AI

What Does Artificial Intelligence Actually Mean? (qz.com) 61

An anonymous reader writes: A new bill (pdf) drafted by senator Maria Cantwell asks the Department of Commerce to establish a committee on artificial intelligence to advise the federal government on how AI should be implemented and regulated. Passing of the bill would trigger a process in which the secretary of commerce would be required to release guidelines for legislation of AI within a year and a half. As with any legislation, the proposed bill defines key terms. In this, we have a look at how the federal government might one day classify artificial intelligence. Here are the five definitions given:

A) Any artificial systems that perform tasks under varying and unpredictable circumstances, without significant human oversight, or that can learn from their experience and improve their performance. Such systems may be developed in computer software, physical hardware, or other contexts not yet contemplated. They may solve tasks requiring human-like perception, cognition, planning, learning, communication, or physical action. In general, the more human-like the system within the context of its tasks, the more it can be said to use artificial intelligence.
B) Systems that think like humans, such as cognitive architectures and neural networks.
C) Systems that act like humans, such as systems that can pass the Turing test or other comparable test via natural language processing, knowledge representation, automated reasoning, and learning.
D) A set of techniques, including machine learning, that seek to approximate some cognitive task.
E) Systems that act rationally, such as intelligent software agents and embodied robots that achieve goals via perception, planning, reasoning, learning, communicating, decision-making, and acting.

Microsoft

Microsoft Releases Free Preview of Its Quantum Development Kit (zdnet.com) 30

Microsoft is releasing a free preview version of its Quantum Development Kit. "The kit includes the Q# programming language and compiler and a local quantum computing simulator, and is fully integrated with Visual Studio," reports ZDNet. "There's also an Azure-based simulator that allows developers to simulate more than 40 logical qubits of computing power, plus documentation libraries, and sample programs, officials said in their December 11 announcement." From the report: Quantum computers are designed to process in parallel, thus enabling new types of applications across a variety of workloads. They are designed to harness the physics of subatomic particles to provide a different way to store data and solve problems compared to conventional computers, as my ZDNet colleague Tony Baer explains. The result is that quantum computers could solve certain high-performance-computing problems more efficiently. Microsoft officials have said applications that developers create for use with the quantum simulator ultimately will work on a quantum computer, which Microsoft is in the process of developing. Microsoft's goal is to build out a full quantum computing system, including both the quantum computing hardware and the related full software stack.
AI

AI-Assisted Fake Porn Is Here and We're All Screwed (vice.com) 280

New submitter samleecole shares a report from Motherboard: There's a video of Gal Gadot having sex with her stepbrother on the internet. But it's not really Gadot's body, and it's barely her own face. It's an approximation, face-swapped to look like she's performing in an existing incest-themed porn video. The video was created with a machine learning algorithm, using easily accessible materials and open-source code that anyone with a working knowledge of deep learning algorithms could put together. It's not going to fool anyone who looks closely. Sometimes the face doesn't track correctly and there's an uncanny valley effect at play, but at a glance it seems believable. It's especially striking considering that it's allegedly the work of one person -- a Redditor who goes by the name 'deepfakes' -- not a big special effects studio that can digitally recreate a young Princess Leia in Rouge One using CGI. Instead, deepfakes uses open-source machine learning tools like TensorFlow, which Google makes freely available to researchers, graduate students, and anyone with an interest in machine learning. Anyone could do it, and that should make everyone nervous.
Businesses

Fired Tech Workers Turn To Chatbots for Counseling (bloomberg.com) 96

An anonymous reader shares a Bloomberg report: For months Lovkesh Joshi was quietly terrified of losing his job as a manager at a top Indian tech services company. Joshi didn't want to burden his wife or friends so he turned to a chatbot therapist called Wysa. Powered by AI, the app promises to be "loyal, supportive and very private," and encourages users to divulge their feelings about a recent major event or big change in their lives. "I could open up and talk," says the 41-year-old father of two school-age children, who says his conversations with the bot flowed naturally. "I felt heard and understood." Joshi moved to a large rival outsourcer two months ago. The upheaval in India's $154 billion tech outsourcing industry has prompted thousands of Indians to seek solace in online therapy services. People accustomed to holding down prestigious jobs and pulling in handsome salaries are losing out to automation, a shift away from long-term legacy contracts and curbs on U.S. work visas. McKinsey & Co says almost half of the four million people working in India's IT services industry will become "irrelevant" in the next three to four years. Indians, like people the world over, tend to hide their mental anguish for fear of being stigmatized. That's why many are embracing the convenience, anonymity and affordability of online counseling startups, most of which use human therapists.
NASA

Google's Machine Learning Is Analyzing Data From NASA's Kepler Space Telescope (nasa.gov) 27

NASA writes: NASA will host a media teleconference at 1 p.m. EST Thursday, Dec. 14, to announce the latest discovery made by its planet-hunting Kepler space telescope. The discovery was made by researchers using machine learning from Google. Machine learning is an approach to artificial intelligence, and demonstrates new ways of analyzing Kepler data... When Kepler launched in March 2009, scientists didn't know how common planets were beyond our solar system. Thanks to Kepler's treasure trove of discoveries, astronomers now believe there may be at least one planet orbiting every star in the sky.
Space.com adds: Kepler spots alien worlds by noticing the tiny brightness dips they cause when they cross the face of their host star from the spacecraft's perspective. Kepler is the most accomplished planet hunter in history. It has found more than 2,500 confirmed alien worlds -- about 70 percent of all known exoplanets -- along with a roughly equal number of "candidates" that await confirmation by follow-up observations or analyses. The vast majority of these discoveries have come via observations that Kepler made during its original mission, which ran from 2009 to 2013. Study of these data sets is ongoing; over the past few years, researchers have used improved analysis techniques to spot many exoplanets in data that Kepler gathered a half-decade ago or more.
Space.com describes Thursday's announcement as an exoplanet discovery. (Earlier they reported on the discovery of "a possibly habitable alien world" about 2.2 times the size of earth orbiting a dwarf star "within the range of distances where liquid water could exist on a world's surface".)

Slashdot reader schwit1 points out that other less-credible sites speculate NASA's announcement will be "a major discovery about life beyond earth."
AI

Emotion Recognition Systems Could Be Used In Job Interviews (techtarget.com) 145

dcblogs writes: Emotion recognition software identifies micro-expressions through video analysis. These are expressions that may be as fast as 1/25 of a second and invisible to the human eye, but a close analysis of video can detect them. These systems are being used in marketing research, but some employers may be interested in using them to assess job candidates.

Vendors claim these systems can be used to develop a personality profile and discover a good cultural fit. The technology raises concerns, illustrated earlier this year who showed that face-reading technology could use photographs to determine sexual orientation with a high degree of accuracy.

One company has already added face recognition into their iPad-based time clock, which the company's CEO thinks could be adapted to also detect an employee's mood when they're clocking out. Yet even he has his reservations. While he thinks it could provide more accurate feedback from employees, he also admits that "There's something very Big Brother about it."
AI

Elon Musk Says Tesla Is Building Dedicated Chips For Autopilot (theregister.co.uk) 32

Elon Musk says Tesla is developing its own chip to run the Autopilot system in future vehicles from the firm. The news was revealed at a Tesla party that took place at the intelligence conference NIPS. Attendees at the party told The Register that Musk said, "I wanted to make it clear that Tesla is serious about AI, both on the software and hardware fronts. We are developing custom AI hardware chips." From the report: Musk offered no details of his company's plans, but did tell the party that "Jim is developing specialized AI hardware that we think will be the best in the world." "Jim" is Jim Keller, a well-known chip engineer who was lead architect on a range of silicon at AMD and Apple and joined Tesla in 2016. Keller later joined Musk on a panel discussing AI at the Tesla Party alongside Andrej Karpathy, Tesla's Director of AI and chaired by Shivon Zilis, a partner and founding member at Bloomberg Beta, a VC firm. Musk is well known for his optimism about driverless cars and pessimism about whether AI can operate safely. At the party he voiced a belief that "about half of new cars built ten years from now will be autonomous." He added his opinion that artificial general intelligence (AGI) will arrive in about seven or eight years.
Hardware

Nvidia Announces 'Nvidia Titan V' Video Card: GV100 for $3000 (anandtech.com) 51

Nvidia has announced the Titan V, the "world's most powerful PC GPU." It's based on Nvidia's Volta, the same architecture as the Nvidia Tesla V100 GPUs behind Amazon Web Service's recently launched top-end P3 instances, which are dedicated to artificial-intelligence applications. From a report: A mere 7 months after Volta was announced with the Tesla V100 accelerator and the GV100 GPU inside it, Nvidia continues its breakneck pace by releasing the GV100-powered Titan V, available for sale today. Aimed at a decidedly more compute-oriented market than ever before, the 815 mm2 behemoth die that is GV100 is now available to the broader public. [...] The Titan V, by extension, sees the Titan lineup finally switch loyalties and start using Nvidia's high-end compute-focused GPUs, in this case the Volta architecture based V100. The end result is that rather than being Nvidia's top prosumer card, the Titan V is decidedly more focused on compute, particularly due to the combination of the price tag and the unique feature set that comes from using the GV100 GPU. Which isn't to say that you can't do graphics on the card -- this is still very much a video card, outputs and all -- but Nvidia is first and foremost promoting it as a workstation-level AI compute card, and by extension focusing on the GV100 GPU's unique tensor cores and the massive neural networking performance advantages they offer over earlier Nvidia cards.
Android

Qualcomm Snapdragon 845 Boosts CPU and GPU Performance, Delivers 4K HDR Capture, 3x Faster AI Processing (hothardware.com) 48

MojoKid writes: Qualcomm provided a deep-dive view today of its Snapdragon 845 mobile processor platform that it recently announced, highlighting key advancements in what the company is referring to as a completely new silicon design. The new chip now employs a Kyro 385 CPU with four high performance cores at 2.8GHz (25 percent faster than the previous gen Snapdragon 835) and four "efficiency" cores operating at 1.7GHz. The new chip also includes the new Spectra 280 image signal processor (ISP). Compared to its predecessor, the image signal processor (ISP) in the Snapdragon 845 promises a 64x uplift in the ability to capture high dynamic range (HDR) color information for 4K HDR video capture and playback. The chip's new Adreno 630 GPU promises a 30 percent boost in gaming performance compared to its predecessor, along with room-scale VR/AR experiences that support 6 degrees of freedom along with simultaneous localization and mapping, or SLAM. Finally, the new SoC platform incorporates Qualcomm's second-generation gigabit LTE modem: the Snapdragon X20. This Cat 18 modem supports peak download speeds of 1.2Gbps along with 5x carrier aggregation, 4x4 MIMO, and Dual SIM-Dual VoLTE. Qualcomm says that the first Snapdragon 845 processors will begin shipping in production devices in early 2018.
AI

AI Can Beat Humans Only One Game At a Time (axios.com) 60

An anonymous reader shares a report: Despite all of the potential for artificial intelligence to solve our most vexing problems, it's still in a primitive state, according to a new report by Stanford University. But a separate paper, this one by Alphabet's DeepMind, suggests again that it has made some of its best progress in the narrow realm of games. Why it matters: Those advances are important, but life isn't a game. AI progress outside of these areas has been harder to define and track. "The most important thing for AI is to go from exceptional promise to use in actual everyday life," Martial Hebert, director of the Robotics Institute at Carnegie Mellon University, tells Axios.
AI

Google's DeepMind AI Becomes a Superhuman Chess Player In a Few Hours (theverge.com) 93

An anonymous reader quotes a report from The Verge: In a new paper published this week, DeepMind describes how a descendant of the AI program that first conquered the board game Go has taught itself to play a number of other games at a superhuman level. After eight hours of self-play, the program bested the AI that first beat the human world Go champion; and after four hours of training, it beat the current world champion chess-playing program, Stockfish. Then for a victory lap, it trained for just two hours and polished off one of the world's best shogi-playing programs named Elmo (shogi being a Japanese version of chess that's played on a bigger board). One of the key advances here is that the new AI program, named AlphaZero, wasn't specifically designed to play any of these games. In each case, it was given some basic rules (like how knights move in chess, and so on) but was programmed with no other strategies or tactics. It simply got better by playing itself over and over again at an accelerated pace -- a method of training AI known as "reinforcement learning."
AI

Inside Baidu's Bid To Lead the AI Revolution (wired.com) 12

mirandakatz writes: China's search giant missed mobile: As WeChat and Alibaba deftly transformed their companies to suit mobile, Baidu stayed stuck in browser mode. It can't afford to make that mistake with the AI revolution -- and, as Jessi Hempel writes at Backchannel, it just might have an edge in its bid to come out on top. There's huge governmental support for AI in China, including a plan to make the country the world leader in AI by 2030, and it has double the number of people online than America does -- AKA vast quantities of raw data. Hempel traveled to Beijing to chronicle this tenuous moment in Baidu's history, and has delivered a deep look at Baidu's AI be on AI, speaking with key leaders including CEO Robin Li and COO Qi Lu. She writes that 'Robin Li is doubling down on a future beyond 2017. In that future, Baidu is not a series of products, but rather an engine that belongs inside everything -- an engine that powers Baidu back to dominance in China, and possibly far beyond.'
The Almighty Buck

'We Could Fund a Universal Basic Income With the Data We Give Away To Facebook and Google' (thenextweb.com) 583

Tristan Greene reports via The Next Web: A universal basic income (UBI), wherein government provides a monthly stipend so citizens can afford a home and basic necessities, is something experts believe would directly address the issue of unemployment and poverty, and possibly even eliminate hundreds of other welfare programs. It may also be the only real solution to the impending automation bonanza. According to AI expert Steve Fuller, the problem is, giving people money when they lose jobs won't fix the issue, it's a temporary solution and we need permanent ones. Sounds fair, and he even has some ideas on how to accomplish this end: "We could hold Google and Facebook and all those big multinationals accountable; we could make sure that people, like those who are currently 'voluntarily' contributing their data to pump up companies' profits, are given something that is adequate to support their livelihoods in exchange."

It's an interesting idea, but difficult to imagine it's implementation. If the government isn't assigning a specific stipend value, we'll have to be compensated individually by companies. One way to do this, is by emulating the old coal mining company scrip scams of early last century. Employees working for companies would be paid in currency only redeemable at the company store. This basically created a system where a company could tax its own workers for profit. Google, for example, could use a system like that and say "opt-in for $10 worth of Google Play music for free," if they wanted to. Which doesn't help pay the bills when machines replace you at work, but at least you'll be able to voice search for your favorite songs. Another idea is to charge companies an automation tax, but again there's concerns as to how this would be implemented. A solution that combines government oversight with a tax on AI companies -- a UBI funded by the dividends of our data -- may be the best option. To be blunt: we should make Google, Microsoft, Facebook and other such AI companies pay for it with a simple data tax.

AI

Tencent Says There Are Only 300,000 AI Engineers Worldwide, But Millions Are Needed (theverge.com) 116

An anonymous reader quotes a report from The Verge: It's well-established that talent is in short supply in the AI industry, but a new report from Chinese tech giant Tencent underscores how great the need might be. According to the study, compiled by the Tencent Research Institute, there are just 300,000 "AI researchers and practitioners" worldwide, but the "market demand" is for millions of roles. These are unavoidably speculative figures, and the study does not offer much detail on how they were reached, but as a general trend they fit with other, more anecdotal reports. Around the world, tech giants regularly complain about the difficulty hiring AI engineers, and the demand has pushed salaries to absurd heights. Individuals with just a few year's experience can expect base pay of between $300,000 and $500,000 a year, says The New York Times, while the very best will collect millions. One independent AI lab told the publication that there were only 10,000 individuals worldwide with the right skills to spearhead serious new AI projects.

Tencent's new "2017 Global AI Talent White Paper" suggests the bottleneck here is education. It estimates that 200,000 of the 300,000 active researchers are already employed in various industries (not just tech), while the remaining 100,000 are still studying. Attendance in machine learning and AI courses has skyrocketed in recent years, as has enrollment in online courses, but there is obviously a lag as individuals complete their education.

Security

A Popular Virtual Keyboard App Leaks 31 Million Users' Personal Data (zdnet.com) 65

Zack Whittaker, writing for ZDNet: Personal data belonging to over 31 million customers of a popular virtual keyboard app has leaked online, after the app's developer failed to secure the database's server. The server is owned by Eitan Fitusi, co-founder of AI.type, a customizable and personalizable on-screen keyboard, which boasts more than 40 million users across the world. But the server wasn't protected with a password, allowing anyone to access the company's database of user records, totaling more than 577 gigabytes of sensitive data. The database appears to only contain records on the app's Android users.
Censorship

Apple, Google CEOs Bring Star Power as China Promotes Censorship (bloomberg.com) 38

An anonymous reader shares a Bloomberg report: Apple's Tim Cook and Google's Sundar Pichai made their first appearances at China's World Internet Conference, bringing star power to a gathering the Chinese government uses to promote its strategy of tight controls online. Apple's chief executive officer gave a surprise keynote at the opening ceremony on Sunday, calling for future internet and AI technologies to be infused with privacy, security and humanity. The same day, one of China's most-senior officials called for more aggressive government involvement online to combat terrorism and criminals. Wang Huning, one of seven men on China's top decision-making body, even called for a global response team to go well beyond its borders. It was Cook's second appearance in China in two months, following a meeting with President Xi Jinping in October. The iPhone maker has most of its products manufactured in the country and is trying to regain market share in smartphones against local competitors such as Huawei. "The theme of this conference -- developing a digital economy for openness and shared benefits -- is a vision we at Apple share," Cook said. "We are proud to have worked alongside many of our partners in China to help build a community that will join a common future in cyberspace."
AI

Google's AI Built an AI that Outperforms Any Made By Humans (sciencealert.com) 235

schwit1 quotes ScienceAlert: In May 2017, researchers at Google Brain announced the creation of AutoML, an artificial intelligence (AI) that's capable of generating its own AIs. More recently, they decided to present AutoML with its biggest challenge to date, and the AI that can build AI created a 'child' that outperformed all of its human-made counterparts... For this particular child AI, which the researchers called NASNet, the task was recognising objects -- people, cars, traffic lights, handbags, backpacks, etc. -- in a video in real-time. AutoML would evaluate NASNet's performance and use that information to improve its child AI, repeating the process thousands of times.

When tested on the ImageNet image classification and COCO object detection data sets NASNet was 82.7 percent accurate at predicting images on ImageNet's validation set. This is 1.2 percent better than any previously published results, and the system is also 4 percent more efficient, with a 43.1 percent mean Average Precision (mAP).

AI

Two Technologists Create Black Metal Album Using An AI (theoutline.com) 57

An anonymous reader quotes a report from The Outline: Coditany of Timeness" is a convincing lo-fi black metal album, complete with atmospheric interludes, tremolo guitar, frantic blast beats and screeching vocals. But the record, which you can listen to on Bandcamp, wasn't created by musicians. Instead, it was generated by two musical technologists using a deep learning software that ingests a musical album, processes it, and spits out an imitation of its style. To create Coditany, the software broke "Diotima," a 2011 album by a New York black metal band called Krallice, into small segments of audio. Then they fed each segment through a neural network -- a type of artificial intelligence modeled loosely on a biological brain -- and asked it to guess what the waveform of the next individual sample of audio would be. If the guess was right, the network would strengthen the paths of the neural network that led to the correct answer, similar to the way electrical connections between neurons in our brain strengthen as we learn new skills.
AI

Stephen Hawking: 'I Fear AI May Replace Humans Altogether' (wired.co.uk) 282

dryriver writes: Wired magazine recently asked physicist Stephen Hawking what he thinks of everything from AI to the Anti Science Movement. One of the subjects touched on was the control large corporations have over information in the 21st Century. In Hawking's own words: "I worry about the control that big corporations have over information. The danger is we get into the situation that existed in the Soviet Union with their papers, Pravda, which means "truth" and Izvestia, which means "news". The joke was, there was no truth in Pravda and no news in Izvestia. Corporations will always promote stories that reflect well on them and suppress those that don't." And since this is Slashdot, here's what Stephen Hawking said about Artificial Intelligence: "The genie is out of the bottle. We need to move forward on artificial intelligence development but we also need to be mindful of its very real dangers. I fear that AI may replace humans altogether. If people design computer viruses, someone will design AI that replicates itself. This will be a new form of life that will outperform humans."
AI

AI Goes Bilingual -- Without a Dictionary (sciencemag.org) 99

sciencehabit shares a report from Science Magazine: Automatic language translation has come a long way, thanks to neural networks -- computer algorithms that take inspiration from the human brain. But training such networks requires an enormous amount of data: millions of sentence-by-sentence translations to demonstrate how a human would do it. Now, two new papers show that neural networks can learn to translate with no parallel texts -- a surprising advance that could make documents in many languages more accessible.

The two new papers, both of which have been submitted to next year's International Conference on Learning Representations but have not been peer reviewed, focus on another method: unsupervised machine learning. To start, each constructs bilingual dictionaries without the aid of a human teacher telling them when their guesses are right. That's possible because languages have strong similarities in the ways words cluster around one another. The words for table and chair, for example, are frequently used together in all languages. So if a computer maps out these co-occurrences like a giant road atlas with words for cities, the maps for different languages will resemble each other, just with different names. A computer can then figure out the best way to overlay one atlas on another. Voila! You have a bilingual dictionary.
The studies -- "Unsupervised Machine Translation Using Monolingual Corpora Only" and "Unsupervised Neural Machine Translation" -- were both submitted to the e-print archive arXiv.org.

Slashdot Top Deals