We present a hierarchical architecture and learning algorithm for visual recognition and other visual inference tasks such as imagination, reconstruction of occluded images, and e...
Abstract. Compared to normal learning algorithms, for example backpropagation, the optimal bounded ellipsoid (OBE) algorithm has some better properties, such as faster convergence,...
A number of key areas in IP network engineering, management and surveillance greatly benefit from the ability to dynamically identify traffic flows according to the applications re...
Sebastian Zander, Thuy T. T. Nguyen, Grenville J. ...
— We study the cross-layer problem of combining routing and cooperative diversity in multi-hop, bandwidthconstrained networks with dedicated multiple access. Previous work in coo...
We suggest a recurrent neural network (RNN) model with a recurrent part corresponding to iterative function systems (IFS) introduced by Barnsley 1] as a fractal image compression ...