Here's how Puny Astronaut describes Skye on the back of the game's information card: "Glide through a gentle and charming world that couldn't be happier to see you." And it's true -- in Skye, there are no evil monsters out to destroy the world, no weapons to find, no traps to avoid and no enemies to slaughter. In fact, there's no way to lose Skye at all. This is a hug in video game form. The game itself involves traversing the land as a flying dragon, playing with giant suspended pianos, solving puzzles, soaring through fields of flowers and chatting with the townsfolk.

"I'm a nerd!" Jana Eggers tells us in a tone that suggests she is very much at ease with the description. Realistically, this airy room in the Berlin offices of the state of Baden-Württemberg, where a major conference on artificial intelligence (AI) took place on Wednesday and Thursday, is probably full of self-confessed nerds unlikely to be too upset by the moniker. Considering the tasks many of them have taken on in their professional lives -- the understanding and developing of artificial intelligence systems -- that brain power is needed. In effect, they are trying to build tech that mirrors the functioning of that most remarkable of natural organs, the human brain. Read more: Teachers for AI -- can robots create more jobs than they retire?

Robotic arms perform inner frame welds for 2018 Honda Accord vehicles during production at the Honda of America Manufacturing Inc. Marysville Auto Plant in Marysville, Ohio, U.S.. But humans are still an integral part of the assembly process. President Trump might think that the way to protect workers in the U.S. is to wage trade wars with countries that he believes are undercutting the prices of domestically-produced goods. But it is increasingly obvious that the real issue is the latest wave of automation. Of course, reports like those from the McKinsey Global Institute and Oxford University have been warning for a while that many of the jobs we know today are at risk of disappearing as artificial intelligence becomes more sophisticated and widespread.

This is a guest post. The views expressed in this article are solely those of the author and do not represent positions of IEEE Spectrum or the IEEE. Engineers are on the cusp of on-device machine learning, as evidenced by the first NIPS workshop on the subject in late 2017, and the advent of new neural processors, such as Kirin 970 from Huawei and Snapdragon 845 from Qualcomm. Thus far, progress in artificial intelligence has been fueled primarily by the availability of data and more computing power. Classical machine learning has mostly been built on a single central node (usually in a data center) with full access to a global dataset and a massive amount of storage and computing power.

Slides can be found here: https://buff.ly/2AvJ5eR Visit the conference website to learn more: www.datanatives.io Stay Connected to Data Natives by Email: Subscribe to our newsletter to get the news first about Data Natives 2017: http://bit.ly/1WMJAqS About the Author: Thomas Poetter, Founder & CEO of Compris Technologies AG, graduated in computer science from the University of Kaiserslautern with artificial intelligence and computational linguistics as key focus areas. Working with DFKI (German Research Center for Artificial Intelligence) during his studies, Thomas was one of the first entrepreneurs who received in 1998 a 2 years' startup stipend from the Fraunhofer society, in his case from the Institute for Experimental Software Engineering (IESE).

Legislators for the European Union have announced an investigation after allegations user data of 50 million Facebook accounts were misused. The investigation comes after Christopher Wylie, a whistle-blower who worked for data analytics company Cambridge Analytica, said on Saturday data of the 50 million users were harvested without their knowledge or consent. Antonio Tajani, president of the European Parliament, said on Twitter the allegations, if true, constitute "an unacceptable violation of our citizens' privacy rights". "The European Parliament will investigate fully, calling digital platforms to account," he said. Allegations of misuse of Facebook user data is an unacceptable violation of our citizens' privacy rights.

BBVA has started testing a facial recognition payment application for employees to use at cafeterias and restaurants at the Ciudad BBVA office and business conference complex in Madrid. The mobile app is powered through technology from the startup Veridas, which was a joint venture last year of BBVA and Das-Nano. Restaurant partner Sodexo Iberia is also involved in the project. More than 1,000 employees at Ciudad are using the payments app, which provides the ability to reserve tables and dine without dealing with a bill. Customers inform a waiter they are ready to leave, and the waiter verifies registration in the system and an account linked through facial recognition.

Even todays most advanced machine learning models are easily fooled by almost imperceptible perturbations of their inputs. Foolbox is a new Python package to generate such adversarial perturbations and to quantify and compare the robustness of machine learning models. It is build around the idea that the most comparable robustness measure is the minimum perturbation needed to craft an adversarial example. To this end, Foolbox provides reference implementations of most published adversarial attack methods alongside some new ones, all of which perform internal hyperparameter tuning to find the minimum adversarial perturbation. Additionally, Foolbox interfaces with most popular deep learning frameworks such as PyTorch, Keras, TensorFlow, Theano and MXNet and allows different adversarial criteria such as targeted misclassification and top-k misclassification as well as different distance measures. The code is licensed under the MIT license and is openly available at https://github.com/bethgelab/foolbox . The most up-to-date documentation can be found at http://foolbox.readthedocs.io .

We propose expected policy gradients (EPG), which unify stochastic policy gradients (SPG) and deterministic policy gradients (DPG) for reinforcement learning. Inspired by expected sarsa, EPG integrates across the action when estimating the gradient, instead of relying only on the action in the sampled trajectory. We establish a new general policy gradient theorem, of which the stochastic and deterministic policy gradient theorems are special cases. We also prove that EPG reduces the variance of the gradient estimates without requiring deterministic policies and, for the Gaussian case, with no computational overhead. Finally, we show that it is optimal in a certain sense to explore with a Gaussian policy such that the covariance is proportional to the exponential of the scaled Hessian of the critic with respect to the actions. We present empirical results confirming that this new form of exploration substantially outperforms DPG with the Ornstein-Uhlenbeck heuristic in four challenging MuJoCo domains.

Networks of coupled dynamical systems provide a powerful way to model systems with enormously complex dynamics, such as the human brain. Control of synchronization in such networked systems has far reaching applications in many domains, including engineering and medicine. In this paper, we formulate the synchronization control in dynamical systems as an optimization problem and present a multi-objective genetic programming-based approach to infer optimal control functions that drive the system from a synchronized to a non-synchronized state and vice-versa. The genetic programming-based controller allows learning optimal control functions in an interpretable symbolic form. The effectiveness of the proposed approach is demonstrated in controlling synchronization in coupled oscillator systems linked in networks of increasing order complexity, ranging from a simple coupled oscillator system to a hierarchical network of coupled oscillators. The results show that the proposed method can learn highly-effective and interpretable control functions for such systems.