Even Artificial Neural Networks Can Have Exploitable ‘Backdoors’

August 25, 2017

wired1

Malicious actors can design that behavior to emerge only in response to a very specific, secret signal, as in the case of Garg’s Post-it. Such “backdoors” could be a problem for companies that want to outsource work on neural networks to third parties, or build products on top of freely available neural networks available online. Both approaches have become more common as interest in machine learning grows inside and outside the tech industry. “In general it seems that no one is thinking about this issue,” says Brendan Dolan-Gavitt, an NYU professor who worked with Garg.

Tricking The Eye To Defeat Shoulder Surfing Attacks

August 22, 2017

Tandon

Every ATM or smartphone user can attest to the discomfort of having a stranger standing close enough to observe a financial transaction — and potentially note a PIN or account number. Now researchers at the NYU Tandon School of Engineering have announced a first-of-its-kind application to combat such “shoulder-surfing,” whether in person or via a building’s video camera.

The 5 Coolest Things on Earth This Week

August 25, 2017

“IllusionPIN,” a new technology developed by researchers at New York University’s Tandon School of Engineering, can automatically shield the screen of an ATM, smartphone or other electronic device. While the user is able to clearly see the information displayed on the device, anyone more than a few feet away sees something completely different. … “Our goal was to increase the resilience of PIN authentication without straining the device or compromising user experience,” says Nasir Memon, a member of the research team.

New York University Abu Dhabi Researchers Develop ‘Unhackable’ Computer Chip

August 20, 2017

Researchers at New York University Abu Dhabi said they have created an ‘unhackable’ chip to shore up the defences of computer hardware, in an age of increasing threats to individuals and companies across the globe. … Ozgur Sinanoglu, NYUAD’s associate dean of engineering for academic affairs and head of the university’s Design for Excellence lab, has previously said malicious computer components – so-called Trojans – could be physically installed in factories or manufacturing labs.

Researchers Built an Invisible Backdoor to Hack AI’s Decisions

August 24, 2017

A team of NYU researchers has discovered a way to manipulate the artificial intelligence that powers self-driving cars and image recognition by installing a secret backdoor into the software…“We saw that people were increasingly outsourcing the training of these networks, and it kind of set off alarm bells for us,” Brendan Dolan-Gavitt, a professor at NYU, wrote to Quartz. “Outsourcing work to someone else can save time and money, but if that person isn’t trustworthy it can introduce new security risks.”

New Research Could Help Reveal Who is Buying Online Trafficking Ads

August 17, 2017

A U.S. researcher says she has developed automated ways to identify links between online sex trafficking ads and the digital currency Bitcoin, techniques that may help locate children being sold for sex. … Having automated style and time stamp analyses to identify sex ads by authors and Bitcoin owners is significant, said Damon McCoy, a New York University Tandon School of Engineering assistant professor of computer science and engineering and a co-author of the research.