Individual Neurons - Computational Explorations in Cognitive Neuroscience - page 54

Object Actions Generate .views .event_specs .events View: Object Actions

Disp AutoSc X: −−− Y: −−−

Edit Specs
Val Txt: NONE

COLOR

Disp Md:

Evt Label: Event::name
Pat Label: NONE
Pat Block: Pattern Values

8 9

New Event Delete

Apply Init

4 5 6 7

Events

[0] 0
[1] 1
[2] 2

0 1 2 3

[3] 3
[4] 4

−1 0 0.5 1

−1.000 1.000

Figure 2.19: The EnviroView display of the input events for the detector network.

showed an activity value of 0 for both inputs, meaning
that it was not activated above threshold by these in-
put patterns. Before getting into the nitty-gritty of why
the unit responded this way, let's proceed through the
remaining digits and observe how it responds to other
inputs.

Select the act button in the lower left hand part
of the window (you will probably have to scroll back up,
because act is near the top).

Press the Step button in the control panel to pro-
cess the first event in the environment (the digit 0). You
will not actually see anything happen. Press Step one
more time, and you will see the next digit (1) presented.
Each Step causes the input units to have their acti-
vationvaluesfixedor clamped to the values given in the
input pattern of the event, followed by a settling process
where the activation of the receiving unit is iteratively
updated over a series of cycles according to the point
neuron activation function (just as the unit in the pre-
vious simulation was updated over time). This settling
process continues until the activations in the network
approach an equilibrium (i.e., the change in activation
from one cycle to the next, shown as variable da in the
simulator, is below some tolerance level). The network
view is then updated at the end of settling so that we
may view this equilibrium activation state that resulted
from processing the input.
You should have seen the input pattern of the digits
0 and 1 in the input layer. However, the receiving unit

Continue to press Step until the digit 9 has been
presented.
You should have seen the receiving unit activated
when the digit 8 was presented, with an activation of
zero for all the other digits. Thus, as expected, the re-
ceiving unit acts like an “8” detector.
We can use a GraphLog to view the pattern of receiv-
ing unit activation across the different input patterns.

Press View on the control panel and select
GraphLog .
The graph shows the activation for the unit as a func-
tion of event (and digit) number along the X axis. You
should see a flat line with a single peak at 8.
Now, let's try to understand exactly why the unit re-
sponds as it does. The key to doing so is to understand
the relationship between the pattern of weights and the
input pattern. Thus, we will configure the NetView to