Sorry to spam everyone again; posting the agenda to the group this
time:
For the first week, in addition to meeting all of you and discussing
organization/ect,
I'll present the paper
"Fast global convergence of high-dimensional statistical recovery" by
Agarwal, Negahban
and Wainwright:
http://arxiv.org/PS_cache/arxiv/pdf/1104/1104.4824v1.pdf
In my opinion, this paper has some very interesting implications for
the relationship
between statistical inference and optimization in high-dimensional
spaces, and also for
the practicality of solving the computational problems that arise in
this area. I'll try
to highlight some of these in the talk.