MOSUL, Iraq--Clashes between advancing Iraqi forces and ISIS militants in Mosul's Intisar and Aden neighborhoods continued on Sunday as the troops carried out a house-to-house search for remaining militants and explosives in liberated quarters. Nobody noticed the heist at the time, but they were caught – after the event. Myerson reckons intelligent agents could have prevented it happening in the first place. Glesni Holland reports from World of Watson 2016 in Las Vegas, where IBM executives unveiled the latest innovations in cognitive computing and demonstrated the impact that Watson can have – and is already having – on some of the world's most important industries. Following the successful launch of Elder Scroll 5: Skyrim Special Edition a few weeks ago, fans are too eager to stop speculating about what to expect from the upcoming game, The Elder Scroll 6.

Infosys has made an investment of R14.5 crore in a Denmark-based artificial intelligence start-up, Unsilo. Founded in 2012, this Danish company is focused on solutions in the area of advanced text analysis. Infosys has earmarked $500 million for its innovation fund which invests in start-ups across the globe. Of this, $250 million has been set aside for start-ups in India. "We will partner with Unsilo to bring their artificial intelligence and machine learning technology to our global clients. They join and expanding portfolio of innovative young companies from around the world that Infosys works with to help enterprises drive their digital transformation," said Ritika Suri, executive vice president & global head of corporate development & ventures at Infosys.

For millennia, we surpassed the other intelligent species with which we share our planet--dolphins, porpoises, orangutans, and the like--in almost all skills, bar swimming and tree-climbing. In recent years, though, our species has created new forms of intelligence, able to outperform us in other ways. One of the most famous of these artificial intelligences (AIs) is AlphaGo, developed by Deepmind. In just a few years, it has learned to play the 4,000-year-old strategy game, Go, beating two of the world's strongest players. Other software developed by Deepmind has learned to play classic eight-bit video games, notably Breakout, in which players must use a bat to hit a ball at a wall, knocking bricks out of it.

The first self-driving cars to be operated by ordinary British drivers will be left deliberately unmarked so that other drivers will not be tempted to "take them on", a senior car industry executive has revealed. One of the biggest fears of an ambitious project to lease the first autonomous vehicles to everyday motorists is that other road users might slam on their brakes or drive erratically in order to force the driverless cars into submission, he said. This is why the first 100 self-driving 4x4 vehicles to be leased to motorists as part of a pilot scheme on busy main roads into London will look no different than other Volvos of the same model, said Erik Coelingh, senior technical leader at Volvo Cars. The scheme will start in 2018. "From the outside you won't see that it's a self-driving car. From a purely scientific perspective it would be interesting to have some cars that are marked as self-driving cars and some that are not and see whether other road users react in a different way," Coelingh told the Observer.

This is the text version of a talk I gave on November 11, 2016, at the Direction conference in Sydney. When John Allsopp invited me here, I told him how excited I was discuss a topic that's been heavy on my mind: accountability in automated systems. But then John explained that in order for the economics to work, and for it to make sense to fly me to Australia, there needed to actually be an audience. Let's start with the most obvious answer--the military. This is the Predator, the forerunner of today's aerial drones. Those things under its wing are Hellfire missiles. These two weapons are the chocolate and peanut butter of robot warfare. In 2001, CIA agents got tired of looking at Osama Bin Laden through the camera of a surveillance drone, and figured out they could strap some missiles to the thing. And now we can't build these things fast enough. We're now several generations in to this technology, and soldiers now have smaller, portable UAVs they can throw like a paper airplane. You launch them in the field, and they buzz around and give you a safe way to do reconaissance. There are also portable UAVs with explosives in their nose, so you can fire them out of a tube and then direct them against a target--a group of soldiers, an orphanage, or a bunker–and make them perform a kamikaze attack. The Army has been developing unmanned vehicles that work on land, little tanks that roll around with a gun on top, with a wire attached for control, like the cheap remote-controlled toys you used to get at Christmas. Here you see a demo of a valiant robot dragging a wounded soldier to safety. The Russians have their own versions of these things, of course. I imagine it asking you who you are in a heavy Slavic accent before firing its many weapons into your fleeing body. Not all these robots are intended as weapons. The Army is trying to automate transportation, sometimes in weird-looking ways like this robotic dog monster.

Over the past few years, neural networks have re-emerged as powerful machine-learning models, yielding state-of-the-art results in fields such as image recognition and speech processing. More recently, neural network models started to be applied also to textual natural language signals, again with very promising results. This tutorial surveys neural network models from the perspective of natural language processing research, in an attempt to bring natural-language researchers up to speed with the neural techniques. The tutorial covers input encoding for natural language tasks, feed-forward networks, convolutional networks, recurrent networks and recursive networks, as well as the computation graph abstraction for automatic gradient computation.

We introduce an Embarrassingly Parallel Search (EPS) method for solving constraint problems in parallel, and we show that this method matches or even outperforms state-of-the-art algorithms on a number of problems using various computing infrastructures. EPS is a simple method in which a master decomposes the problem into many disjoint subproblems which are then solved independently by workers. Our approach has three advantages: it is an efficient method; it involves almost no communication or synchronization between workers; and its implementation is made easy because the master and the workers rely on an underlying constraint solver, but does not require to modify it. This paper describes the method, and its applications to various constraint problems (satisfaction, enumeration, optimization). We show that our method can be adapted to different underlying solvers (Gecode, Choco2, OR-tools) on different computing infrastructures (multi-core, data centers, cloud computing). The experiments cover unsatisfiable, enumeration and optimization problems, but do not cover first solution search because it makes the results hard to analyze. The same variability can be observed for optimization problems, but at a lesser extent because the optimality proof is required. EPS offers good average performance, and matches or outperforms other available parallel implementations of Gecode as well as some solvers portfolios. Moreover, we perform an in-depth analysis of the various factors that make this approach efficient as well as the anomalies that can occur. Last, we show that the decomposition is a key component for efficiency and load balancing.

Abstract--A number of important applied problems in engineering, finance and medicine can be formulated as a problem of anomaly detection based on a one-class classification. A classical approach to this problem is to describe a normal state using a one-class support vector machine. Then to detect anomalies we quantify a distance from a new observation to the constructed description of the normal class. In this paper we present a new approach to one-class classification. We formulate a new problem statement and a corresponding algorithm that allow taking into account privileged information during the training phase. We evaluate performance of the proposed approach using synthetic datasets, as well as the publicly available Microsoft Malware Classification Challenge dataset. Anomaly detection refers to the problem of finding patterns in data that do not conform to an expected behaviour.

We invite applications for a joint postdoctoral position at the GRAPPA center of excellence in Astroparticle Physics at the U. of Amsterdam, and at the Lorentz Institute at the U. of Leiden. Preference will be given to candidates with expertise in the application of advanced statistical methods and/or machine learning, to astronomy and/or particle physics. The appointment is for two years with a starting date in the Fall 2017. Instructions to apply: Interested candidates should submit the application material through the general GRAPPA postdoctoral search at https://academicjobsonline.org/ajo/jobs/8324.

Data Science for Internet of Things provides a concise introduction to the application of Predictive learning algorithms to Internet of Things. This mini book is based on my teaching at Oxford University, UPM(University of Madrid) and also working with consulting clients.We first outline the key issues involved and then explores three key areas: Stream processing, Deep Learning and Sensor fusion for IoT. The book is also a recommended material for the Stanford University course: Building a Successful Business for the Internet of Things and Mobile (BUS20) $11.99