In many pattern recognition applications, high-dimensional feature vectors impose a high computational cost as well as the risk of "overfitting". Feature Selection addre...
A simple algorithm for surface interpolation is proposed. Its central idea is quadratic interpolation of a curved segment from the position and normal vectors at the end points, w...
Based on studies and experiments on the loss term of SVMs, we argue that 1-norm measurement is better than 2-norm measurement for outlier resistance. Thus, we modify the previous 2...
We address the possibility of overloaded vector precoding in single user MIMO channels, i.e. the number of data streams is larger than the minimum of the number of antennas at tran...
Beamformers are spatial filters that pass source signals in particular focused locations while suppressing interference from elsewhere. The widely-used minimum variance adaptive b...