Breaking News
News & Analysis

Neural nets make a comeback at Darpa

8/10/2009 08:00 PM EDT
1 Comment
NO RATINGS
More Related Links
View Comments: Oldest First | Newest First | Threaded View
RidgeRat
User Rank
Rookie
re: Neural nets make a comeback at Darpa
RidgeRat   8/14/2009 11:16:37 PM
NO RATINGS
Let's wish DARPA luck this round... The problem with neural networks is not so much what they can do as what they will do... which is learn the information provided to them, not what you want them to learn. Should I repeat a classic and probably apocryphal story? Oh heck, why not. The goal was to train a self-guided vehicle to navigate along a road. Shouldn't be hard, on each side there is a curb, with nice green grass beyond that, just stay between the lines. Only it turned out that when they took the vehicle outside to prepare for a demo, the light spectrum from REAL grass was nowhere close to pure green used in the laboratory to train the neural navigation logic. Successful demo or lose funding... what to do? Well, to make a long story short, they went out and spray painted all the grass green...

Top Comments of the Week
August Cartoon Caption Winner!
August Cartoon Caption Winner!
"All the King's horses and all the KIng's men gave up on Humpty, so they handed the problem off to Engineering."
5 comments
Like Us on Facebook

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)
EE Times on Twitter
EE Times Twitter Feed
Radio
LATEST ARCHIVED BROADCAST
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.
Flash Poll