4 ms·
If I had to guess: neural networks have to operate on pixel data whereas real neurons don't. Brains and eyes have evolved in tandem. Perhaps what makes them so
by chongli 10y ago
If I had to guess: neural networks have to operate on pixel data whereas real neurons don't. Brains and eyes have evolved in tandem. Perhaps what makes them so efficient is that the eyes handle some of the processing as a consequence of their physical shape and characteristics.
Look at the eyes of bees. Very different from our own (and from the cameras we build) and perhaps very specialized to the limited set of tasks that bees carry out?
- saosebastiao 10y agoThat's much of a topology thing...our brain may be a CPU, but we have neurons throughout much of our body, processing extremely low level sensory inputs into richer information that brains can process at a higher level. And different species have very different neural topologies: I've heard it described that Octopoda act more like an eight-member swarm intelligences than an intelligent eight-armed creature. Because their neural density is much more diffuse in their arms, they do so much more processing there, and the central brain acts more like a coordination unit than anything.
- lacampbell 10y agoMonolithic brain or modular microbrains...it's the Torvalds/Tannenbaum debate all over again!
- throwitawayday 10y agoOn eyes in general: the retina is not like a camera pixel array, rather, it extracts all kinds of features and detects motion: "Eye smarter than scientists believed: neural computations in circuits of the retina." Gollisch, Tim, and Markus Meister. Neuron 65.2 (2010): 150-164. http://www.sciencedirect.com/science/article/pii/S0896627309009994 http://www.sciencedirect.com/science/article/pii/S0896627309...
- davidwihl 10y agoThe neurons in the eyes pre-process a huge amount of information converting data from about 126M rods and cones to 1M ganglion cells - a pretty amazing amount of data compression. Ref: https://www.amazon.com/Visual-Thinking-Kaufmann-Interactive-Technologies/dp/0123708966 https://www.amazon.com/Visual-Thinking-Kaufmann-Interactive-...
- cr0sh 10y agoFrom my limited understanding, CNNs seem to approximate this functionality in image processing tasks, where the convolution layers are akin to the retina extracting info (lines, shapes, etc) and compression (kernel striding and pooling), while the later fully-connected neural network works for categorization or whatnot. That probably doesn't answer the question of how a bee's eyes work, though. ???
- lawpoop 10y agoI guess that's true for mammalian eyes, but what about insect eyes? If the bumble brain only hold 200k neurons, how many do their eyes have on board?
- tom_wilde 10y ago~7000 elements per compound eye according to this page: https://www.google.co.uk/amp/s/brookfieldfarmhoney.wordpress.com/2013/01/30/the-eyes-have-it-honeybee-eyes/amp/ https://www.google.co.uk/amp/s/brookfieldfarmhoney.wordpress...