This paper presents eight PAC-Bayes bounds to analyze the generalization performance of multi-view classifiers. These bounds adopt data dependent Gaussian priors which emphasize classifiers with high view agreements. The center of the prior for the first two bounds is the origin, while the center of the prior for the third and fourth bounds is given by a data dependent vector. An important technique to obtain these bounds is two derived logarithmic determinant inequalities whose difference lies in whether the dimensionality of data is involved. The centers of the fifth and sixth bounds are calculated on a separate subset of the training set. The last two bounds use unlabeled data to represent view agreements and are thus applicable to semi-supervised multi-view learning. We evaluate all the presented multi-view PAC-Bayes bounds on benchmark data and compare them with previous single-view PAC-Bayes bounds. The usefulness and performance of the multi-view bounds are discussed.

Video Friday is your weekly selection of awesome robotics videos, collected by your edgy Automaton bloggers. We'll also be posting a weekly calendar of upcoming robotics events for the next two months; here's what we have so far (send us your events!): Let us know if you have suggestions for next week, and enjoy today's videos. IHMC has managed to get their ATLAS balancing on the edge of cinder blocks, balancing itself with outstretched arms as it does so. The robot is able to detect and explore partial footholds (in this case line contacts).

Google's secretive AI division is working on a'big red button' that can halt its artificial intelligence software. Researchers have previously warned that AI could threaten humanity, with doomsday scenarios of AIs taking over, with one expert involved in the new paper admitting Google's historic win over Go world champion proves AI can be'unpredictable and immoral'. Now the DeepMind team say they have the answer - an off switch. Google's DeepMind team say AI agents are'unlikely to behave optimally all the time' and have called for'safe interruptibility' to be built into systems. Google has set up an ethics board to oversee its work in artificial intelligence.

Shipping giant Maersk Group made headlines earlier this year when it used a drone to deliver a bucket of cookies to a tanker at sea. That was just the beginning of an effort the company thinks could dramatically change its business. The company is evaluating ways to expand its use of drones and plans a bigger test later this year, said Markus Kuhn, a supply chain manager at Maersk, at the Drones Data X conference in San Francisco. In January, the company made a drone fly 250 meters from one of its barges to a tanker and drop off a batch of cookies. It's now looking for a drone-making partner for a test flight that would haul a 10-kilogram package for 10 kilometers. That test would illustrate a key use case for Maersk, that of resupplying its fleet of tankers, oil platforms, container ships and other craft.

Almost two decades ago, Google was just a search engine with a textbox and a promising algorithm that curated the ever growing internet into a list of blue links. As technology paced further, the Mountain View based company invested in a bunch of rising platforms and ideas that revolved around the future of the World Wide Web. Spreading across soon enough, their growth skyrocketed eliminating major tech leaders from the market. Moreover, industry's dependencies grew more on Google when they inhibited the responsibility of improving 80% of the smartphones around the globe. However, in the past year or so, smartphone manufacturers have been unable to maintain the wow factor in their products.

"About 6,000 languages are currently spoken in the world today," says Elizabeth Salesky of MIT Lincoln Laboratory's Human Language Technology (HLT) Group. "Within the law enforcement community, there are not enough multilingual analysts who possess the necessary level of proficiency to understand and analyze content across these languages," she continues. This problem of too many languages and too few specialized analysts is one Salesky and her colleagues are now working to solve for law enforcement agencies, but their work has potential application for the Department of Defense and Intelligence Community. The research team is taking advantage of major advances in language recognition, speaker recognition, speech recognition, machine translation, and information retrieval to automate language processing tasks so that the limited number of linguists available for analyzing text and spoken foreign languages can be used more efficiently. "With HLT, an equivalent of 20 times more foreign language analysts are at your disposal," says Salesky.

And pretty soon, they'll come for us. That seems to be the story today, whether from Hollywood or in breathless articles in popular tech magazines about artificial intelligence and nanotechnology. In a world where machines can learn, once humans push the "on" button, there's no stopping our robot overlords, right? When machines become more intelligent, humans are freed to become more creative. That opens doors to completely new possibilities.

Awesome Con is a place to celebrate comic books, movies, television, toys, and games – and beginning this year, science and technology are joining the party at the inaugural Awesome Con Science Fair presented by Science Channel! Maybe you wanna learn more about cloning. You're probably terrified of robotics. The worlds of science fiction and science fact overlap, and we're excited to now expand Awesome Con and create new awareness, advocacy, and interest in all things nerdy, techy, wonky, and smart! The Awesome Con Science Fair presented by Science Channel's participants include Science Channel (obviously), NASA, the Smithsonian, the Department of Energy, the National Institute of Standards and Technology, the National Nanotechnology Coordination Office, the Society for Science and the Public, and Nerd Nite!

Unlike robotic AIs that are being built to emulate the human brain, UNU works with existing human intelligence instead of replicating it. The platform, which is open to the public, allows a group of people to converge on an answer in real time. While a swarm of seven predicted the Oscars, the Derby decision came from 20 people. For the AMA at 1 p.m. EDT today, the group that will make political predictions is expected to range from 100 to 200 people. The participants will come from UNU's user base, which has already answered fantasy football and cooking queries.

DeepMind, Google's artificial intelligence company, catapulted itself into fame when its AlphaGo AI beat the world champion of Go, Lee Sedol. However, DeepMind is working to do a lot more than beat humans at chess and Go and various other games. Indeed, its AI algorithms were developed for something far greater: To "solve intelligence" by creating general purpose AI that can be used for a host of applications and, in essence, learn on their own. This, of course, raises some concerns. Namely, what do we do if the AI breaks…if it gets a virus…if it goes rogue?