The goal in supervised learning is to learn from examples a function that predicts well new data. Nonparametric methods are often crucial since the functions to be learned can be nonlinear and complex Kernel methods are probably the most popular among nonparametric learning methods, but despite excellent theoretical properties, they have limited applications in large scale learning because of time and memory requirements, typically at least quadratic in the number of data points. Overcoming these limitations has motivated a variety of practical approaches including gradient methods, as well accelerated, stochastic and preconditioned extensions, to improve time complexity [1, 2, 3, 4, 5, 6]. Random projections provide an approach to reduce memory requirements, popular methods including Nyström [7, 8], random features [9], and their numerous extensions. From a theoretical perspective a key question has become to characterize statistical and computational tradeoffs, that is if, or under which conditions, computational gains come at the expense of statistical accuracy.

Axios recently reported that it had discovered a document that revealed something very interesting: The Trump administration was considering a government-run 5G network. According to the memo, this was in order to fight China's upcoming dominance in the wireless 5G space, and would ensure a safe network for self-driving cars, AI, VR and other cutting-edge technologies. This kind of state-run network is completely antithetical to the administration's public stance on deregulation and privatization. It even prompted FCC Chairman Ajit Pai to come down strongly in opposition. It turns out, however, that the document was outdated, and the Trump administration strongly denies it ever seriously considered such a proposal.

With artificial intelligence weapons on both sides, are we in a new cold war? It is easy to confuse the current geopolitical situation with that of the 1980s. The United States and Russia each accuse the other of interfering in domestic affairs.…

While drones and driverless cars dominate the headlines, another breakthrough--robot dexterity--is likely to have an even greater impact in both business and everyday life. "Robot manipulation is the next shoe to drop," says Robert Platt, computer science professor and head of the Helping Hands robotics lab at Northeastern. "Imagine a robot that can do things with it's hands in the real world--anything from defusing a bomb to doing your laundry. This has been a dream in the research community for decades, but now we're finally getting to the point where it could actually happen." Recent advances in machine learning, Big Data, and robot perception have put us on the threshold of a quantum leap in the ability of robots to perform fine motor tasks and function in uncontrolled environments, says Platt.

We have a patch to unpatch your patch, and Dell might reverse-merger itself back into existence as a publicly-traded company. Yeah, you read that correctly. Please apply to the damaged area.Microsoft's new Windows 10 Spectre patch disables Intel's fix Unfortunately, Intel's recent patch for the Spectre CPU issue caused spontaneous reboots, so now Microsoft has released a Windows patch that essentially undoes the fix. If you've already applied Intel's update, it should solve the rebooting problem until Intel applies a new, better patch. You know what a reverse merger is, right?Dell may sell itself to VMware The thing is, Dell owns 80 percent of VMware.

WASHINGTON – Donald Trump's misspellings have become legendary -- "covfefe," anyone? Tickets to the prime-time speech before a joint session of Congress were printed inviting lawmakers and guests to the State of the "Uniom," lawmakers and officials said Monday. "Just received my ticket for the State of the Union. Looks like @BetsyDeVosED was in charge of spell checking… #SOTUniom," tweeted House Democrat Raul Grijalva, referring to Education Secretary Betsy DeVos, who has advocated policies that critics say undermine the public education system. Grijalva's tweet also included a photograph of the offending ticket.

Some of the tickets issued for President Donald Trump's first State of the Union address to Congress on Tuesday had a glaring mistake and provided invitees a welcome to the "State of the Uniom." An Arizona House member Monday tweeted a picture of his own ticket for President Trump's address and pointed out that the word "Union" was misspelled as "Uniom." Rep. Raul Grijalva, a Democrat, joked in his post saying that the often criticized secretary of education was responsible for the incident. "Just received my ticket for the State of the Union. Looks like @BetsyDeVosEd was in charge of spell checking… #SOTUniom, " he wrote on Twitter.

It is easy to confuse the current geopolitical situation with that of the 1980s. The United States and Russia each accuse the other of interfering in domestic affairs. Russia has annexed territory over U.S. objections, raising concerns about military conflict. As during the Cold War after World War II, nations are developing and building weapons based on advanced technology. AI can also be used to control non-nuclear weapons including unmanned vehicles like drones and cyberweapons, the expert says. Unmanned vehicles must be able to operate while their communications are impaired – which requires onboard AI control.

NEW YORK – Telecommunications regulators and industry groups voiced opposition Monday to a government-built wireless network that the Trump administration is reportedly considering . The news website Axios reported Sunday that national security officials may want a government-built next-generation "5G" mobile network because of concerns about China and cybersecurity. A White House spokesman referred inquiries to the National Security Council, which did not immediately respond to questions. The telecom industry, which is powerful in Washington, is already working on 5G, which heralds better internet on smartphones as well as potential applications for self-driving cars and other new technology. The new standard is already being tested and could be widely available by 2020.

Adversarial examples are perturbed inputs designed to fool machine learning models. Adversarial training injects such examples into training data to increase robustness. To scale this technique to large datasets, perturbations are crafted using fast single-step methods that maximize a linear approximation of the model's loss. We show that this form of adversarial training converges to a degenerate global minimum, wherein small curvature artifacts near the data points obfuscate a linear approximation of the loss. The model thus learns to generate weak perturbations, rather than defend against strong ones. As a result, we find that adversarial training remains vulnerable to black-box attacks, where we transfer perturbations computed on undefended models, as well as to a powerful novel single-step attack that escapes the non-smooth vicinity of the input data via a small random step. We further introduce Ensemble Adversarial Training, a technique that augments training data with perturbations transferred from other models. On ImageNet, Ensemble Adversarial Training yields models with strong robustness to black-box attacks. In particular, our most robust model won the first round of the NIPS 2017 competition on Defenses against Adversarial Attacks (Kurakin et al., 2017c).