An item in the current TRN issue about computing via visual perception has me wondering if the method could be adapted to take advantage of one of the great strengths of the human brain: pattern recognition. Is it possible to pose a problem mathematically so that the correct answer has a discernible pattern when all of the possible answers are represented visually?

Imagine scanning through a three-dimensional cloud of random dots, lines or shapes — the solution space in math jargon — until you come to a part of the cloud that has a pattern. The coordinates of the pattern within the cloud would map to the mathematical representation of the solution.

Today researchers struggle with translating complicated pattern recognition problems into the stepwise logic of computers in order to give the machines humanlike vision and language understanding. Perhaps someday researchers will work on translating numbercrunching problems into the perceptual logic of the human brain.

The news item also reminded me of a distopian science fiction story that horrified me as a kid. A captive blind girl was periodically brought to a room where a machine plugged itself into her eye sockets. If I remember right, she was forced to react to or manipulate blurry shapes, and both the interface and the process were painful. The suggestion was that humans, or at least some humans, had literally become cogs in the machine, or what today we might call human coprocessors.

Can anyone tell me the author and title of that story? I haven’t been able to track it down. Let me know at eric [at] trnmag.com.