TIT
1998
13 years 11 months ago
1998
We show that if the closureof a function class F under the metric induced by some probability distribution is not convex, then the sample complexity for agnostically learning F wi...
EC
2008
13 years 11 months ago
2008
We present a statistical model of empirical optimization that admits the creation of algorithms with explicit and intuitively defined desiderata. Because No Free Lunch theorems di...
COLT
1994
Springer
14 years 3 months ago
1994
Springer
We examine the relationship between the VCdimension and the number of parameters of a smoothly parametrized function class. We show that the VC-dimension of such a function class ...
COLT
2001
Springer
14 years 4 months ago
2001
Springer
We investigate the use of certain data-dependent estimates of the complexity of a function class, called Rademacher and Gaussian complexities. In a decision theoretic setting, we ...
CDC
2009
IEEE
14 years 4 months ago
2009
IEEE
—TD learning and its refinements are powerful tools for approximating the solution to dynamic programming problems. However, the techniques provide the approximate solution only...