changeset 549:c14545acdfba

minor edit.
author Robert McIntyre <rlm@mit.edu>
date Fri, 02 May 2014 03:34:13 -0400
parents 0b891e0dd809
children b1d8d9b4b569
files thesis/cortex.org
diffstat 1 files changed, 8 insertions(+), 7 deletions(-) [+]
line wrap: on
line diff
     1.1 --- a/thesis/cortex.org	Thu May 01 23:41:41 2014 -0400
     1.2 +++ b/thesis/cortex.org	Fri May 02 03:34:13 2014 -0400
     1.3 @@ -3209,13 +3209,14 @@
     1.4  
     1.5    After about 1 minute of manual training, I was able to achieve 95%
     1.6    accuracy on manual testing of the worm using =init-interactive= and
     1.7 -  =test-empathy-accuracy=. The majority of errors are near the
     1.8 -  boundaries of transitioning from one type of action to another.
     1.9 -  During these transitions the exact label for the action is more open
    1.10 -  to interpretation, and disagreement between empathy and experience
    1.11 -  is essentially irrelevant at this point, giving a practical
    1.12 -  identification accuracy of even higher than 95%. When I watch this
    1.13 -  system myself, I generally see no errors in action identification.
    1.14 +  =test-empathy-accuracy=. The majority of disagreements are near the
    1.15 +  transition boundaries from one type of action to another.  During
    1.16 +  these transitions the exact label for the action is often unclear,
    1.17 +  and disagreement between empathy and experience is practically
    1.18 +  irrelevant. Thus, the system's effective identification accuracy is
    1.19 +  even higher than 95%. When I watch this system myself, I generally
    1.20 +  see no errors in action identification compared to my own judgment
    1.21 +  of what the worm is doing.
    1.22  
    1.23  ** Digression: Learning touch sensor layout through free play
    1.24