It is often important to be able to track the numerical value of a unit, weight, etc. One way to do this is with a value object. To create a value, select the value object in the Object Menu, select the object that you want to track and click on the workpace where you want the value to appear. Many objects have more than one numeric value associated with them. In these cases you must use the value object's parameter menu to indicate which of the parameters this value should track.
Exercise 5: Create a value for the word unit in order to track its value. Label the value CAT. Your network should now look like Figure 4.
Figure 4: Adding a value for the word unit.
Exercise 6: Click on the unit to change the activation of the word unit to 1.0. Notice that the value of the value also changes to 1.0 (see Figure 5).
Figure 5: The network with the word unit toggled to 1.
To create a weight, select the weight object from the Object Menu, select the units from which the weights should come and double select the units to which the weights should go. Then click on any unoccupied part of the workspace. A weight will be connected from each selected unit to each double selected unit. The strength of the weight is represented by the colour of the line. Red lines indicate positive weights and blue lines indicate negative weights. Arrow marks indicate the direction of the connection and whether or not a weight is frozen If a weight is frozen (indicated by pink arrow marks) it's value will not change during learning.
When a network has a large number of weights, it is convenient to be able to make the weights invisible, since they can obscure objects in the workspace (especially units). The default setting for BrainWave is weights visible. To make the weights invisible, unselect the weights visible item in the Options menu.
Exercise 7: Create a set of weights from the letter units to the word unit. Set the value of the weights to be 0.1. The small positive weights should turn red. Your network should now look like Figure 6. Practice making the weights invisible/visible.
Figure 6: Adding weights to the network.
[Next Section: Sets]
