A Batch Problem is the equivalent of a problem in NeuroShell 2. In the Batch Processor, though, a batch problem is a row entry in the Batch Grid. There are two different kinds of batch problems: base problems and alias problems. The base problem will always entered in the first row in the Batch Grid. It is from this base problem that alias problems are built. After the base problem, all alias problems will appear in subsequent rows.
For example, suppose you are trying to predict the weather and your base problem is using a three-layer feed forward Backpropagation network. You would like to find out if a recurrent network would yield better results. In the Batch Processor, you would set up an alias problem that looks exactly like the base problem, except that it uses a recurrent network.