A probabilistic deformable model for the representation of brain structures is described. The statistically learned deformable model represents the relative location of head (skull...
The mammalian visual system is one of the most intensively investigated sensory systems. However, our knowledge of the typical input it is operating on is surprisingly limited. To ...
Abstract. The Gram matrix plays a central role in many kernel methods. Knowledge about the distribution of eigenvalues of the Gram matrix is useful for developing appropriate model...
This paper proposes a semi-supervised boosting approach to improve statistical word alignment with limited labeled data and large amounts of unlabeled data. The proposed approach ...
Results of neural network learning are always subject to some variability, due to the sensitivity to initial conditions, to convergence to local minima, and, sometimes more dramat...