A major difficulty in building Bayesian network models is the size of conditional probability tables, which grow exponentially in the number of parents. One way of dealing with th...
: A novel approach to artificial neural networks is presented. The philosophy of this approach is based on two aspects: the design of task-specific networks, and a new neuron model...
Benedito Dias Baptista F. Filho, Eduardo Lobo Lust...
We address the well-known problem of determining the capacity of constrained coding systems. While the onedimensional case is well understood to the extent that there are technique...
Selectiveattentionis a mechanismsused to sequentiallyselectthe spatiallocationsof salientregionsin the sensor’sfieldof view. This mechanism overcomesthe problem of flooding limi...
We describe a new approach for understanding the difficulty of designing efficient learning algorithms. We prove that the existence of an efficient learning algorithm for a circui...