[Ml-stat-talks] Fwd: [Theory-Read] ML reading group (special talk on Monday)

David Blei blei at CS.Princeton.EDU
Fri Jan 31 15:04:37 EST 2014


elad hazan from the technion is giving a special seminar on monday
2/3.  this looks fascinating, especially to those of you interested in
optimization on large data sets.  see below.

note: elad hazan is one of the co-authors of the adagrad algorithm,
which is a game-changer in stochastic optimization, and received his
phd from our CS department.


---------- Forwarded message ----------
From: Pranjal Awasthi <pawashti at princeton.edu>
Date: Fri, Jan 31, 2014 at 1:12 PM
Subject: [Theory-Read] ML reading group (special talk on Monday)
To: "ml-reading-group at lists.cs.princeton.edu"
<ml-reading-group at lists.cs.princeton.edu>
Cc: "ehazan at ie.technion.ac.il" <ehazan at ie.technion.ac.il>,
"theory-read at lists.cs.princeton.edu"
<theory-read at lists.cs.princeton.edu>

Hello all,

We will have a special seminar on Monday Feb 3rd at 12:30pm. We will
meet in CS 302 and our speaker will be Elad Hazan.
Please find the talk title and abstract below. Hope to see you there.


Projection-free online learning
Modern large data sets prohibit any super-linear time operations. This
motivates the study of iterative optimization algorithms with low
complexity per iteration. The computational bottleneck in applying
state-of-the-art iterative methods is usually the so-called
"projection step".
We consider projection-free optimization that replaces projections by
more efficient linear optimization steps. We give the first
linearly-converging algorithm of this type, and describe applications
to online learning and stochastic optimization.

Theory-Read mailing list
Theory-Read at lists.cs.princeton.edu

More information about the Ml-stat-talks mailing list