We all know how traffic lights work.

If we would assign numbers to the activation sequence it would look like
this:
1, 2, 3, 1, 2, 3, 1, 2, ... where 1 = Green, 2 = Orange and 3 = Red

According to CLA, predictions are purely mathematical and derive from
simple probability algorithms.
Which means that the temporal connections between cells don't count for
predictions.

In the traffic lights example when Orange lights up (2) given a sequence
of: 1, 2, 3, 1, 2, 3, 1, 2 ...
the prediction union is:

1 - Green - 3 / 8= 37,5%
2 - Orange - 3 / 8 = 37,5%
3 - Red - 2 / 8 = 25%

So, after Orange the most likely to light up next  is either Green or
Orange again which will result in
the entire human species gone extinct by car accidents.

Why are we still alive?

Is it because in reality traffic lights are being perceived differently
(after all they are not simple sequences
but tools of man that we learned how they work) or is it simply because I
have misinterpreted how CLA
makes predictions?

Can we understand traffic lights without being told that it's a repeating
sequence of 1, 2, 3, ... ?
_______________________________________________
nupic mailing list
[email protected]
http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org

Reply via email to