2011-03-05 #2

Adaptron Test Run – 2 - Date: 5th Mar 2011

After modifying the attention algorithm I ended up with this scenario for test case #41. The sequences on a single sense continue to grow even though it has paid attention to a larger P-habit combination. This is because they are all still sequentially dependent stimuli.

Does this make sense? In the comments on run-1 I had expected the STMs on the other senses to throw away their accumulated sequences when attention switched to a different sense of P-Habit. Is this realistic? If one is concentrating on a sequence on one sense and another starts on a second sense does that terminate the one on the 1st sense? It would seem that the sequence on the 1st sense is independent of the second sense because it has changed on the 1st sense while no change has occurred on the second sense. Thus the 1st sense sequence should end, stay sequentially dependent and a new sequence start incorporating both senses. This would still give the same stimuli pattern as in this test run up to input #5.

After #5 the question becomes does the same approach apply? When a sequence of several senses has been attended to and one sense becomes dormant (no changes) do the sequences accumulated so far on the other senses get thrown away? If this is to remain symmetric the answer is yes. Also the sense that goes from changing to dormant is a change just as much as if it had been dormant and started changing.