\magnification=1200
\baselineskip=20pt
\nopagenumbers
\font\big=cmr12 scaled \magstep2

\centerline{\bf STANFORD UNIVERSITY}
\centerline{\bf DEPARTMENT OF STATISTICS}
\centerline{\big DEPARTMENTAL SEMINAR}
\bigskip

\baselineskip=12pt
\centerline{4:15 p.m., Tuesday, October 2, 2001}
\centerline{Sequoia Hall Rm. 200}
\centerline{(Cookies at 3:45 in 1st Floor Lounge)}

\bigskip
\baselineskip=15pt
\centerline{\sl Peter Bartlett}
\centerline{\sl Department of Statistics}
\centerline{\sl University of California, Berkeley}
\bigskip
\centerline{\bf Error Estimation with Rademacher and Gaussian Complexities}

\bigskip

We investigate model selection strategies based on penalized empirical
loss minimization. We point out a tight relationship between error
estimation and data-based complexity penalization: any good error estimate
may be converted into a data-based penalty function and the performance
of the estimate is governed by the quality of the error estimate.
We focus in particular on Rademacher and gaussian complexities, and show
how estimates of these complexities can be computed.  We give examples
of the application of these techniques in finding data-dependent error
estimates for decision trees, neural networks and support vector machines.

(Joint work with Stephane Boucheron, Gabor Lugosi, and Shahar Mendelson.)

\bye