Samsung Electronics Co. says it will spend 75 million euro ($89 million) in TTTech, a Vienna, Austria-based company that provides technologies for auto networks and safety controls to Audi cars and others. The Thursday announcement to invest in a partner company of Audi AG comes as the South Korean company completed its acquisition of Harman. Samsung said it created a business unit at Harman to be tasked with autonomous driving and advanced driver assistance systems. Samsung's investment in TTTech is the first investment from its $300 million fund to strengthen its technology pool for autonomous driving, such as artificial intelligence and security. Samsung is the world's largest maker of memory chips and smartphones.

"Clearly humans can't be allowed to have any interaction with their customers…" We've seen a couple of artificial intelligence (AI) shout-outs this week with Microsoft targeting telcos' worryingly poor customer service record and European scientists potentially creating the building blocks for Skynet, reports Telecoms.com Starting in the Microsoft offices, the team has paired up with Amdocs to provide an intelligent solution to one of the biggest questions which has faced the telco space in years; why are we so poor at customer services? In the Net Promoter (NP) annual report, the team put together a ranking of all the verticals in the US. At the top of the list were physical retail stores, and at the bottom of the list were internet service providers. Cable/Satellite providers were second to last, and cellular providers were 17 out of 23.

Two University of Oxford biomedical researchers are calling for robots to be built with real human tissue, and they say the technology is there if we only choose to develop it. Writing in Science Robotics, Pierre-Alexis Mouthuy and Andrew Carr argue that humanoid robots could be the exact tool we need to create muscle and tendon grafts that actually work. That's where robots come in. The researchers propose a "humanoid-bioreactor system" with "structures, dimensions, and mechanics similar to those of the human body." As the robot interacted with its environment, tissues growing on its body would receive the typical strains and twists that they would if they grew on an actual human.

On the roster: The scariest stat you'll see all day -Trump tries outreach to Dems on DREAMers - Dems set demands on taxes - Flynn pushed plan that profited his client - 'You're my boy, Blue' THE SCARIEST STAT YOU'LL SEE ALL DAY When you consider the fact that a third of American adults cannot name a single branch of their federal government, you cease to wonder why things are so bad and begin to wonder why they are not already worse. In a poll conducted for the Annenberg Public Policy Center ahead of this weekend's celebration of the 229th anniversary of the ratification of the Constitution, only 26 percent of respondents could identify the executive, legislative and judicial branches, while 40 percent could name only one or two. Americans talk openly and often about the dumbing down of our culture, what we refer to as "The Big Stupid." It is a lament, but also something of a brag for people not clutched by ignorance of this magnitude. But it's easy to be an intellectual elite in a nation where not even half of the people know what kind of government they have.

In modern computer vision problems such as facial recognition [1] and object tracking [2], researchers have found success applying the union of subspaces (UoS) model, in which data vectors lie near one of several subspaces. Under this model, the goal is to simultaneously identify these underlying subspaces and cluster the points according to their nearest subspace. Algorithms designed to solve this problem fall under the category of subspace clustering, a topic that has received a great deal of attention in recent years [3] due to its efficacy on real-world datasets such as the Extended Yale Face Database B [4] and the MNIST handwritten digit database [5]. One of the earliest approaches to solving the subspace clustering problem involves an iterative method in the spirit of K-means, known as K-subspaces (KSS) [6], [7], [8], which alternates between assigning points to clusters and estimating the subspace basis associated with each cluster. As this algorithm is only guaranteed to converge to a local minimum, in practice one runs many instances of the algorithm and chooses the final clustering as the one that produces the minimum cost. Although its empirical performance is limited, KSS continues to serve as a benchmark for subspace clustering algorithms, in part due to its computational efficiency and simplicity. Therefore, a deeper understanding of this method is an important contribution in the area of subspace clustering and a contribution of this paper. While the KSS cost function and alternating algorithm are perhaps the most natural approach for the subspace clustering problem, it is known that there is a set of initializations of nonzero measure from which the algorithm will convergence to a point other than the global minimizer.

We address the problem of multi-class classification in the case where the number of classes is very large. We propose a double sampling strategy on top of a multi-class to binary reduction strategy, which transforms the original multi-class problem into a binary classification problem over pairs of examples. The aim of the sampling strategy is to overcome the curse of long-tailed class distributions exhibited in majority of large-scale multi-class classification problems and to reduce the number of pairs of examples in the expanded data. We show that this strategy does not alter the consistency of the empirical risk minimization principle defined over the double sample reduction. Experiments are carried out on DMOZ and Wikipedia collections with 10,000 to 100,000 classes where we show the efficiency of the proposed approach in terms of training and prediction time, memory consumption, and predictive performance with respect to state-of-the-art approaches.

Moscow-based Kaspersky makes anti-virus software for computers. On Wednesday, the U.S. government banned federal agencies from using it over concerns about the company's ties to the Kremlin and Russian spy operatives. Kaspersky has denied any unethical ties with Russia or any government.

BOSTON – Self-driving cars may not hit the road in earnest for many years -- but autonomous boats could be just around the pier. Spurred in part by the car industry's race to build driverless vehicles, marine innovators are building automated ferry boats for Amsterdam canals, cargo ships that can steer themselves through Norwegian fjords and remote-controlled ships to carry containers across the Atlantic and Pacific. The first such autonomous ships could be in operation within three years. One experimental workboat spent this summer dodging tall ships and tankers in Boston Harbor, outfitted with sensors and self-navigating software and emblazoned with the words "UNMANNED VESSEL" across its aluminum hull. "We're in full autonomy now," said Jeff Gawrys, a marine technician for Boston start-up Sea Machines Robotics, sitting at the helm as the boat floated through a harbor channel.

As robots get more advanced and perform more complicated tasks, such as conducting a philharmonic orchestra in Italy, there is still one thing the machines certainly cannot do: feel. However, if scientists from the University of Houston have anything to say about it, that may change soon, at least insofar as the sense of touch is concerned. Cunjiang Yu, an assistant professor at the university and three other researchers created "a semiconductor in a rubber composite format" that can stretch and still retain functionality, allowing a robotic hand to feel temperature differences and distinguish between hot and cold. Writing in the journal Science Advances, they described "a new mechanism for producing stretchable electronics, a process that relies upon readily available materials and could be scaled up for commercial production," according to a statement on the university' website. Semiconductors are usually brittle and incorporating them into stretchable materials usually involves complicated procedures, Yu said in the statement, making the resulting materials both less stable and more expensive than the new material created by his team.

Since the introduction of 3-D Secure in 2007, adoption rates have been slow, and have varied between countries. Concerns about lower conversions and rising customer friction are preventing higher penetration rates, but new fraud prevention software solutions offer ways to balance customer satisfaction and security requirements. Roger Lester, Payments Expert at Featurespace, explores how advanced machine learning and adaptive behavioral analytics technology help balance 3-D Secure checks against futureproofing an organization's fraud defenses. Why has adoption of 3-D Secure been mixed? Designed to be an additional layer of security for online payments, the 3-D ("3 Domain") Secure protocol has been adopted by all major card networks--more familiar to many of us under commercial names, such as "Verified by Visa" and "MasterCard SecureCode."