Kottke points to a great article by Columbia's Duncan Watts...
The social context of content has everything to do with it's meaning. It's one of the reasons that I think that a purely pixel-based algorithmic approach to, say, image recognition is doomed. In optimistic moments, I've said that the computer vision community may produce a 98% reliable dog detector... But what we really want is a "funny" detector... or "cool" detector... that's gonna be a long-time coming... or maybe it's already here but involves analyzing people's actions around the pixels v. just the pixels in isolation.
Subscribe to:
Post Comments (Atom)
1 comment:
"computer vision community may produce a 98% reliable dog detector"
Your definition of 'funny' is subjective and isn't necessarily mine, I'd be happy with the dog detector.
"but involves analyzing people’s actions around the pixels v. just the pixels in isolation"
And people get their 'doggy' reaction from seeing the picture. I'm searching for a dog and you give me what *other* people say is a dog. I think that's a hearsay search. Why would you give me a result of what people think is a dog if you could tell if its a dog yourself?
"I think that a purely pixel-based algorithmic approach to, say, image recognition is doomed."
How can a baby learn that a dog is a dog? It has no pre-knowledge of a dog, yet it learns to identify and distinguish dogs in it's vision.
Even deaf babies born with only 1 function eye can do that, even deaf babies born with only 1 function eye and no sense of smell can recognize the same dog in a *static* picture.
It follows there is enough information there to identify the dog. Even in the static picture.
Those RGB photo-detectors on your retina are just pixels, can you recognize images of dogs?
Post a Comment