Posted on Monday, 11th March 2013

Read KEB16.pdf through example 16.2 of Section 16.1.5, and then read KordingWolpert04.pdf and post a comment.

Posted in Class | Comments (3)

  1. amweinst Says:
    March 18th, 2013 at 11:55 am

    I was a little confused about model 3 so I was hoping that we could go over the models in a bit more detail in class. I understand that model 1 assumes that pointing variability, but not mean, would change with uncertainty and that model 2 assumes that the uncertainty information is combined with prior information in a Bayesian manner. I think that model 3 is modeling more of an error minimization of prior trials, but then why would this model not also include prior information as well, if it is including prior error information?

  2. dunovank Says:
    March 18th, 2013 at 5:55 pm

    In the last paragraph on pg. 245, the authors refer to estimating the average visual uncertainty experienced across subjects for each level of the imposed uncertainty (O, M, and L). I’m having difficulty determining how they’re calculating visual uncertainty separately from the imposed uncertainty. Is it just a simple contrast between the trials on which they got feedback and those that didn’t include visual feedback?

  3. rvistein Says:
    March 18th, 2013 at 6:08 pm

    I’m not comfortable with the logic in the paper. I will agree that of the models presented the Baysian matches the pattern best, but that does not exclude other models not tested here in. They rule out the other two systems clearly, by showing learning without the direct feedback required for mapping and that the deviation from the true offset does not average to zero but the epistemology of their conclusions seems too exclusive.

Leave a Reply

You must be logged in to post a comment.