Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. Also, Slashdot's Facebook page has a chat bot now. Message it for stories and more. ×
Google

Google Releases an AI Tool For Publishers To Spot and Weed Out Toxic Comments (bbc.com) 177

Google today launched a new technology to help news organizations and online platforms identify and swiftly remove abusive comments on their websites. The technology, called Perspective, will review comments and score them based on how similar they are to comments people said were "toxic" or likely to make them leave a conversation. From a report on BBC: The search giant has developed something called Perspective, which it describes as a technology that uses machine learning to identify problematic comments. The software has been developed by Jigsaw, a division of Google with a mission to tackle online security dangers such as extremism and cyberbullying. The system learns by seeing how thousands of online conversations have been moderated and then scores new comments by assessing how "toxic" they are and whether similar language had led other people to leave conversations. What it's doing is trying to improve the quality of debate and make sure people aren't put off from joining in.
AI

Microsoft Research Developing An AI To Put Coders Out of a Job (mspoweruser.com) 301

jmcbain writes: Are you a software programmer who voted in a recent Slashdot poll that a robot/AI would never take your job? Unfortunately, you're wrong. Microsoft, in collaboration with the University of Cambridge, is developing such an AI. This software "can turn your descriptions into working code in seconds," reports MSPoweruser. "Called DeepCoder, the software can take requirements by the developer, search through a massive database of code snippets and deliver working code in seconds, a significant advance in the state of the art in program synthesis." New Scientist describes program synthesis as "creating new programs by piecing together lines of code taken from existing software -- just like a programmer might. Given a list of inputs and outputs for each code fragment, DeepCoder learned which pieces of code were needed to achieve the desired result overall." The original research paper can be read here.
AI

Tinder Wants AI To Set You Up On a Date (bbc.com) 133

Dave Lee, writing for BBC: Tinder is growing up. It's now a serious technology company tackling one of life's most important matters, and is by far the most popular dating app worldwide. After a lot of boardroom musical chairs, Mr Rad is the chairman of both Tinder and Swipe Ventures, the arm of the company designed to buy other dating-related technologies. One of which is artificial intelligence. And its collision with dating might be the most intriguing application of AI yet. "I think this might sound crazy," Mr Rad said on Tuesday at tech conference Start-Up Grind. "In five years time, Tinder might be so good, you might be like "Hey [Apple voice assistant] Siri, what's happening tonight?' "And Tinder might pop up and say 'There's someone down the street you might be attracted to. She's also attracted to you. She's free tomorrow night. We know you both like the same band, and it's playing -- would you like us to buy you tickets?'... and you have a match. "It's a little scary."
Transportation

College Senior Turns His Honda Civic Into a Self-Driving Car Using Free Hardware, Software (technologyreview.com) 128

holy_calamity writes: University of Nebraska student Brevan Jorgenson swapped the rear-view mirror in his 2016 Honda Civic for a home-built device called a Neo, which can steer the vehicle and follow traffic on the highway. Jorgenson used hardware designs and open-source software released by Comma, a self-driving car startup that decided to give away its technology for free last year after receiving a letter asking questions about its functionality from the National Highway Traffic Safety Administration (NHTSA). Jorgenson is just one person in a new hacker community trying to upgrade their cars using Comma's technology. "A Neo is built from a OnePlus 3 smartphone equipped with Comma's now-free Openpilot software, a circuit board that connects the device to the car's electronics, and a 3-D-printed case," reports MIT Technology Review. The report notes that Neodriven, a startup based in Los Angeles, has recently started selling a pre-built Neo device that works with Comma's Openpilot software, but it costs $1,495.
Transportation

Self-Driving Car Speed Race Ends With A Crash (electrek.co) 87

An anonymous reader writes:On a professional track in Buenos Aires, fans watched the first Formula E auto race with self-driving electric cars. "Roborace's two test vehicles battled it out on the circuit at a reasonably quick 115MPH," reports Engadget, "but one of the cars crashed after it took a turn too aggressively. The racing league was quick to tout the safety advantages of crashing autonomous cars ('no drivers were harmed'), but it's clear that the tech is still rough around the edges." Electrek is reporting that the cars "still have a cabin for a driver but neither car's cabin was occupied during the event." The ultimate goal is to have several teams racing the exact same self-driving car, while letting each team customize its car's driving software.
An Argentinian journalist shared footage of the race cars on Twitter, and apparently at one point a dog wandered out in front of an oncoming race car. But the real question is how the fans are going to feel about watching a speed race between cars with no drivers?
Microsoft

Bill Gates: The Robot That Takes Your Job Should Pay Taxes (qz.com) 387

In a recent interview with Quartz, Bill Gates said he believes that governments should tax companies that use robots who are taking human jobs, as a way to at least temporarily slow the spread of automation and to fund other types of employment. The money gained from taxing robots could then be used to finance jobs taking care of elderly people or working with kids in schools -- jobs which humans are particularly well suited for. Quartz reports: [Gates] argues that governments must oversee such programs rather than relying on businesses, in order to redirect the jobs to help people with lower incomes. The idea is not totally theoretical: EU lawmakers considered a proposal to tax robot owners to pay for training for workers who lose their jobs, though on Feb. 16 the legislators ultimately rejected it. "You ought to be willing to raise the tax level and even slow down the speed" of automation, Gates argues. That's because the technology and business cases for replacing humans in a wide range of jobs are arriving simultaneously, and it's important to be able to manage that displacement. "You cross the threshold of job replacement of certain activities all sort of at once," Gates says, citing warehouse work and driving as some of the job categories that in the next 20 years will have robots doing them. You can watch Gates' remarks in a video here, or read the transcript embedded in Quartz' report.
AI

Japan Unveils Next-Generation, Pascal-Based AI Supercomputer (nextplatform.com) 121

The Tokyo Institute of Technology has announced plans to launch Japan's "fastest AI supercomputer" this summer. The supercomputer is called Tsubame 3.0 and will use Nvidia's latest Pascal-based Tesla P100 GPU accelerators to double its performance over its predecessor, the Tsubame 2.5. Slashdot reader kipperstem77 shares an excerpt from a report via The Next Platform: With all of those CPUs and GPUs, Tsubame 3.0 will have 12.15 petaflops of peak double precision performance, and is rated at 24.3 petaflops single precision and, importantly, is rated at 47.2 petaflops at the half precision that is important for neural networks employed in deep learning applications. When added to the existing Tsubame 2.5 machine and the experimental immersion-cooled Tsubame-KFC system, TiTech will have a total of 6,720 GPUs to bring to bear on workloads, adding up to a total of 64.3 aggregate petaflops at half precision. (This is interesting to us because that means Nvidia has worked with TiTech to get half precision working on Kepler GPUs, which did not formally support half precision.)
AI

EU Moves To Bring In AI Laws, But Rejects Robot Tax Proposal (newatlas.com) 72

An anonymous reader quotes a report from New Atlas: The European Parliament has voted on a resolution to regulate the development of artificial intelligence and robotics across the European Union. Based on a raft of recommendations drafted in a report submitted in January to the legal affairs committee, the proposed rules include establishing ethical standards for the development of artificial intelligence, and introducing an insurance scheme to cover liability for accidents involving driverless cars. Not every element in the broad-ranging report was accepted by the Parliament though, with a recommendation to institute a "robot tax" roundly rejected. The robot tax proposal was designed to create a fund that manages the repercussions and retraining of workers made redundant through the increased deployment of industrial and service robots. But those in the robotics industry were supportive of the Parliamentary rejection, with the International Federation of Robotics suggesting to Reuters a robot tax would have been harmful to the burgeoning industry, stifling innovation and competitiveness. The European Parliament passed the resolution comfortably with 396 votes to 123, with 85 abstentions.
AI

Zuckerberg Shares Facebook's Plan to Bring Community Together, Edits Out a Questionable Sentence Minutes Later (mashable.com) 104

Facebook CEO and founder Mark Zuckerberg wants to bring people closer together. He published a 6,000-word letter on his Facebook page Thursday to outline his vision for the kind of world he thinks Facebook can help create. The free-wielding note included few specifics, but offered a number of broad, ambitious goals for how the tech giant can contribute to a better understanding of everything from terrorism to fake news. Interestingly, minutes after the post was published, Zuckerberg edited out a sentence from the letter. Mashable adds: In the post, Zuckerberg briefly touches on how artificial intelligence can be used to detect terrorist propaganda. "Right now, we're starting to explore ways to use AI to tell the difference between news stories about terrorism and actual terrorist propaganda so we can quickly remove anyone trying to use our services to recruit for a terrorist organization," he wrote in the post published Thursday. That sounds like a straightforward enough application of AI -- one that's in line with what Zuckerberg and other executives have discussed in the past -- but it's different from what the CEO had originally written. In an earlier version of the missive, which was shared with a number of news outlets in advance of its publication on Facebook, Zuckerberg took the idea farther. The "long-term promise of AI," he wrote, is that it can be used used to "identify risks that nobody would have flagged at all, including terrorists planning attacks using private channels." Here's an expanded version of the quote from the Associated Press (emphasis ours). "The long term promise of AI is that in addition to identifying risks more quickly and accurately than would have already happened, it may also identify risks that nobody would have flagged at all "including terrorists planning attacks using private channels, people bullying someone too afraid to report it themselves, and other issues both local and global. It will take many years to develop these systems." That's different from what was described in the final version that was shared Thursday, which made no mention of private communication in relation to AI and terrorism.
AI

Google Releases TensorFlow 1.0 With New Machine Learning Tools (venturebeat.com) 20

An anonymous reader shares a VentureBeat report: At Google's inaugural TensorFlow Dev Summit in Mountain View, California, today, Google announced the release of version 1.0 of its TensorFlow open source framework for deep learning, a trendy type of artificial intelligence. Google says the release is now production-ready by way of its application programing interface (API). But there are also new tools that will be part of the framework, which includes artificial neural networks that can be trained on data and can then make inferences about new data. Now there are more traditional machine learning tools, including K-means and support vector machines (SVMs), TensorFlow's engineering director, Rajat Monga, said at the conference. And there's an integration with the Python-based Keras library, which was originally meant to ease the use of the Theano deep learning framework. And there are now "canned estimators," or models, Monga said, including simple neural networks to start using quickly.
AI

AI Software Juggles Probabilities To Learn From Less Data (technologyreview.com) 49

moon_unit2 quotes a report from MIT Technology Review: You can, for instance, train a deep-learning algorithm to recognize a cat with a cat-fancier's level of expertise, but you'll need to feed it tens or even hundreds of thousands of images of felines, capturing a huge amount of variation in size, shape, texture, lighting, and orientation. It would be lot more efficient if, a bit like a person, an algorithm could develop an idea about what makes a cat a cat from fewer examples. A Boston-based startup called Gamalon has developed technology that lets computers do this in some situations, and it is releasing two products Tuesday based on the approach. Gamalon uses a technique that it calls Bayesian program synthesis to build algorithms capable of learning from fewer examples. Bayesian probability, named after the 18th century mathematician Thomas Bayes, provides a mathematical framework for refining predictions about the world based on experience. Gamalon's system uses probabilistic programming -- or code that deals in probabilities rather than specific variables -- to build a predictive model that explains a particular data set. From just a few examples, a probabilistic program can determine, for instance, that it's highly probable that cats have ears, whiskers, and tails. As further examples are provided, the code behind the model is rewritten, and the probabilities tweaked. This provides an efficient way to learn the salient knowledge from the data.
Google

Engineers On Google's Self-Driving Car Project Were Paid So Much That They Quit (theverge.com) 95

According to a new report from Bloomberg, most of the money Google spent on it self-driving car project, now spun off into a new entity called Waymo, has gone to engineers and other staff. While it has helped retain a lot of influential and dedicated workers in the short run, it has resulted in many staffers leaving the company in the long run due to the immense financial security. The Verge reports: Bloomberg says that early staffers "had an unusual compensation system" that multiplied staffers salaries and bonuses based on the performance of the self-driving project. The payments accumulated as milestones were reached, even though Waymo remains years away from generating revenue. One staffer eventually "had a multiplier of 16 applied to bonuses and equity amassed over four years." The huge amounts of compensation worked -- for a while. But eventually, it gave many staffers such financial security that they were willing to leave the cuddly confines of Google. Two staffers that Bloomberg spoke to called it "F-you money," and the accumulated cash allowed them to depart Google for other firms, including Chris Urmson who co-founded a startup with ex-Tesla employee Sterling Anderson, and others who founded a self-driving truck company called Otto which was purchased by Uber last year, and another who founded Argo AI which received a $1 billion investment from Ford last week.
Robotics

Elon Musk: Humans Need To Merge With Machines Else They Will Become Irrelevant in AI Age (cnbc.com) 251

Billionaire Elon Musk is known for his futuristic ideas. So it didn't come as a surprise when on Monday at the World Government Summit in Dubai, he predicted that over time we will see a "closer merger of biological intelligence and digital intelligence." He added, via a CNBC report: "It's mostly about the bandwidth, the speed of the connection between your brain and the digital version of yourself, particularly output." Musk explained what he meant by saying that computers can communicate at "a trillion bits per second", while humans, whose main communication method is typing with their fingers via a mobile device, can do about 10 bits per second. In an age when AI threatens to become widespread, humans would be useless, so there's a need to merge with machines, according to Musk. "Some high bandwidth interface to the brain will be something that helps achieve a symbiosis between human and machine intelligence and maybe solves the control problem and the usefulness problem," Musk explained.
Wikipedia

34 'Highly Toxic Users' Wrote 9% of the Personal Attacks On Wikipedia (bleepingcomputer.com) 174

Researchers used machine learning to analyze every single comment left on Wikipedia in 2015. An anonymous reader shares their results: 34 "highly toxic users" were responsible for 9% of all the personal attacks in the comments on Wikipedia, according to a research team from Alphabet's Jigsaw and the Wikimedia Foundation. They concluded that "significant progress could be made by moderating a relatively small number of frequent attackers." But at the same time, in Wikipedia's comments "less than half of attacks come from users with little prior participation; and perhaps surprisingly, approximately 30% of attacks come from registered users with over a 100 contributions. These results suggest the problems associated with personal attacks do not have an easy solution... the majority of personal attacks on Wikipedia are not the result of a few malicious users, nor primarily the consequence of allowing anonymous contributions."

The researchers "developed a machine learning algorithm that was able to identify and distinguish different forms of online abuse and personal attacks," reports Bleeping Computer, adding that the team "hopes that Wikipedia uses their study to build a comments monitoring dashboard that could track down hotspots of abusive personal attacks and help moderators ban or block toxic users." The paper describes it as a method "that combines crowdsourcing and machine learning to analyze personal attacks at scale."

Government

Face Recognition + Mandatory Police Body Cameras = Mass Surveillance? (siliconvalley.com) 110

Facial recognition software is already in use, and it has privacy advocates worried. An anonymous reader quotes the Bay Area Newsgroup. Southern California-based FaceFirst sells its facial recognition technology to retail stores, which use it to identify shoplifters who have been banned from the store, and alert management if they return. Corporate offices and banks also use the software to recognize people who are wanted by police... Several local law enforcement agencies have expressed interest in the technology, but so far none have had the budget for it. FaceFirst sells software police officers can install on their smartphones and use to identify people in the field from up to 12 feet away.

Some privacy experts worry facial recognition technology will show up next in police body cameras, with potentially dangerous consequences... The problem, say privacy advocates, is that all kinds of people come into contact with police, including many who are never suspected of any crimes. So lots of innocent people could be caught up in a police database fed by face-recognizing body cameras. The body cameras could turn into a "massive mobile surveillance network," said Jeramie Scott, national security counsel for the Electronic Privacy Information Center.

One-third of America's police departments use body cameras. (And just in San Jose, there's already 450 neighborhood cameras that have also agreed to share their footage for police investigations.) The new technologies concern the ACLU's policy director for technology and civil liberties. "You have very powerful systems being purchased, most often in secret, with little-to-no public debate and no process in place to make sure that there are policies in place to safeguard community members."
Transportation

Ford Just Invested $1 Billion In Self-Driving Cars (usatoday.com) 113

An anonymous reader quote USA Today: Ford Motor is betting $1 billion on the world's self-driving car future. The Detroit automaker announced Friday that it would allocate that sum over five years to a new autonomous car startup called Argo AI, which is headquartered in Pittsburgh, Pa., and will have offices in Michigan and California. Ford's financial outlay is part of a continuing investment strategy anchored to transforming the car and truck seller into a mobility company with a hand in ride-hailing, ride-sharing and even bicycle rentals.
Lucas123 writes: Argo AI founders CEO Bryan Salesky, and COO Peter Rander are alumni of Carnegie Mellon National Robotics Engineering Center and former leaders on the self-driving car teams of Google and Uber, respectively. Argo AI's team will include roboticists and engineers from inside and outside of Ford working to develop a new software platform for Ford's fully autonomous vehicle, expected in 2021. Ford said it could also license the software to other carmakers.
Microsoft

Microsoft Debuts Customizable Speech-To-Text Tech, Releases Some Cognitive Services Tools To Developers (geekwire.com) 23

Microsoft is readying three of its 25 Cognitive Services tools for wider release to developers. From a report on GeekWire: Microsoft's AI and Research Group, a major new engineering and research division formed last year inside the Redmond company, is debuting a new technology that lets developers customize Microsoft's speech-to-text engine for use in their own apps and online services. The new Custom Speech Service is set for release today as a public preview. Microsoft says it lets developers upload a unique vocabulary -- such as alien names in Human Interact's VR game Starship Commander -- to produce a sophisticated language model for recognizing voice commands and other speech from users. It's the latest in a series of "cognitive services" from Microsoft's Artificial Intelligence and Research Group, a 5,000-person division led by Microsoft Research chief Harry Shum. The company says it has expanded from four to 25 cognitive services in the last two years, including 19 in preview and six that are generally available.
AI

Are Gates, Musk Being 'Too Aggressive' With AI Concerns? (xconomy.com) 311

gthuang88 reports on a talk titled "Will Robots Eat Your Job?" Bill Gates and Elon Musk are sounding the alarm "too aggressively" over artificial intelligence's potential negative consequences for society, says MIT professor Erik Brynjolfsson. The co-author of The Second Machine Age argues it will take at least 30 to 50 years for robots and software to eliminate the need for human laborers. In the meantime, he says, we should be investing in education so that people are prepared for the jobs of the future, and are focused on where they still have an advantage over machines -- creativity, empathy, leadership, and teamwork.
The professor acknowledges "there are some legitimate concerns" about robots taking jobs away from humans, but "I don't think it's a problem we have to face today... It can be counterproductive to overestimate what machines can do right now." Eventually humankind will reach a world where robots do practically everything, the professor believes, but with a universal basic income this could simply leave us humans with more leisure time.
Security

Can The Mayhem AI Automate Bug-Patching? (technologyreview.com) 23

"Now when a machine is compromised it takes days or weeks for someone to notice and then days or weeks -- or never -- until a patch is put out," says Carnegie Mellon professor David Brumley. "Imagine a world where the first time a hacker exploits a vulnerability he can only exploit one machine and then it's patched." An anonymous reader quotes MIT Technology Review: Last summer the Pentagon staged a contest in Las Vegas in which high-powered computers spent 12 hours trying to hack one another in pursuit of a $2 million purse. Now Mayhem, the software that won, is beginning to put its hacking skills to work in the real world... Teams entered software that had to patch and protect a collection of server software, while also identifying and exploiting vulnerabilities in the programs under the stewardship of its competitors... ForAllSecure, cofounded by Carnegie Mellon professor David Brumley and two of his PhD students, has started adapting Mayhem to be able to automatically find and patch flaws in certain kinds of commercial software, including that of Internet devices such as routers.

Tests are underway with undisclosed partners, including an Internet device manufacturer, to see if Mayhem can help companies identify and fix vulnerabilities in their products more quickly and comprehensively. The focus is on addressing the challenge of companies needing to devote considerable resources to supporting years of past products with security updates... Last year, Brumley published results from feeding almost 2,000 router firmware images through some of the techniques that powered Mayhem. Over 40%, representing 89 different products, had at least one vulnerability. The software found 14 previously undiscovered vulnerabilities affecting 69 different software builds. ForAllSecure is also working with the Department of Defense on ideas for how to put Mayhem to real world use finding and fixing vulnerabilities.

Facebook

Facebook's AI Unlocks the Ability To Search Photos By What's in Them (techcrunch.com) 62

An anonymous reader shares a TechCrunch report: Initially used to improve the experience for visually impaired members of the Facebook community, the company's Lumos computer vision platform is now powering image content search for all users. This means you can now search for images on Facebook with keywords that describe the contents of a photo, rather than being limited by tags and captions. To accomplish the task, Facebook trained an ever-fashionable deep neural network on tens of millions of photos. Facebook's fortunate in this respect because its platform is already host to billions of captioned images. The model essentially matches search descriptors to features pulled from photos with some degree of probability.

Slashdot Top Deals