3718 search results - page 265 / 744 » On learning with dissimilarity functions |
268
click to vote
ICCV
2011
IEEE
14 years 6 months ago
2011
IEEE
We propose a novel way to induce a random field from an energy function on discrete labels. It amounts to locally injecting noise to the energy potentials, followed by finding t...
236
click to vote
SIAMCOMP
2011
15 years 1 months ago
2011
Abstract. In the 2nd Annual FOCS (1961), Chao-Kong Chow proved that every Boolean threshold function is uniquely determined by its degree-0 and degree-1 Fourier coefficients. These...
187
click to vote
CVPR
2005
IEEE
16 years 8 months ago
2005
IEEE
Our objective is to model the visual manifold of object appearance corresponding to geometric transformation. We learn a generative model for object appearance where the appearanc...
148
click to vote
ICML
2003
IEEE
16 years 7 months ago
2003
IEEE
When the training instances of the target class are heavily outnumbered by non-target training instances, SVMs can be ineffective in determining the class boundary. To remedy this...
145
click to vote
AHS
2007
IEEE
16 years 1 months ago
2007
IEEE
Machine Learning algorithms allow to create highly adaptable systems, since their functionality only depends on the features of the inputs and the coefficients found during the tr...