When the new Star Wars video game, Battlefront II, was made public in a final testing session before a general release, it didn't receive quite the reception its publisher, Electronic Arts, was hoping for. It featured a confusing mixture of virtual collectibles and randomised rewards that could be used to unlock characters within the game, meaning it would take 40 hours of continuous play to access just one top-tier character such as Luke Skywalker. The system, though, could be shortcut with cash: players were able to spend real money buying so-called "loot crates" full of the required rewards and credits. Just a few thousand dollars was all it would take to unlock every character in the game. A bargain! What's more, these loot crates were also randomised, with users not knowing what they were getting before buying. The fan backlash to this system, on social news site Reddit and elsewhere, was immediate and furious.

The New Jersey bill would make operating a drone under the influence of alcohol a disorderly persons offense, which carries a sentence of up to six months in prison, a $1,000 fine or both. It also would make using a drone to hunt wildlife and endanger people or property a similar offense.

The multi-armed bandit (MAB) problem is a classic example of the exploration-exploitation dilemma. It is concerned with maximising the total rewards for a gambler by sequentially pulling an arm from a multi-armed slot machine where each arm is associated with a reward distribution. In static MABs, the reward distributions do not change over time, while in dynamic MABs, each arm's reward distribution can change, and the optimal arm can switch over time. Motivated by many real applications where rewards are binary counts, we focus on dynamic Bernoulli bandits. Standard methods like $\epsilon$-Greedy and Upper Confidence Bound (UCB), which rely on the sample mean estimator, often fail to track the changes in underlying reward for dynamic problems. In this paper, we overcome the shortcoming of slow response to change by deploying adaptive estimation in the standard methods and propose a new family of algorithms, which are adaptive versions of $\epsilon$-Greedy, UCB, and Thompson sampling. These new methods are simple and easy to implement. Moreover, they do not require any prior knowledge about the data, which is important for real applications. We examine the new algorithms numerically in different scenarios and find out that the results show solid improvements of our algorithms in dynamic environments.

Modeling interactions between features improves the performance of machine learning solutions in many domains (e.g. recommender systems or sentiment analysis). In this paper, we introduce Exponential Machines (ExM), a predictor that models all interactions of every order. The key idea is to represent an exponentially large tensor of parameters in a factorized format called Tensor Train (TT). The Tensor Train format regularizes the model and lets you control the number of underlying parameters. To train the model, we develop a stochastic Riemannian optimization procedure, which allows us to fit tensors with 2^160 entries. We show that the model achieves state-of-the-art performance on synthetic data with high-order interactions and that it works on par with high-order factorization machines on a recommender system dataset MovieLens 100K.

This paper introduces a probabilistic framework for k-shot image classification. The goal is to generalise from an initial large-scale classification task to a separate task comprising new classes and small numbers of examples. The new approach not only leverages the feature-based representation learned by a neural network from the initial task (representational transfer), but also information about the classes (concept transfer). The concept information is encapsulated in a probabilistic model for the final layer weights of the neural network which acts as a prior for probabilistic k-shot learning. We show that even a simple probabilistic model achieves state-of-the-art on a standard k-shot learning dataset by a large margin. Moreover, it is able to accurately model uncertainty, leading to well calibrated classifiers, and is easily extensible and flexible, unlike many recent approaches to k-shot learning.

An automated army of pro-Donald J. Trump chatbots overwhelmed similar programs supporting Hillary Clinton five to one in the days leading up to the presidential election, according to a report published Thursday by researchers at Oxford University. The chatbots -- basic software programs with a bit of artificial intelligence and rudimentary communication skills -- would send messages on Twitter based on a topic, usually defined on the social network by a word preceded by a hashtag symbol, like #Clinton. Their purpose: to rant, confuse people on facts, or simply muddy discussions, said Philip N. Howard, a sociologist at the Oxford Internet Institute and one of the authors of the report. If you were looking for a real debate of the issues, you weren't going to find it with a chatbot. "And a lot of what they pass around is false news."

Chatbots have invaded the workplace. Or at least, invaded the conversations around the digital workplace. In the last month, I have seen some outstanding presentations on the use of conversational interfaces (aka chatbots, aka bots) in the digital workplace/intranet context. While a number of interesting articles have been published on the topic in the last year, the presentations had a hands-on, "lets write a simple bot" element to them that made the abstract real. What follows is an overview of some of these presentations to get you thinking about how bots might benefit your digital workplace.

Pretend for a minute you're the captain of a ship that's being attacked by enemy cannons. Now – say something funny. Making up jokes on the spot is a real test of wits. Yet two comedians have developed an improv show in which many of the ad-libbed gags are delivered by a toy robot. In the last couple of years this unlikely comedy trio – known as HumanMachine – has performed 30 times to nearly 3000 people at comedy festivals in the UK and Canada.

Netherlands-based data intelligence company SciSports is hoping to change world football through data, motion tracking and machine learning. Using data and machine learning, the company produces a "SciSkill Index" – an objective ranking of current ability, potential and influence of thousands of footballers across hundreds of different competitions around the world. The score is determined by the SciSports' existing data library and from 3D data collected from stadium cameras, which converts movements in practice or during the match into useful information in real time. "It is the first system that allows you to compare James Troisi with Neymar and check if Milos Degenek has the potential to become as good as David Luiz," a company spokesperson told Which-50. "This will enable clubs to increase their scouting scope, decrease their risk of signing the wrong player and enlarge the change of finding the right talent."

The Basel committee rules have been an ongoing international response to the 2007-2009 financial crisis that saw the bankruptcy of U.S. investment bank Lehman Brothers and taxpayer bailouts of big banks. The financial crisis was the prelude to the Great Recession that saw many people lose their jobs and homes. Governments in the United States, Europe and elsewhere were pushed to rescue banks to prevent a cutoff of credit to businesses that would further harm the economy and increase unemployment.