Group items tagged

"He never got the chance to find out. A week later, he opened up his digital "wallet," where the artwork would supposedly be available, and was faced with an ominous banner reading, "This page has gone off grid. We've got a 404 error and explored deep and wide, but we can't find the page you're looking for."  The artwork, which he expected to be on the page, had disappeared entirely. "There was no history of my ever purchasing it, or ever owning it," he said. "Now there's nothing. My money's gone.""

"Learning from rewards seems like the simplest thing. I make coffee, I sip coffee, I'm happy. My brain registers "brewing coffee" as an action that leads to a reward. That's the guiding insight behind deep reinforcement learning, a family of algorithms that famously smashed most of Atari's gaming catalog and triumphed over humans in strategy games like Go. Here, an AI "agent" explores the game, trying out different actions and registering ones that let it win."

"It's fast getting to the point where it feels unreasonable to solely blame the famous and the industries that promote them. These days, people are going to plastic surgeons wanting to resemble their own modified avatars from selfies, rather than celebrities. If you like, the fiction of Hollywood perfection has been democratised. Indeed, it's interesting how, even as "improved" celebrities are mocked, or, as with Pike, call it out themselves, the modification of our own images continues unhindered, save for the occasional "#nofilter" humblebrag. It's gone beyond old-school catfishing (pretending to be someone else) to the point where people are essentially deep-faking themselves. And it's all just a bit of fun. Until it isn't. The desire to look better is all too human but are we inexorably moving towards the moment when we lose our grip on what we actually look like?"

"So when I discovered the pocket of Instagram where you can find out what it thinks you're interested in (on the app, you'll find it under Settings> Security> Access data > Ads), I obviously felt it my duty as a netizen to see what dark insights it had into my private soul. Here goes: jewellery; luxury goods; electronic music; love; emotions; fashion design; crafts. I mean: no offence, Kraftwerk (and loved ones) but I could not name eight things I am less interested in. Maybe oxbow lakes."

""This is the first time such machine learning tools have been used in this context," says Fluri, "and we found that the deep artificial neural network enables us to extract more information from the data than previous approaches. We believe that this usage of machine learning in cosmology will have many future applications.""

"Rather than being "taught" medical knowledge, the AI was asked to learn using unsupervised deep neural networks, known as autoencoders, without being given any medical knowledge. The researchers developed a method for translating the features found by the AI-only numbers initially-into high-resolution images that can be understood by humans."

"A New York Times deep-dive into a facial recognition AI tool sold to law enforcement agencies uncovered that the company has amassed more than three billion images. Those images are scraped from all corners of the internet from social media sites to companies' "About Us" pages.  That's way more than the typical police or even FBI database. "

"But now Silicon Valley seems to be under a little-noticed threat. Amid Covid-19, the deep recession, and renewed antitrust pressure from Congress and regulators, the Valley faces a very different challenge - the disruption of its very essence, the serendipitous encounter. The culprit is a rush by many of the Valley's leading companies to permanently lock in the coronavirus-led shift to remote work."

""How are we going to trust anything electronically mediated in the very near future? What do we do when anyone can imitate anyone else, for any reason that suits them?""

"What Hamilton and his colleagues found surprising was that this approach could also be applied to helping find problems with existing deep networks, related to the surge of "deepfakes" that have recently cropped up. They applied this data structure to find areas where probabilistic models, such as the generative adversarial networks (GANs) that are often used to create deepfakes, break down. They coined these problematic areas "blind spots," and note that they give us insight into how GANs can be biased. Such blind spots further show that GANs struggle to represent particular areas of a dataset, even if most of their fakes can fool a human. "

""The Vastaamo data breach is a shocking act which hits all of us deep down," the country's interior minister, Maria Ohisalo, wrote on her website on Monday. Finland must be a country where "help for mental health issues is available and it can be accessed without fear", she added. Ministers met for crisis talks this weekend, with further emergency discussions tabled for the coming week over the data breach."

"But those same services have also enabled the creation of what one professional factchecker calls a "perfect storm for misinformation". And with real-life interaction suppressed to counter the spread of the virus, it's easier than ever for people to fall deep down a rabbit hole of deception, where the endpoint may not simply be a decline in vaccination rates or the election of an unpleasant president, but the end of consensus reality as we know it. What happens when your basic understanding of the world is no longer the same as your neighbour's? And can Facebook stop that fate coming to us all?"

"Elon Musk wants Tesla to be seen as "much more than an electric car company." On Thursday's Tesla AI Day, the CEO described Tesla as a company with "deep AI activity in hardware on the inference level and on the training level" that can be used down the line for applications beyond self-driving cars, including a humanoid robot that Tesla is apparently building."

""Their position is that because Mr Musk is famous and might be more of a target for deep fakes, his public statements are immune," Pennypacker wrote, adding that such arguments would allow Musk and other famous people "to avoid taking ownership of what they did actually say and do"."

"There's a similar problem in artificial intelligence: The people who develop AI are increasingly having problems explaining how it works and determining why it has the outputs it has. Deep neural networks (DNN)-made up of layers and layers of processing systems trained on human-created data to mimic the neural networks of our brains-often seem to mirror not just human intelligence but also human inexplicability."

"Sources familiar with how AI-based systems have been integrated into the IDF's operations said such tools had significantly sped up the target creation process. "We prepare the targets automatically and work according to a checklist," a source who previously worked in the target division told +972/Local Call. "It really is like a factory. We work quickly and there is no time to delve deep into the target. The view is that we are judged according to how many targets we manage to generate." A separate source told the publication the Gospel had allowed the IDF to run a "mass assassination factory" in which the "emphasis is on quantity and not on quality". A human eye, they said, "will go over the targets before each attack, but it need not spend a lot of time on them". For some experts who research AI and international humanitarian law, an acceleration of this kind raises a number of concerns."

"What's fascinating about all this is how much of it comes down, not to finance or technology, but to people and what they know. In that sense the FT's deep dive into TSMC's travails reminded me of a striking piece of research conducted decades ago by the philosopher of science Harry Collins when he was a PhD student. Collins was interested in how knowledge gets transferred and intrigued by a particular piece of technology, the TEA laser. This was a device that was comprehensively documented in the physics literature but which research laboratories were unable to replicate. What Collins discovered was that "nobody could make the laser work if they hadn't spent time in a laboratory that already had a working laser. There was very good information in the journals about how to build such a laser. But anybody who tried to put one together using written articles failed. They had something that looked like a laser on their bench, but it wouldn't lase.""

""We erase your past" declares the company's tagline. Eliminalia, which has offices in several cities including Barcelona and Kyiv, is part of a growing industry that will clean up your online profile. Officially the company performs "a deep search across the internet for all information - whether it be an article, a blog, social media posts or even a mistaken identity". It then endeavours, on behalf of its clients, to get any negative information removed."

"What do today's AI-generated lesson plans look like? AI-generated lesson plans are already better than many people realise. Here's an example generated through the GPT-3 deep learning language model:"

"Renée: What I find most alarming is that people have the ability to just create reality by making something trend, to reinforce over and over and over again these conspiracy theories. You do have this increasingly divergent set of realities where there's a deep conviction built up over many, many years of reinforcing the same tropes and stories. You can't just correct that with a fact check."