A regularized risk minimization procedure for regression function estimation is introduced that achieves near optimal accuracy and confidence under general conditions, including heavy-tailed predictor and response variables. The procedure is based on median-of-means tournaments, introduced by the authors in [8]. It is shown that the new procedure outperforms standard regularized empirical risk minimization procedures such as lasso or slope in heavy-tailed problems.

A new short film from the Campaign Against Killer Robots warns of a future where weaponised flying drones target and assassinate certain members of the public, using facial recognition technology to identify them. Is this a realistic threat that could rightly spur an effective ban on the technology? Or is it an overblown portrayal designed to scare governments into taking simplistic, unnecessary and ultimately futile action? We asked two academics for their expert opinions. A new short film from the Campaign Against Killer Robots warns of a future where weaponised flying drones target and assassinate certain members of the public, using facial recognition technology to identify them.

With the rise of technology, so too comes the rise of a new category of intimacy. Digisexuals, or people who primarily use technology for sexual satisfaction, could soon become more prolific in society, according to experts. "It is safe to say the era of immersive virtual sex has arrived," said Neil McArthur, the director of the Center for Professional and Applied Ethics at the University of Manitoba and the author of a new scientific study on digisexuality. McArthur and a team from the university published a new report in the Journal of Sexual and Relationship Therapy detailing the need to be prepared for a rise in digisexuality. "As these technologies advance, their adoption will grow and many people will come to identify themselves as'digisexuals' – people whose primary sexual identity comes through the use of technology, McArthur said, according to the Telegraph. "Many people will find that their experiences with this technology become integral to their sexual identity and some will prefer them to direct sexual interactions with humans." Catalan nanotechnology engineer Sergi Santos holds the head of Samantha, a sex doll packed with artificial intelligence providing her the capability to respond to different scenarios and verbal stimulus, in his house in Rubi, north of Barcelona, Spain, Mar. Technology like virtual reality has become increasingly realistic and now includes more ways than ever to satisfy sexual appetites. Some pornography channels have begun to offer three-dimensional role-playing games. Perhaps most notably, realistic sex robots have become available to purchase. The company Abyss Creations created an "Android Love Doll" that boasts 50 different sexual positions and comes complete with an app that learns as the user interacts with it. Buyers are also able to handpick a doll to their tastes, altering hair color and other body parts. Many dolls have realistic silicone skin and some form of artificial intelligence that allow them to interact with a user. Others go as far as simulating an orgasm during sex. And while some companies sell such robots for more than $20,000, that price will likely decrease over time as technology becomes more accessible. "There is no question that sexbots are coming," he said. "People will form an intense connection with their robot companions.

Many self-driving car tests are conveniently run in warm, sunny climates where the road conditions are rarely less than ideal. But what about that significant chunk of the planet that gets snowfall? The Russian internet giant has started testing its autonomous Prius cars in winter conditions around Moscow's suburbs to see how they fare when snow obscures the roads and ice makes traction difficult. The video you see here is highly edited, but it suggests that the driverless machines are up to the job -- they can stay in their lanes, come to smooth stops and brake for pedestrians. These aren't the worst conditions a self-driving car could face.

The pick station will be trialled at the firm's high-tech depot in Andover, southern England, which already features other automated technologies.

Perhaps you remember the iconic theme of the globally popular Kung Fu Panda movies, "You are the secret ingredient!" This meant that self-belief is important and with it great things can be achieved--Po, for example, became the Dragon Warrior. My meaning here is that computer science is both a powerful enabler of rapid advances in all intellectual fields and a disruptor driving furious revolutions in commerce and society worldwide. Computer science is more important and potent than ever! Computing is driving unprecedented rapid change.

The brain of a honeybee is tiny -- the size of a pin head -- and contains less than a million neurons, compared to the 85 billion in our own brains. Yet with that sliver of brain, bees can do some extraordinary things. They can count and interpret abstract patterns. Most famously, bees have the ability to communicate the location of flowers to other bees in the hive. When a foraging bee has found a source of nectar and pollen, it can let others in the hive know by performing a peculiar figure-of-eight dance called the waggle dance.

In the present work we propose a Deep Feed Forward network architecture which can be trained according to a sequential learning paradigm, where tasks of increasing difficulty are learned sequentially, yet avoiding catastrophic forgetting. The proposed architecture can re-use the features learned on previous tasks in a new task when the old tasks and the new one are related. The architecture needs fewer computational resources (neurons and connections) and less data for learning the new task than a network trained from scratch

The study of networks has witnessed an explosive growth over the past decades with several ground-breaking methods introduced. A particularly interesting -- and prevalent in several fields of study -- problem is that of inferring a function defined over the nodes of a network. This work presents a versatile kernel-based framework for tackling this inference problem that naturally subsumes and generalizes the reconstruction approaches put forth recently by the signal processing on graphs community. Both the static and the dynamic settings are considered along with effective modeling approaches for addressing real-world problems. The herein analytical discussion is complemented by a set of numerical examples, which showcase the effectiveness of the presented techniques, as well as their merits related to state-of-the-art methods.

Learning distributed representations of documents has pushed the state-of-the-art in several natural language processing tasks and was successfully applied to the field of recommender systems recently. In this paper, we propose a novel content-based recommender system based on learned representations and a generative model of user interest. Our method works as follows: First, we learn representations on a corpus of text documents. Then, we capture a user's interest as a generative model in the space of the document representations. In particular, we model the distribution of interest for each user as a Gaussian mixture model (GMM). Recommendations can be obtained directly by sampling from a user's generative model. Using Latent semantic analysis (LSA) as comparison, we compute and explore document representations on the Delicious bookmarks dataset, a standard benchmark for recommender systems. We then perform density estimation in both spaces and show that learned representations outperform LSA in terms of predictive performance.