Connect with us

Published

on

Load up the website This Person Does Not Exist and it’ll show you a human face, near-perfect in its realism yet totally fake. Refresh and the neural network behind the site will generate another, and another, and another. The endless sequence of AI-crafted faces is produced by a generative adversarial network (GAN)—a type of AI that learns to produce realistic but fake examples of the data it is trained on. 

But such generated faces—which are starting to be used in CGI movies and ads—might not be as unique as they seem. In a paper titled This Person (Probably) Exists, researchers show that many faces produced by GANs bear a striking resemblance to actual people who appear in the training data. The fake faces can effectively unmask the real faces the GAN was trained on, making it possible to expose the identity of those individuals. The work is the latest in a string of studies that call into doubt the popular idea of an AI model as a “black box” that reveals nothing about what’s going on inside.

To reveal the hidden training data, Ryan Webster and his colleagues at the University of Caen Normandy in France used a type of attack called a membership attack, which can be used to find out whether certain data was used to train an AI model. These attacks typically take advantage of subtle differences between the way a model treats data it was trained on—and has thus seen thousands of times before —and unseen data. For example, a model might identify a previously unseen image accurately, but with slightly less confidence, than one it was trained on. A second, attacking model can learn to spot such tells in the first model’s behavior and use them to predict when certain data, such as a photo, is in the training set or not. 

Such attacks can lead to serious security leaks. For example, finding out that someone’s medical data was used to train a model associated with a disease might reveal that this person has that disease.

Webster’s team extended this idea so that instead of identifying the exact photos used to train a GAN, they identified photos in the GAN’s training set that were not identical but appeared to portray the same individual—in other words, faces with the same identity. To do this, the researchers first generated faces with the GAN and then used a separate facial-recognition AI to detect whether the identity of these generated faces matched the identity of any of the faces seen in the training data.

The results are striking. In many cases, the team found multiple photos of real people in the training data that appeared to match the fake faces generated by the GAN, revealing the identity of individuals the AI had been trained on.

The left-hand column in each block shows faces generated by a GAN. These fake faces are followed by three photos of real people identified in the training data

The work raises some serious privacy concerns. “The AI community has a misleading sense of security when sharing trained deep neural network models,” says Jan Kautz, vice president of learning and perception research at Nvidia. 

In theory this kind of attack could apply to other data tied to an individual, such as biometric or medical data. On the other hand, Webster points out that the technique could also be used by people to check if their data has been used to train an AI without their consent.

An artist could check if their work had been used to train a GAN in a commercial tool, he says: “You could use a method such as ours for evidence of copyright infringement.”

The process could also be used to make sure GANs don’t expose private data in the first place. The GAN could check if its creations resembled real examples in its training data, using the same technique developed by the researchers, before releasing them.

Yet this assumes that you can get hold of that training data, says Kautz. He and his colleagues at Nvidia have come up with a different way to expose private data, including images of faces and other objects, medical data and more, that does not require access to training data at all.

Instead, they developed an algorithm that can recreate the data that a trained model has been exposed to by reversing the steps that the model goes through when processing that data. Take a trained image-recognition network: to identify what’s in an

Read More

————

By: Will Douglas Heaven
Title: AI fake-face generators can be rewound to reveal the real faces they trained on
Sourced From: www.technologyreview.com/2021/10/12/1036844/ai-gan-fake-faces-data-privacy-security-leak/
Published Date: Tue, 12 Oct 2021 09:15:23 +0000

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Tech

LATAM crypto exchange Bitso and FMF launch NFT of Mexico’s National Team jerseys

Published

on

Bitso, a leading cryptocurrency platform operating in Latin America, and the Mexican Football Federation (FMF), today announced the joint launch of the first collectible NFT of the Mexico National Team’s jerseys that was acquired in cryptocurrencies.

This morning through their social media platforms, the FMF and Bitso announced the opportunity to acquire the new official National Team fan jerseys ahead of the team’s participation in the 2022 World Cup. In just 20 minutes, the entire collection sold out.

bitsonftjeysey
The NFTs of the jerseys have an exclusive design for the metaverse – each is unique on the blockchain and can be resold by its owner in subsequent transactions.

The collection consisted of 100 official physical jerseys, each with a corresponding NFT version of the jersey that fans’ avatars can wear within the Decentraland metaverse. Each physical and NFT jersey set sold for the equivalent of $1,800 MXN in ethers.

“Our mission is to make cryptocurrency useful in the everyday life of Mexicans; we are committed to spreading the technology through innovative opportunities that help people throughout the country familiarize themselves with this new world. We are very excited to offer the incredible, historic opportunity for the fans of our National Team so that through their Bitso account, they can wear the colors of the National Team on and ‘off’ the field in the metaverse.”
– Bárbara González Briseño, General Director of Bitso México

Jersey NFTs

Created by Bitso, the virtual jersey sports the official colors of Mexico and the new National Team shield, characteristics that will make it stand out when users wear it in the virtual world of Decentraland.

The post LATAM crypto exchange Bitso and FMF launch NFT of Mexico’s National Team jerseys appeared first on CryptoNinjas.

Read More

————

By: CryptoNinjas.net
Title: LATAM crypto exchange Bitso and FMF launch NFT of Mexico’s National Team jerseys
Sourced From: www.cryptoninjas.net/2022/07/29/latam-crypto-exchange-bitso-and-fmf-launch-nft-of-mexicos-national-team-jerseys/
Published Date: Fri, 29 Jul 2022 15:19:02 +0000

Did you miss our previous article…
https://www.mansbrand.com/the-download-a-breakthrough-climate-bill-and-twitters-terrible-trends/

Continue Reading

Tech

Long-running crypto exchange EXMO unveils “lively” rebrand amidst growth

Published

on

EXMO, a crypto exchange platform operating since 2014, announced this week a rebranded visual identity with includes a new logo, brand colors, and design features. This new branding comes as EXMO continues to grow its crypto platform while also seeking to expand its presence in other jurisdictions.

Some new developments underway at EXMO:

Soon, users will be able to earn passive income from EXMO’s new staking platform.Plans to launch an EXMO crypto debit card.Expansion of its services in international markets with the opening of offices in Poland and Lithuania.

EXMO’s new logo

The rationale for the re-brand:

“At EXMO, we have a vision of a world where crypto is in every wallet. Hassle-free. We want to achieve this by making crypto as simple and accessible to everyone as possible. And we know that you already appreciate EXMO for offering user-friendly services and helpful support. Also for the opportunity to trade anywhere and anytime, closing deals in just a few taps. Such important changes required a rethinking of our corporate style, which has long needed a massive upgrade. So today we are introducing a new brand identity for EXMO with a completely new visual concept. We are launching a new logo, brand colors, and design elements. Our key design principles are simplicity, boldness, and a pinch of fun. But most importantly, we have changed our logo. Simple and easily recognizable, it represents the humanity of our brand. The logo stands out due to the wavy letter ‘m’ which symbolizes exchange rate charts and also resembles a spring that will launch you into the crypto world.”
– The EXMO Team regarding the re-branding

The post Long-running crypto exchange EXMO unveils “lively” rebrand amidst growth appeared first on CryptoNinjas.

Read More

————

By: CryptoNinjas.net
Title: Long-running crypto exchange EXMO unveils “lively” rebrand amidst growth
Sourced From: www.cryptoninjas.net/2022/07/26/long-running-crypto-exchange-exmo-unveils-lively-rebrand-amidst-growth/
Published Date: Tue, 26 Jul 2022 08:10:38 +0000

Did you miss our previous article…
https://www.mansbrand.com/stitching-together-the-grid-will-save-lives-as-extreme-weather-worsens/

Continue Reading

Tech

Stitching together the grid will save lives as extreme weather worsens

Published

on

The blistering heat waves that set temperature records across much of the US in recent days have strained electricity systems, threatening to knock out power in vulnerable regions of the country. 

The electricity has largely stayed online so far this summer, but there have been scattered problems and close calls already. 

Heavy use of energy-sucking air-conditioners is the biggest problem. But intense heat can also reduce the output of power plants, blow transformers, and force power lines to sag. Severe droughts across large parts of the country have also significantly reduced the availability of hydroelectric power, according to the North American Electric Reliability Corporation (NERC). 

It’s unlikely to get better soon. A number of grid operators may struggle to meet peak summer demand, creating the risk of rolling blackouts, the NERC report notes.

The nation’s isolated and antiquated grids are in desperate need of upgrades to keep the lights, heat, and air-conditioning on in the midst of extreme weather events that climate change is making more common, severe, and dangerous. One clear way to ease many of these issues is to more tightly integrate the country’s regional grids, stitching them together with more long-range transmission lines. 

If electricity generated in one area can be more easily shared across much wider regions, power can simply flow to where it’s needed at those moments when customers crank up air-conditioners en masse, or when power plants or fuel supply lines fail amid soaring temperatures, wildfires, hurricanes, or other events, says Liza Reed, a research manager focused on transmission at the Niskanen Center, a Washington, DC, think tank.  

The problem is it’s proved difficult to build more long-range transmission and grid interconnections for a variety of reasons, including the permitting challenges of erecting wires through private and public lands across cities, counties, and states and the reluctance of local authorities to forfeit control or submit to greater federal oversight.

The case of Texas

The unreliability of the US grid is not a new problem. Severe heat and winter storms have repeatedly exposed the frailty of electricity systems in recent years, leaving thousands to millions of people without power as temperatures spiked or plunged.

One of the fundamental challenges is that the grids today are highly fragmented. There are three main electricity networks within the US: the Eastern Grid, the Western Grid, and the Electric Reliability Council of Texas (ERCOT). But there are numerous regional transmission organizations within those first two systems, including the California Independent System Operator, Southwest Power Pool, PJM Interconnection, New York ISO, and more. 

These grids form a complex web of networks operating under different regulators, rules and market structures, and often with limited connections between them.

Map USA grid
A variety of regional transmission organizations oversee different parts of the nation’s aging and fragmented grids, which operate under different rules and with often limited connections between them.
legend

ERCOT is especially isolated, in part because of the desire among local politicians, citizens, and power companies to avoid added competition, the hassle of following other states’ rules, and oversight from the Federal Energy Regulatory Commission (FERC). But the state offers a case study in why that can be a serious problem amid increasingly harsh climate conditions, Reed says.

The Texas grid operator pleaded with customers several times earlier this month to cut electricity use as blistering summer temperatures created  demand surges that threatened to outstrip supply and require rolling blackouts.  Low wind conditions, cloud cover, and outages at fossil-fuel power plants added to the strains.

Shutting off the electricity needed to run air-conditioning in triple-digit temperatures

Read More

————

By: James Temple
Title: Stitching together the grid will save lives as extreme weather worsens
Sourced From: www.technologyreview.com/2022/07/28/1056483/stitching-together-the-grid-will-save-lives-as-extreme-weather-worsens/
Published Date: Thu, 28 Jul 2022 08:00:00 +0000

Continue Reading

Trending