The Retail Sector Is in Upheaval, and Women Are Taking the Brunt of the Job Losses

Computer Vision Algorithms Are Still Way Too Easy to Trick

AI image recognition has made some stunning advances, but as new research shows, the systems can still be tripped up by examples that would never fool a person.

Labsix, a group of MIT students who recently tricked an image classifier developed by Google… Read more

AI image recognition has made some stunning advances, but as new research shows, the systems can still be tripped up by examples that would never fool a person.

Labsix, a group of MIT students who recently tricked an image classifier developed by Google to think a 3-D-printed turtle was a rifle, released a paper on Wednesday that details a different technique that could fool systems even faster. This time, however, they managed to trick a “black box,” where they only had partial information on how the system was making decisions.

The team’s new algorithm starts with an image it wants to use to trick another system—in the example from their paper, it’s a dog—and then starts altering pixels to make the image look more like the source image; in this case, skiers. As it works, the “adversarial” algorithm challenges the image recognition system with versions of the picture that quickly move into territory any human would recognize as skiers (check out the gif, above). But all the while, the algorithm maintains just the right combination of sabotaged pixels to make the system think it’s looking at a dog.

The researchers tested their method on Google’s Cloud Vision API—a good test case in part because Google has not published anything about how the computer vision software works, or even all the labels the system uses to classify images. The team says that they’ve only tried foiling Google’s system so far, but that their technique should work on other image recognition systems as well.

There are plenty of researchers working on counteringadversarial examples like this, but for safety-critical uses, such as autonomous vehicles, artificial intelligence won’t be trusted until adversarial attacks are impossible, or at least much more difficult, to pull off.

Image credit:

Let's block ads! (Why?)



More artcles

"Women" im Kunst Museum Winterthur

Women’s Australian Open: Olivia Cowan unter den Top 20

Sokol/Post und Linz-Steg im AVL Women-Halbfinale – Aich/Dob schlägt Waldviertel in Volley League

Yoko Ono - so viel mehr als John Lennons Witwe

Künstlerin und Lennon-Witwe

Yoko Ono wird 85. Foto: Sashenka Gutiérrez

Women in Tech: „Zweifelt nicht an euren Fähigkeiten!“

New Bonn swimming pool to have curtain for Muslim women

New Bonn swimming pool to have curtain for Muslim women

Farrah Abraham shows ample cleavage wearing plunging dress in Las Vegas after kissing new boyfriend Simon Saran at toga party



Related links

Le Specs Women's Jealous Games Sunglasses, Streaky Tort/Brown Mono, One Size

$69.00

Le Specs Women's Jealous Games Sunglasses, Streaky Tort/Brown Mono,...

QUAY AUSTRALIA Women's Kitti Black/Smoke Lens Sunglasses

$50.00

QUAY AUSTRALIA Women's Kitti Black/Smoke Lens Sunglasses

Gerber Little Girls' Toddler Two-Piece Sundress and Hat Set, Sunglasses, 5T

$10.99

Gerber Little Girls' Toddler Two-Piece Sundress and Hat Set,...

Tifosi Women's Alpe 2.0 Wrap Sunglasses, Crystal Brown, 128 mm

$33.95

Tifosi Women's Alpe 2.0 Wrap Sunglasses, Crystal Brown, 128 mm

Pacific Coast Original KD's Biker Sunglasses (Black Frame/Green Lens)

$10.95

Pacific Coast Original KD's Biker Sunglasses (Black Frame/Green Lens)

Duduma Polarized Sports Sunglasses for Running Cycling Fishing Golf Tr90 Unbreakable Frame (blue/blue)

$43.98

Duduma Polarized Sports Sunglasses for Running Cycling Fishing Golf...

Related links

Brand New 2018 Valentino Women Sunglasses VA 2016 3005/87 Authentic Frame Italy

$292.73

Brand New 2018 Valentino Women Sunglasses VA 2016 3005/87 Authentic...