We introduce a new family of margin-based regret guarantees for adversarial contextual bandit learning. Our results are based on multiclass surrogate losses. Using the ramp loss, we derive a universal margin-based regret bound in terms of the sequential metric entropy for a benchmark class of real-valued regression functions. The new margin bound serves as a complete contextual bandit analogue of the classical margin bound from statistical learning. The result applies to large nonparametric classes, improving on the best known results for Lipschitz contextual bandits (Cesa-Bianchi et al., 2017) and, as a special case, generalizes the dimension-independent Banditron regret bound (Kakade et al., 2008) to arbitrary linear classes with smooth norms. On the algorithmic side, we use the hinge loss to derive an efficient algorithm with a $\sqrt{dT}$-type mistake bound against benchmark policies induced by $d$-dimensional regression functions. This provides the first hinge loss-based solution to the open problem of Abernethy and Rakhlin (2009). With an additional i.i.d. assumption we give a simple oracle-efficient algorithm whose regret matches our generic metric entropy-based bound for sufficiently complex nonparametric classes. Under realizability assumptions our results also yield classical regret bounds.

Learning to infer Bayesian posterior from a few-shot dataset is an important step towards robust meta-learning due to the model uncertainty inherent in the problem. In this paper, we propose a novel Bayesian model-agnostic meta-learning method. The proposed method combines scalable gradient-based meta-learning with nonparametric variational inference in a principled probabilistic framework. During fast adaptation, the method is capable of learning complex uncertainty structure beyond a point estimate or a simple Gaussian approximation. In addition, a robust Bayesian meta-update mechanism with a new meta-loss prevents overfitting during meta-update. Remaining an efficient gradient-based meta-learner, the method is also model-agnostic and simple to implement. Experiment results show the accuracy and robustness of the proposed method in various tasks: sinusoidal regression, image classification, active learning, and reinforcement learning.

Modern AI and robotic systems are characterized by a high and ever-increasing level of autonomy. At the same time, their applications in fields such as autonomous driving, service robotics and digital personal assistants move closer to humans. From the combination of both developments emerges the field of AI ethics which recognizes that the actions of autonomous machines entail moral dimensions and tries to answer the question of how we can build moral machines. In this paper we argue for taking inspiration from Aristotelian virtue ethics by showing that it forms a suitable combination with modern AI due to its focus on learning from experience. We furthermore propose that imitation learning from moral exemplars, a central concept in virtue ethics, can solve the value alignment problem. Finally, we show that an intelligent system endowed with the virtues of temperance and friendship to humans would not pose a control problem as it would not have the desire for limitless self-improvement.

Semantic vectors are learned from data to express semantic relationships between elements of information, for the purpose of solving and informing downstream tasks. Other models exist that learn to map and classify supervised data. However, the two worlds of learning rarely interact to inform one another dynamically, whether across types of data or levels of semantics, in order to form a unified model. We explore the research problem of learning these vectors and propose a framework for learning the semantics of knowledge incrementally and online, across multiple mediums of data, via binary vectors. We discuss the aspects of this framework to spur future research on this approach and problem.

In this paper, we study the problem of learning vision-based dynamic manipulation skills using a scalable reinforcement learning approach. We study this problem in the context of grasping, a longstanding challenge in robotic manipulation. In contrast to static learning behaviors that choose a grasp point and then execute the desired grasp, our method enables closed-loop vision-based control, whereby the robot continuously updates its grasp strategy based on the most recent observations to optimize long-horizon grasp success. To that end, we introduce QT-Opt, a scalable self-supervised vision-based reinforcement learning framework that can leverage over 580k real-world grasp attempts to train a deep neural network Q-function with over 1.2M parameters to perform closed-loop, real-world grasping that generalizes to 96% grasp success on unseen objects. Aside from attaining a very high success rate, our method exhibits behaviors that are quite distinct from more standard grasping systems: using only RGB vision-based perception from an over-the-shoulder camera, our method automatically learns regrasping strategies, probes objects to find the most effective grasps, learns to reposition objects and perform other non-prehensile pre-grasp manipulations, and responds dynamically to disturbances and perturbations.

Ryan Calo's "law and Technology" Viewpoint "Is the Law Ready for Driverless Cars?" (May 2018) explored the implications, as Calo said, of " ... genuinely unforeseeable categories of harm" in potential liability cases where death or injury is caused by a driverless car. He argued that common law would take care of most other legal issues involving artificial intelligence in driverless cars, apart from such "foreseeability." Calo also said the courts have worked out problems like AI before and seemed confident that AI foreseeability will eventually be accommodated. One can agree with this overall judgment but question the time horizon. AI may be quite different from anything the courts have seen or judged before for many reasons, as the technology is indeed designed to someday make its own decisions.

If Elon Musk doesn't like something, he'll create his own version. That's exactly what he's done for his children's education by starting a radical ultra-exclusive school at his SpaceX headquarters in Hawthorne, California. For the past four years, the non-profit'experimental' school has been educating the billionaire's five sons, children of some SpaceX employees and a number of gifted students from Los Angeles. The school has some unconventional teaching methods. Reports suggest it allows students to skip subjects they don't like, build flamethrowers and'defeat evil AIs'.

On April 12, at offices of the healthcare incubator MATTER at the Merchandise Mart in downtown Chicago, as well as streamed live online, Andrew Burt, chief privacy officer and legal engineer at the data management company Immuta, delivered a lecture entitled "Regulating Artificial Intelligence: How to Control the Unexplainable" in which he focused on the ethical, legal, and regulatory issues surrounding the deployment of machine learning systems. Sponsored by all three UChicago Graham School Professional Masters degree programs--Biomedical Informatics (MScBMI), Analytics (MScA), and Threat and Response Management (MScTRM)--the catalyst for the occasion was Sam Volchenboum, MD, PhD, MS, director of the Center for Research Informatics at UChicago, and faculty director for the BMI program, whose encounter with Burt at a recent South by Southwest conference led to an exchange of ideas he saw as immediately relevant to the Graham School programs. "As a physician, I'm seeing the use of machine learning algorithms all over the hospital and all over medicine," Dr. Volchenboum said. "We just plow ahead with developing our models and our predictions. But it wasn't until I spoke with Andrew that I really stopped and thought about the implications of these algorithms and how they can be used in both good and also bad ways. It was an eye-opening experience and I've been really excited about bringing Andrew here to talk ever since."

Earlier this year, Mattel announced that it was partnering with Tynker to bring Barbie-themed coding lessons to young kids. As of today, six free coding experiences are now available as is a new STEM-themed doll -- Robotics Engineer Barbie. The lessons are geared towards beginners, kindergarten-aged and older, and aim to teach logic, problem-solving and the basics of coding. While they learn, kids can also take on different career roles alongside Barbie, including musician, astronaut, pastry chef, robotics engineer, farmer and beekeeper. "Our mission is to empower youth to become the makers of tomorrow through coding, and the Barbie brand is an ideal partner to help us introduce programming to a large number of kids in a fun, engaging way," Tynker CEO Krishna Vedati said in a statement.

Taught by a Stanford-educated, ex-Googler and an IIT, IIM – educated ex-Flipkart lead analyst. This team has decades of practical experience in quant trading, analytics and e-commerce. This course is a gentle yet thorough introduction to Data Science, Statistics and R using real life examples. If you want to learn how to start building professional, career-boosting mobile apps and use Machine Learning to take things to the next level, then this course is for you. The Complete iOS Machine Learning Masterclass is the only course that you need for machine learning on iOS.