The Great Fashion Computing Challenge

["Michael McGeachie" <mmcgeach@>]

At GSB last week a very important topic came up.  (No, not the Economic
Theory of Dating at MIT.)  (No, also not the upcoming inaugural season of
CSAIL-American Idol.)  It was, the terrible state of fashion adoption in
CSAIL.  That motivated myself and Marvin Minsky to suggest the following
Grand Challenge, through this week's GSB announcement.  (And yes, most Grand
Challenges are issued through GSB mailings.)

        --  The Great Fashion Computing Challenge  --

Design learning algorithms that understand fashion and proposes wardrobes.
With the preponderance of fashion magazines and fashion ads, there are
countless examples to feed to learning algorithms.  Minor image-segmentation
issues notwithstanding, the following questions should be answerable by your
algorithm:

- Given any CSAIL lab member, predict the decade he or she experienced
fashion-sense arrest.

- Using advanced clustering techniques, determine the best dress shirt for
wearing a "Nerd Pride" pocket-protector.

- Define a "hipness" metric, H(n), and determine the minimum hipness
required to wear thick black rimmed glasses.

- With an understanding of retro-ology, predict when plaid flannels will
again be acceptable in a) rock clubs, b) software shops, c) Japan.

- Using photographs of the Stata Center and determine what human outfits are
most "Gehry."

- Partition recent CSAIL graduates into those capable of interviewing in New
York and those that must be relegated to California.



And speaking of recent CSAIL graduates, GSB follows graduation today.
Anyone graduating gets a free drink!  Sample our selection of non-alcoholic
Ginger Beer and Root Beer, while supplies last!


            +-                                                  -+
              girl scout benefit -+-  5:30 pm  -+- 32-G9 lounge
            +-                                                  -+

               For those coming from elsewhere: Building 32 is
                <http://whereis.mit.edu/map-jpg?selection=32>