It was shown recently that the $K$ L1-norm principal components (L1-PCs) of a real-valued data matrix $\mathbf X \in \mathbb R^{D \times N}$ ($N$ data samples of $D$ dimensions) can be exactly calculated with cost $\mathcal{O}(2^{NK})$ or, when advantageous, $\mathcal{O}(N^{dK - K + 1})$ where $d=\mathrm{rank}(\mathbf X)$, $K

Discrete combinatorial optimization plays a central role in many scientific disciplines, however for hard problems we lack linear time algorithms that would allow us to solve very large instances. Moreover it is still unclear what are the key features that make a discrete combinatorial optimization problem hard to solve. Here we study random K-satisfiability problems with K 3, 4 which are known to be very hard close to the SAT-UNSAT threshold, where problems stop having solutions. We show that the Backtracking Survey Propagation algorithm, in a time practically linear in the problem size, is able to find solutions very close to the threshold, in a region unreachable by any other algorithm. All solutions found have no frozen variables, thus supporting the conjecture that only unfrozen solutions can be found in linear time, and that a problem becomes impossibile to solve in linear time when all solutions contain frozen variables. Optimization problems with discrete variables are widespread among scientific disciplines and often among the hardest to solve.

Austin--Microsoft Ventures has invested in Austin artificial intelligence startup CognitiveScale, filling out its Series B investment round to more than 25 million, the company reported Tuesday. Cognitive Scale would not reveal the amount Microsoft has invested. In August, the startup announced it had raised a total of 21.8 million in the Series B round from investors such as Intel Capital and Norwest Venture Partners. The Microsoft investment will go towards developing artificial intelligence software for the computing giant's Microsoft HoloLens and Microsoft Azure products, according to a press release. "Our goal is to embed cognitive systems of intelligence into a whole range of new personal computing applications and business processes, from customer engagement, to procurement and regulatory compliance, using hyper-personalized holographic projections and mixed reality," CognitiveScale CEO Akshay Sabhikhi said in the press release.

Dean Pomerleau can still remember his first tussle with the black-box problem. The year was 1991, and he was making a pioneering attempt to do something that has now become commonplace in autonomous-vehicle research: teach a computer how to drive. This meant taking the wheel of a specially equipped Humvee military vehicle and guiding it through city streets, says Pomerleau, who was then a robotics graduate student at Carnegie Mellon University in Pittsburgh, Pennsylvania. With him in the Humvee was a computer that he had programmed to peer through a camera, interpret what was happening out on the road and memorize every move that he made in response. Eventually, Pomerleau hoped, the machine would make enough associations to steer on its own.

Miniature robots that doctors could guide through a patient's body to kill cancer cells are closer to reality thanks to winners of this year's Nobel Prize for Chemistry. Three winners share the 727,000 prize for developing nanoscale machines--1000th the width of a human hair--that pave the way for applications in medicine, computing and engineering. The winners were Jean-Pierre Sauvage of the University of Strasbourg in France, Fraser Stoddart of Northwestern University in Illinois, USA, and Bernard Feringa of the University of Groningen in the Netherlands. Each devised different groups of molecules with moving parts that they could control remotely, despite their tiny size. "It's early days, but once you can control movement, you have many possibilities," said Feringa, interviewed after receiving notification of the prize.

Arthur C. Clarke famously stated that "any sufficiently advanced technology is indistinguishable from magic." No current technology embodies this statement more than neural networks and deep learning. And like any good magic it not only dazzles and inspires but also puts fear into people's hearts. One known property of artificial neural networks (ANNs) is that they are universal function approximators. This means that any mathematical function can be represented by a neural network.

The 2016 Nobel Prize in chemistry has been awarded for the design and synthesis of the world's smallest machines. The work has overtones of science fiction, but holds huge promise in fields as diverse as medicine, materials and energy. This is especially true of efforts to develop nano-scale machines (1,000 times smaller than the width of a human hair), which are always destined to remain tiny however big our ambitions for them grow. It's difficult to trace the development of molecular machines to one person or scientific step. But a 1959 lecture by the celebrated physicist Richard Feynman is as good a point as any.

We know you're passionate about technology and machine learning is one of the most exciting topics in the industry today. We are currently investigating machine learning and artificial intelligence, and would like to gather your feedback and suggestions. Our objective is to identify the needs of our community on machine learning and artificial intelligence. It should take less than 5 minutes to fill out the survey. Survey responses will be anonymous.

The Initiative for a Synthesis in Studies of Awareness will organize a two-week Summer School, with plenary lectures in the morning and parallel sessions in the afternoon, in which the lecturers will lead study groups that are aimed at producing original research of publishable quality. The lectures will cover topics in various aspects of neuroscience, experimental as well as computational; theoretical physics; logic and philosophy; and various other fields in cognitive science and the study of complex systems, including artificial intelligence, artificial life, and robotics. We invite graduate students and postdoctoral researchers to participate in the summer school. Organizers will provide lodging for all accepted students and travel support for selected students. Applications will be open until December 25, 2016.

Hackers demonstrated they can take over a Tesla from miles away if it connects to a malicious Wi-Fi hotspot. Dmitri Dolgov, a longtime veteran of Google's seven-year self-driving car effort, recently took over as technical lead, replacing Chris Urmson. SAN FRANCISCO -- Google's self-driving cars have hit another milestone on the road to the automotive future, notching two million miles on the autonomous-testing odometer. That mark, which the Alphabet-owned company announced Wednesday, was hit as other companies spent the summer dominating the self-driving headlines. Uber recently began picking up Pittsburgh passengers in its small fleet of driverless (though driver-monitored) vehicles, while Ford announced plans to sell transportation that lacked a steering wheel and pedals by 2021.