Vision Opportunities

CAs cross area boundaries (Pulvermuler)
(Being provacative) the reason that VIP has problem with object recognition, is that you are working on the wrong level
By using recurrent networks, you will be able to get a handle on more sophisticated processing.
From the CA perspective neurons act as parts of memories.
From standard visual processing, they act as transformers.
Can we combine these. How?
Similarly, I need to hook CAs into the real world. Can you help here?