third: Cross-lingual Word Sense Disambiguation. In the relative subpage more details, but generally speaking the results were very good (sometimes better than the participants to SemEval). One of the best personal achievement, in addition to have had excellent results with fairly simple algorithms and a good amount of data manipulation, was to have learnt a very good framework: Weka.
I’ve found myself working again on ML problems during my MSc, for an homonymous course. I’ve learned a toolkit called NLTK and worked mostly on sentiment analysis and information retrieval. No projects of interest, but in case I decide to work again with this stuff I’ll create a subsection.