Tuesday, July 16, 2013

MS Faculty Summit Day One

Day one of the faculty summit was a great deal of fun.  Primarily, I found I enjoyed catching up with people -- I saw a number of past/current collaborators, as well as many people who I enjoy talking to but don't get to see often enough.  The summit is well attended -- several hundreds of people -- so there's plenty of people to see.

The morning keynote session included a large chunk of time with Bill Gates.  He talked briefly and took a large number of questions.  In his remarks, he spoke about the bright future he saw for software, particularly the potential in making large advances in big science problems.  We're able to do so much more now, we can be much more ambitious about what we can do.  Then he spoke about the areas the Bill Gates Foundation is focused on.  Education -- the Bill Gates Foundation funds the Khan Academy and several MOOC projects.  He sees MOOCs as a way of increasing personalization in education.  He talked about fighting disease, and in particular disease spread modeling.  He also talked about genetic modification of crops to improve disease resistance, drought resistance, and nutritional value, and finally digital microfinance tools.  There weren't any particularly controversial thoughts or questions;  the most memorable to me involved a question about the patent/copyright/IP system, which Bill Gates tried to strongly defend.

Our morning session on verification for cloud computing systems went well.  (All the sessions seemed to start a bit late, so we were a bit pressed for time.)  Bryan Parno and Michael Walfish gave excellent presentations, and I thought our talks fit nicely together in terms of giving a pretty complete picture of goings-on in the area.  The after-lunch session on the economics of computing was very good -- Dan Huttenlocher talked about the theory of badging (giving a model for how badges on things like StackExchange can motivate people to different behaviors, in both theory and practice), Muthu talked about a new advertising market he was interested in, and Eva Tardos talked about composable and efficient mechanisms (and the class of smooth mechanisms).  I thought both sessions were slightly underattended, though -- the bulk of the attendees seemed to be focused on machine learning and related work.  That must be where the "action" is these days.  

The evening offered a boat cruise -- and a heat wave throughout the US translates into beautiful sunny weather and quite comfortable temperatures for a boat cruise in Seattle.  The trip included great views of Mount Ranier, bridges rising to scoot out of our way, and apparently we boated by (one of?) Bill Gates's houses, although I missed that.  Lots of fun conversations (although punctuated by important issues -- how universities handle parental leave, how to handle cheating in classes, growth and hiring) and catching up on various goings-on.


Anonymous said...

Would you say machine learning is possibly the most popular area of computer science today? I noticed that UW hired 7 new faculty members for this year, and many work in machine learning.

Michael Mitzenmacher said...

I don't personally have the stats to back that, but anecdotally it does seem to be the "killer area" right now. Perhaps both because the number of applications where it's proving successful is increasing, and it offers a methodology that provides hope for dealing with "big data" problems.

Anon said...

I think "Kahn Academy" in the post is a typos and should be "Khan Academy" ( I don't think there is any such thing as a "Kahn Academy"). Please feel free to delete this comment after the correction.

Anon said...

"Muphry's law" strikes again: "typos" in my comment should have been "typo".