Jonathan Goodman gave the departmental seminar yesterday (10 Dec 2012) and I was amused by an extended analogy he made. After his (very clear) intro, he said that math talks were like action movies. The overall theorem and its applications provide the plot, and the proofs provide the action scenes.
Bob:
what interests me is the category “action movie.” My impression is that there are a lot of action movies nowadays but before Raiders of the Lost Ark it wasn’t really a genre at all.
Maybe it skipped a generation. In the early days of silent films, we had Buster Keaton and Harold Lloyd, pulling Jackie Chan style stunts.
There were always lots of movies with action. Sci-fi has always been action packed, as have war films and westerns. And don’t forget the kung fu films!
That’s great! Makes me want to see one of his talks.
Somewhat related: Are the charts for your dept seminars publicly-available? Nicolai Meinshausen’s talk is of particular interest. There are a couple of other interesting sounding ones too.
Related to Meinshausen’s lasso publication: Has anyone done “regression shrinkage” (“sparsity promotion” in my neck of the woods) using a multivariate t-distribution rather than L1? Just curious. My experience (hyperspectral data analysis) is that L1 is a bit soft as a penalty term.