What id really like to see htm do, is record labelled footage and be able to change the labels as it fed back, (as in maybe left right and strike?) that way you could get a game out of the memory system, with no code. Would be funny if it was real footage. :)
I imagine all it would take was a little motor input coming in with the eye, and these are the cells you hack during inference before every feedback, and it should twist the video to obey the motor?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
What id really like to see htm do, is record labelled footage and be able to change the labels as it fed back, (as in maybe left right and strike?) that way you could get a game out of the memory system, with no code. Would be funny if it was real footage. :)
I imagine all it would take was a little motor input coming in with the eye, and these are the cells you hack during inference before every feedback, and it should twist the video to obey the motor?