Seeing
ain’t easy. About one-third of your brain is devoted to vision. And that
complexity makes vision one of the thorniest problems in artificial intelligence.
Computer scientist Rob Fergus and his colleagues aim to mimic or even surpass
the capabilities of the human brain. But getting a computer to correctly
identify a visual object (a cat, for instance) was nearly impossible until
recently.
A technique
called deep learning has radically improved the ability of devices to
recognize images and interpret speech. Deep learning uses computerized neural
networks to process information, and their performance at visual recognition
is rapidly approaching that of humans.
Where is
machine vision taking us? The Secret Science Club is ready to leap into the future.
Dr. Fergus asks: What if you could tell
your computer to use its “eyes” to organize your photo collection? Can space
telescopes be trained to look for Earth-like planets? Are there medical
conditions that can be screened for or diagnosed with machine vision? Could
your personal robot valet’s eyes get so good that he never gives you mismatched
socks ever again?
Rob Fergus is associate professor of computer
science at NYU’s Courant Institute
and a research scientist at Facebook’s
new Artificial Intelligence lab.
Before & After
--Try our pixelated cocktail of the night, the Machine Dream!
--Do the
robot to data-driven grooves
--Stick
around for the wet-wired Q&A
This open source edition of the Secret Science Club meets Thursday, September 11, 8 pm @ the Bell
House, 149 7th St.
(between 2nd and 3rd avenues) in Gowanus, Brooklyn .
Subway: F or G to 4th Ave.
Doors open at 7:30 pm. Please bring
ID: 21+. No cover. Just bring your smart self!
PLUS! Check out these three special SSC events coming up at Symphony Space!
No comments:
Post a Comment