AI

Microsoft's Cortana Doesn't Put Up With Sexual Harassment (hothardware.com) 153

MojoKid writes: Not long after Apple unveiled its Siri personal assistant to the world, it took very little time before people began asking her outrageous questions, sometimes inappropriate or just humorous, if for no other reason than they just could. When creating Cortana, Microsoft was well-aware of what its digital assistant was going to have to deal with, so, believe it or not, it was designed in such a way to handle abuse in a specific manner. According to Microsoft's Deborah Harrison, who is one of eight writers for Cortana, a chunk of the earliest queries were about Cortana's sex life. A specific goal was to make sure Cortana wasn't treated as a subservient. If she's insulted, she doesn't apologize or back down. She handles it with tact, so as to reduce the chance of further abuse.
AI

Financial Advisers Disrupted By AI (bloomberg.com) 71

schwit1 writes: Banks are watching wealthy clients flirt with robo-advisers, and that's one reason the lenders are racing to release their own versions of the automated investing technology this year, according to a consultant. Robo-advisers, which use computer programs to provide investment advice online, typically charge less than half the fees of traditional brokerages, which cost at least 1 percent of assets under management.
AI

Harnessing Artificial Intelligence To Build an Army of Virtual Analysts 41

An anonymous reader writes: PatternEx, a startup that gathered a team of AI researcher from MIT CSAIL as well as security and distributed systems experts, is poised to shake up things in the user and entity behavior analytics market. Their goal was to make a system capable of mimicking the knowledge and intuition of human security analysts so that attacks can be detected in real time. The platform can go through millions of events per day and can make an increasingly better evaluation of whether they are anomalous, malicious or benign.
Transportation

Jaguar Land Rover To Test Autonomous Cars In 'Living Lab' (thestack.com) 24

An anonymous reader writes: British automaker Jaguar Land Rover has announced its £5.5 million investment in a 'living lab' for the testing and development of connected and self-driving car technologies. The UK Connected Intelligent Transport Environment (CITE) will span 41-miles of public roads around Coventry and Solihull, and will be used to test new connected and autonomous vehicle (CAV) systems in real-life conditions. The company is planning to install roadside sensor equipment around the lab route to monitor vehicle-to-vehicle and vehicle-to-infrastructure communications. The fleet will include 100 CAV cars, which will test four different connectivity technologies; 4G long-term evolution (LTE) and its more advanced version LTE-V, dedicated short-range communication (DSRC), and local Wi-Fi hotspots.
Robotics

Let's Tear Down a Kiva Bot! (robohub.org) 22

Ben Einstein, writes new submitter Robofenix2, has torn down a Kiva bot -- a mobile ground-based warehouse delivery drone, aka Amazon's busiest employee. These robotic systems have revolutionised the warehouse distribution industry helping deliver packages. Ben was able to get his hands on an older generation, end-of-life Kiva bot and cracked open its bright orange shell to expose a brilliant piece of engineering; this post shares the fruits of Kiva's hard work. This 2011 video is also worth viewing, not least to see Kiva's shelf-lifting corkscrew action.
Classic Games (Games)

Computer Beats Go Champion 149

Koreantoast writes: Go (weiqi), the ancient Chinese board game, has long been held up as one of the more difficult, unconquered challenges facing AI scientists... until now. Google DeepMind researchers, led by David Silver and Demis Hassabis, developed a new algorithm called AlphaGo, enabling the computer to soundly defeat European Go champion Fan Hui in back-to-back games, five to zero. Played on a 19x19 board, Go players have more than 300 possible moves per turn to consider, creating a huge number of potential scenarios and a tremendous computational challenge. All is not lost for humanity yet: DeepMind is scheduled to face off in March with Lee Sedol, considered one of the best Go players in recent history, in a match compared to the Kasparov-Deep Blue duels of previous decades.
AI

Marvin Minsky, Pioneer In Artificial Intelligence, Dies at 88 (nytimes.com) 76

An anonymous reader sends word that Marvin Lee Minsky, co-founder of the Massachusetts Institute of Technology's AI laboratory, has died. The Times reports: "Marvin Minsky, who combined a scientist’s thirst for knowledge with a philosopher’s quest for truth as a pioneering explorer of artificial intelligence, work that helped inspire the creation of the personal computer and the Internet, died on Sunday night in Boston. He was 88. Well before the advent of the microprocessor and the supercomputer, Professor Minsky, a revered computer science educator at M.I.T., laid the foundation for the field of artificial intelligence by demonstrating the possibilities of imparting common-sense reasoning to computers."
AI

Microsoft Releases Its Deep Learning Toolkit On GitHub (microsoft.com) 53

An anonymous reader writes: Microsoft is moving its machine learning Computational Network Toolkit (CNTK) from its own hosting site, CodePlex, to GitHub. They're also putting it under the MIT open source license. The move marks an effort to make it easier for developers to collaborate on building their own deep learning applications using the CNTK. Under the CodePlex license, access was restricted to academics only, and it was wholly targeted to that audience. Now that it's opening the project to everyone, Microsoft hopes to attract a greater number of developers, and a wider variety as well. This follows similar releases from Google and Baidu.
AI

Google Launches Free Course On Deep Learning (blogspot.com) 16

An anonymous reader writes: In November, Google open sourced TensorFlow, its machine learning platform. Now, the company is following up by teaching people how to use it. They've launched a free course at Udacity that "provides you with all the basic tools and vocabulary to get started with deep learning, and walks you through how to use it to address some of the most common machine learning problems." A series of lectures explains how to set up your data, build training models, and extend those models. It also touches on image recognition and how to use recurrent neural networks. The signup page notes that this is considered an intermediate-to-advanced level course, so you'll probably need some basic machine learning knowledge to get the most out of it.
Businesses

A.I. Startups Building Bots For Businesses (xconomy.com) 25

gthuang88 writes: Virtual digital assistants are gaining popularity with the rise of Siri, Google Now, and Facebook's M service. Now startups are using related artificial intelligence techniques to solve business problems. Talla is building an interactive bot on Slack and HipChat for handling workflows in recruiting and human resources. The software uses natural language processing, word vectors, and some deep learning. Other startups, such as Gamalon, DataRobot, and Sentenai, are focused on probabilistic programming, data science, and machine learning for the Internet of things. Working with private data sets and business apps could help these startups avoid competing with the big players, at least for now.
AI

How Robotaxis Might Mitigate Electric Car Depreciation (robohub.org) 111

Hallie Siegel writes: Autonomous car expert Brad Templeton argues that we're in for a period of about 5 years in electric cars where each year's new model is a lot better, and that could be a problem for people trying to sell them. Further exacerbating Moore's Law for cars is that autonomous features (like traffic jam assist) rely heavily on computers. Unfortunately cars cost a lot more than computers or cell phones, so throwing them away before the end of their lifespan is a bit of a problem. How do get over the depreciation problem while autonomous cars and electric cars are going through this period of rapid development? Templeton suggests that a taxi model could be the answer, since use is so much more intense that with a private ownership model, that the cars are likely to wear out before they become worthless from a resale perspective.
AI

Can Author Obfuscation Trump Forensic Linguistics? (webis.de) 84

An anonymous reader writes: Everyone possesses their own writing style, which may be used to identify authors even if they wish to remain anonymous: linguists employ stylometry to settle disputes over the authorship of historic texts as well as more recent cases, and are called to verify the authors of suicide notes or threatening letters. Computer linguists carry out research on software for forensic text analyses, and a recent study shows many of these approaches to be reproducible. Now, a competition has been announced to develop obfuscation software to hide an author's style with the task: "Given a document, paraphrase it so that its writing style does not match that of its original author, anymore." We'll see what comes out of that. Meanwhile, the question remains: Who will win in the long run? Forensic linguists, or obfuscation technology?
AI

Baidu Releases Open Source Artificial Intelligence Code (thestack.com) 34

An anonymous reader writes: Chinese web services company Baidu has released a new artificial intelligence software called WARP-CTC. The code is apparently capable of speech recognition, particularly for short segments, that exceeds human capability. The source code uses an approach called 'connectionist temporal classification' and has been released on GitHub.
Yahoo!

Yahoo Releases Largest Ever Machine Learning Dataset To Researchers (tumblr.com) 41

An anonymous reader writes: Yahoo Labs has released a record-breaking dataset containing 110 billion interactions from 20 million Yahoo News users in 1.5TB of zipped data. The anonymized data is intended for research initiatives in artificial intelligence, including user-behavior modeling, collaborative filtering techniques and unsupervised learning methods.
Security

Smartwatches Can Be Used To Spy On Your Card's PIN Code (softpedia.com) 50

An anonymous reader writes: A researcher has developed a smartwatch app that can interpret hand motions and translate the movements to specific keystrokes on 12-key keypads, like the ones used at ATMs. The app sends the data to a nearby smartphone, which then relays it to a server, for analysis. The whole AI algorithm on which it's built has a 73% accuracy for touchlogging events, and 59% for keylogging. The entire code is on GitHub, along with his research paper, and a YouTube video.
AI

Coast-To-Coast Autonomous Tesla Trips 2-3 Years Out, Says Elon Musk (google.com) 259

Jalopnik reports that Elon Musk's predicted window for being able (for Tesla owners, that is) to call up your autonomous car and have it find its own way from New York to California, or vice versa, is astonishingly close: 24-36 months from now. From the article: As far as the summoning feature is concerned, Tesla plans for the 33-foot range to greatly expand—soon. Within two years, Musk predicted that owners will be able to summon their car from across the country. “If you’re in New York and your car is in Los Angeles, you can summon your car to you from your phone and tell the car to find you,” Musk said. “It’ll automatically charge itself along the journey. I might be slightly optimistic about that, but not significantly optimistic.” In getting from one place to another, Musk said autopilot “is better than human in highway driving, or at least it will be soon with machine learning.” If it’s not already better than human, Musk said it will be within the coming months. But right now, Musk said the car still needs a human around, just in case. “The car currently has sensors to achieve that cross-country goal,” Musk said. “But you’d need more hardware and software, you’d need more cameras, more radars, redundant electronics, redundant power buses and that sort of thing.
Businesses

Apple Purchases Software Company To Read Users' Expressions (thestack.com) 56

An anonymous reader writes: Apple's first disclosed acquisition of 2016 is software company Emotient, which specializes in reading users' expressions while they operate computers. Emotient uses AI software to break down micro-emotions shown on each face in a video frame and quantify it into three indicators: is the subject paying attention to the advertising, are they emotionally engaged, and are they showing a positive or negative emotion? The faces are pixelated to provide user anonymity without sacrificing the expression.
AI

Tesla Model S Software Updates Lets Car Park Itself With No One Inside It (bgr.com) 145

An anonymous reader writes with a link to this article at Boy Genius Report about a software upgrade now hitting Tesla owners, which begins: Tesla earlier today began pushing out version 7.1 of its software to Model S and Model X owners and, suffice it to say, it's a doozy of a software update. While we'll get to the full changelog shortly, we first wanted to highlight a feature called Summon which enables users to park their cars without having to be inside it. Conversely, it also lets Tesla owners summon their cars that already happen to be parked.
Input Devices

Video Api.ai CEO Ilya Gelfenbeyn Talks About Conversational Voice Interfaces (Video) 32

Api.ai makes an Android voice-controlled utility called Assistant. I have it on my Android phone. It is one of many simiar apps, and I have been trying them a little at a time. Are any of them as good as Siri? Let's just say, "Quality varies."

And Android voice assistants aren't the point of this interview, anyway. It's more about the process of developing interactive, voice-based IO systems. This whole voice/response thing is an area that's going to take off any year now -- and has been in that state for several decades -- but may finally be going somewhere, spurred by intense competition between the many companies working in this field, including Ilya's.
AI

Will Advanced AI Spell the End of Lawyers? 287

HughPickens.com writes: Lawyers have been described as the canaries in the coal mine in the face of a wave of automation now beginning to displace highly skilled white-collar workers as the increasing reliance on so-called "e-discovery" software in lawsuits raises the specter that $35-an-hour paralegals as well as $400-an-hour lawyers could fall victim to programs that could read and analyze legal documents more quickly and accurately than humans. Now John Markoff writes in the NY Times that a new study, "Can Robots Be Lawyers?", by Dana Remus analyzes which aspects of a lawyer's job could be automated and concludes that many of the tasks that lawyers perform fall well within human behavior that cannot be easily codified. "When a task is less structured, as many tasks are," writes Remus, "it will often be impossible to anticipate all possible contingencies."

According to Markoff being a lawyer involves performing a range of tasks including counseling, appearing in court, and persuading juries. Reading documents accounts for a relatively modest portion of a lawyer's activities. Remus estimates that about 13 percent of all legal work might ultimately fall prey to automation. According to Markoff, if that amount of work disappeared in a single year, it would be devastating but implemented over many years, this amount of technological change will be less noticeable. Even in the case of start-ups like LegalZoom and Rocket Lawyer, two sites that can aid in the preparation of legal documents, the impact of automation will more likely be in expanding into underserved markets rather than in displacing existing legal services.. ""A careful look at existing and emerging technologies reveals that it is only relatively structured and repetitive tasks that can currently be automated," concludes Remus. "These tasks represent a relatively modest percentage of lawyers' billable hours."

Slashdot Top Deals