What if you and your wearable could have a conversation? Not a conversation involving command prompts or specific phrases, but a natural, fluid discussion about the weather, or what kind of run you should go on today. That's the idea behind Asteria: claiming to be'True AI', which learns everything about you but doesn't just crunch numbers to throw them back at your face. Rather, it turns the data into personalized pieces of information relevant to you - through a conversation. It sounds like a film or novel, something romanticized and created to whet our appetites for genuine AI, but Asteria co-founders CEO Dan Gailey and COO Nathan Ross believe it's possible - and that they're well on their way to making it real.

As a person who despises the cold, there is nothing worse than being home on a frigid winter night and having your utility company turn off the gas for repairs. Thankfully, robots are here to save us from having to bundle up like the Michelin Man just to stay toasty in our own homes. ULC Robotics has developed a robot, called Cirris XR, that's able to work on live gas mains so utilities don't have to turn off service to do repairs. Using the robot can also increase efficiency and operator safety. According to the company, gas utilities using Cirris XR would be able to repair five or six joints a day versus the one or two joints a human could work on.

James Lovelock's parting words last time we met were: "Enjoy life while you can. Because if you're lucky, it's going to be 20 years before it hits the fan." It was early 2008, and the distinguished scientist was predicting imminent and irreversible global warming, which would soon make large parts of the planet uninhabitably hot or put them underwater. The fashionable hope that windfarms or recycling could prevent global famine and mass migration was, he assured me, a fantasy; it was too late for ethical consumption to save us. Before the end of this century, 80% of the world's population would be wiped out.

This paper considers online convex optimization over a complicated constraint set, which typically consists of multiple functional constraints and a set constraint. The conventional projection based online projection algorithm (Zinkevich, 2003) can be difficult to implement due to the potentially high computation complexity of the projection operation. In this paper, we relax the functional constraints by allowing them to be violated at each round but still requiring them to be satisfied in the long term. This type of relaxed online convex optimization (with long term constraints) was first considered in Mahdavi et al. (2012). That prior work proposes an algorithm to achieve $O(\sqrt{T})$ regret and $O(T^{3/4})$ constraint violations for general problems and another algorithm to achieve an $O(T^{2/3})$ bound for both regret and constraint violations when the constraint set can be described by a finite number of linear constraints. A recent extension in Jenatton et al. (2016) can achieve $O(T^{\max\{\beta,1-\beta\}})$ regret and $O(T^{1-\beta/2})$ constraint violations where $\beta\in (0,1)$. The current paper proposes a new simple algorithm that yields improved performance in comparison to prior works. The new algorithm achieves an $O(\sqrt{T})$ regret bound with finite constraint violations.

During my commute the other day, I ended up on a dark subway car. The train still had power--the air conditioning was on, the announcements were coming through--but all the lights were dead. I live near an above-ground stop, so at first there was morning sunshine coming in through the windows. But when the train went underground, we were plunged into complete darkness. I found myself suddenly in a sea of floating, ghostly faces, illuminated by the glow of smartphones.

Apple is in the final stages of constructing a mammoth new campus in California to house the bulk of its workers. It's been lovingly called the "Spaceship" for years, and the latest drone flyover of the site shows you why. As per usual, two drone pilots have been conducting flyovers of the facility using DJI Inspire 1 and Phantom 4 quadcopters. They show an outer building that's mostly done, with details like landscaping and solar panels being finished at speed. The first video, filmed by Duncan Sinfield with his Inspire 1, does a good job of showing you around campus.

CHINA has long been seen as the "world's factory," churning out vast mountains of low-quality goods, but it is also considered a nation incapable of producing innovative products and ideas. Now, this is beginning to change -- China is closing the innovation gap. From drones to artificial intelligence, the Internet to genetic engineering, innovative Chinese companies are leading global innovation and reshaping the country's technology and business landscape. "There is a huge market with great opportunity," said Dai Xiang, co-founder and general manager of Enpower Energy, a manufacturer of aqueous ion batteries, which are cleaner, safer and more cost-effective than lead-acid batteries. After 20 years in the United States, working as a senior manager for start-ups in Silicon Valley, Dai decided to go back to China and start his own business.

Will "Miss Peregrine's Home for Peculiar Children" have a fairy-tale ending at the box office? While its final chapter has yet to be written, Tim Burton's fantasy film is earning pretty good grades at the multiplex so far. The picture about a group of extraordinary children collected 9 million on Friday, according to an estimate from distributor 20th Century Fox. That means the movie is on track to gross around 27 million by weekend's end -- a so-so start, considering the picture cost the studio 110 million to make. The weekend's other big debut, "Deepwater Horizon," lagged slightly behind in ticket sales Friday, with 7.1 million.

Last year, Disney created the VertiGo, a robot that can climb walls. I wanted to take that concept further, with a powered-up version of the drone, strong enough to carry a person. The Ventooz lends itself to a variety of uses. It could carry people who inspect or repair buildings and structures like dams. It could serve as a rescue vehicle stranded in dangerous situations.

The number of methods available for classification of multi-label data has increased rapidly over recent years, yet relatively few links have been made with the related task of classification of sequential data. If labels indices are considered as time indices, the problems can often be seen as equivalent. In this paper we detect and elaborate on connections between multi-label methods and Markovian models, and study the suitability of multi-label methods for prediction in sequential data. From this study we draw upon the most suitable techniques from the area and develop two novel competitive approaches which can be applied to either kind of data. We carry out an empirical evaluation investigating performance on real-world sequential-prediction tasks: electricity demand, and route prediction. As well as showing that several popular multi-label algorithms are in fact easily applicable to sequencing tasks, our novel approaches, which benefit from a unified view of these areas, prove very competitive against established methods. Keywords: multi-label classification; problem transformation; sequential data; sequence prediction; Markov models 1. Introduction Multi-label classification is the supervised learning problem where an instance is associated with multiple class variables (i.e., labels), rather than with a single class, as in traditional classification problems. See [1] for a review. Corresponding author, jesse.read@polytechnique.edu Preprint submitted to Pattern Recognition September 29, 2016 labels were modelled independently - at the expense of an increased computational cost. The case of binary labels is most common, where a positive class value denotes the relevance of the label (and the negative or null class denotes irrelevance). Typical examples of binary multi-label classification involve categorizing text documents and images, which can be assigned any subset of a particular label set. For example, an image can be associated with both labels beach and sunset. The multi-label classification paradigm has been successfully considered also in many other domains, such as text, video, audio, and bioinformatics - see [1] and references therein for further examples.