Model based learning systems usually face to a problem of forgetting as a result of the incremental learning of new instances. Normally, the systems have to re-learn past instances...
We introduce an algorithm for learning a local metric to a continuous input space that measures distances in terms of relevance to the processing task. The relevance is defined a...
The self-organizing map SOM is widely used as a data visualization method in various engineering applications. It performs a non-linear mapping from a high-dimensional data spac...
Multilayer perceptrons are often trained using error backpropagation (BP). BP training can be done in either a batch or continuous manner. Claims have frequently been made that bat...
We describe a deterministic model of packet delay and use it to derive both the packet pair [2] property of FIFO-queueing networks and a new technique (packet tailgating) for acti...