Design Con 2015
Breaking News
Comments
Newest First | Oldest First | Threaded View
DrQuine
User Rank
CEO
re: Google neural network teaches itself to identify cats
DrQuine   7/1/2012 9:40:22 PM
NO RATINGS
Image recognition insights are often buried in the unreported details of detection thresholds, false positives and false negatives. First (regarding detection thresholds): how does the neuron's sensitivity compare with a human? Secondly (false positives): does the neuron ever fire on an image in which a human viewer cannot see the cat (and a cat is not believed to be present)? If so, are there any cues regarding what makes that image seem catlike? What mimics a cat and confuses the neuron? Finally (false negatives): does the neuron ever miss cats that seem to be within the normal detection sensitivity of the neuron? These instances may lead to some cues regarding the parameters that the neuron is depending upon that we don't use as heavily. These are the cats that are "camouflaged" from the neuron.

Sanjib.A
User Rank
CEO
re: Google neural network teaches itself to identify cats
Sanjib.A   6/29/2012 3:50:28 PM
NO RATINGS
It's really amazing to see the amount of efforts spent to do this...16000 processor cores and 1 billion connections! Just curious to know how much of money might have been spent, any idea?

markwrob
User Rank
Rookie
re: Google neural network teaches itself to identify cats
markwrob   6/28/2012 4:21:10 PM
NO RATINGS
I was also thinking of Skynet, but feel we don't have to worry until the Google network can *herd* cats. If that happens, it's game over for us carbon-based units (to borrow a phrase from another sci-fi dynasty). ;-) All kidding aside, this is an interesting development in cognitive computing and I expect we'll see many more to come from the likes of Google.

Peter Clarke
User Rank
Blogger
re: Google neural network teaches itself to identify cats
Peter Clarke   6/28/2012 11:23:43 AM
NO RATINGS
@Luis Sanchez My understanding is that the network does now know the word "cat." But if you show the network any image that YOU think is of a cat, one neuron, or pattern of neurons always fires in response. And it you show it an image of anything that YOU think is not of a cat, that neuron, or sub-network, does not respond. Nonetheless it is learning by itself, and like neuromorphic system. And the experiment could be extended to cross matching the images with spoken words.

Luis Sanchez
User Rank
Rookie
re: Google neural network teaches itself to identify cats
Luis Sanchez   6/28/2012 9:31:31 AM
NO RATINGS
Interesting! Learning by itself, what a cat is! It's like a baby learning and relating images with words. I suppose the neural system was able to "hear" the audio of the videos and relate "cat" to the images of cats? If no tags were used... how was the relation established? Seems indeed we're in the cognitive computing time.

Peter Clarke
User Rank
Blogger
re: Google neural network teaches itself to identify cats
Peter Clarke   6/27/2012 9:34:45 PM
NO RATINGS
I think Google make the point that most people have made networks with 10s of millions of connections while they have created a network support a billion connections. It may be that scale is important in neuromorphic systems and the good stuff only starts to happen when you get above 1 billion connections.

eewiz
User Rank
CEO
re: Google neural network teaches itself to identify cats
eewiz   6/27/2012 6:17:01 PM
NO RATINGS
Neural Networks trained using a set of test data identifies a pattern! There is nothing radically new here. Just that Google could afford to pay for a large 16KCPU server to support the large number of nueral connections.

wilber_xbox
User Rank
Manager
re: Google neural network teaches itself to identify cats
wilber_xbox   6/27/2012 2:48:04 PM
NO RATINGS
after watching a week of youtube videos if the artificial neuron could only recognize cats then either there are a lots of cat videos on youtube or they have superpowers.

prabhakar_deosthali
User Rank
CEO
re: Google neural network teaches itself to identify cats
prabhakar_deosthali   6/27/2012 1:49:01 PM
NO RATINGS
Finally we are there , able to model a fraction of the human brain!

dec66
User Rank
Rookie
re: Google neural network teaches itself to identify cats
dec66   6/27/2012 12:57:59 PM
NO RATINGS
It's official: Google is Skynet...



Flash Poll
Top Comments of the Week
Like Us on Facebook
EE Times on Twitter
EE Times Twitter Feed

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)
EE Life
Frankenstein's Fix, Teardowns, Sideshows, Design Contests, Reader Content & More
Max Maxfield

Book Review: Deadly Odds by Allen Wyler
Max Maxfield
8 comments
Generally speaking, when it comes to settling down with a good book, I tend to gravitate towards science fiction and science fantasy. Having said this, I do spend a lot of time reading ...

Martin Rowe

No 2014 Punkin Chunkin, What Will You Do?
Martin Rowe
Post a comment
American Thanksgiving is next week, and while some people watch (American) football all day, the real competition on TV has become Punkin Chunkin. But there will be no Punkin Chunkin on TV ...

Rich Quinnell

Making the Grade in Industrial Design
Rich Quinnell
13 comments
As every developer knows, there are the paper specifications for a product design, and then there are the real requirements. The paper specs are dry, bland, and rigidly numeric, making ...

Martin Rowe

Book Review: Controlling Radiated Emissions by Design
Martin Rowe
1 Comment
Controlling Radiated Emissions by Design, Third Edition, by Michel Mardiguian. Contributions by Donald L. Sweeney and Roger Swanberg. List price: $89.99 (e-book), $119 (hardcover).