New tentative contract for Canada's scientists would enshrine right to speak to media Rouge National Urban Park, nearly 95% complete, becomes North America's largest urban park Academics join outcry sparked by Hong Kong's contentious extradition bill Africa's science academy leads push for ethical data use Attend the Journées de la relève en recherche to get tips and tricks to succeed in your graduate studies or to enter the job market!

There is great potential for the use of Artificial intelligence (AI) by cities. With the help of AI, we can provide more responsive services to citizens. However, AI poses ethical issues that need special attention, as Chief Digital Officers of London and Helsinki here we explain why and suggest an approach for city government. The use of automation and machine learning systems is not a new phenomenon in public administration, but it is being transformed in how it is being used -- from automating simple transactions to more complex problem-solving. Today we see adoption across a range of municipal services -- chatbots in customer services, prioritisation of housing repairs, traffic signalling, demand-responsive transport, even library book management systems.

Artificial intelligence is central to the ongoing tech revolution, and it's getting smarter all the time. The driving force behind computer vision, speech analysis and natural language processing, AI impacts industry and society in numerous ways -- and will continue to do so far into the future. It's no surprise, then, that the AI field is rife with career opportunities -- so many of them, in fact, that the sector now faces a unique challenge: There are too many jobs and too few qualified candidates. On the up side, that means it offers virtually guaranteed (and well-paying) employment for those who've got the goods. So how does one get into AI, and what does an artificial intelligence career path look like? We asked some of the field's top experts to share insights from their journey to help guide the way.

With the lag of tech in healthcare, will AI/ML improve patient care or remain a smart idea? Technology experts have promised artificial intelligence (AI) and machine learning (ML) will revolutionize healthcare. Applications have the potential to streamline workflows and reduce human errors, speeding drug discovery, assisting surgery, and provisioning better billing and coding methods. But, in an industry that typically lags in digital maturity by as much as 10 years, according to a 2017 study, is AI in healthcare an empty promise or truly a forward-thinking and innovative reality? Technology experts have promised artificial intelligence (AI) and machine learning (ML) will revolutionize healthcare.

WASHINGTON/DUBAI, UNITED ARAB EMIRATES - U.S. President Donald Trump played down Iran's downing of a U.S. military surveillance drone on Thursday, saying he suspected it was shot by mistake and "it would have made a big difference" to him had the remotely controlled aircraft been piloted. While the comments appeared to suggest Trump was not eager to escalate the latest in a series of incidents with Iran, he also warned: "This country will not stand for it." Tehran said the unarmed Global Hawk surveillance drone was on a spy mission over its territory, but Washington said it was shot down over international airspace. "I think probably Iran made a mistake -- I would imagine it was a general or somebody that made a mistake in shooting that drone down," Trump told reporters at the White House. "We had nobody in the drone. It would have made a big difference, let me tell you, it would have made a big, big difference" if the aircraft had been piloted, Trump said as he met Canadian Prime Minister Justin Trudeau in the Oval Office.

Trump calls the strike a'foolish move'; national security correspondent Jennifer Griffin reports. It's time for the U.S. to take military action against Iran – not to start a war, but to blow some things up in retaliation for Iran shooting down a U.S. surveillance drone Thursday in international air space, just days after setting off explosives that damaged two oil tankers. President Trump gave Iran a pass after the recent tanker attacks. But instead of reassessing their strategy and trying to de-escalate tensions, the Iranians escalated significantly by shooting down the American drone – a high-flying unmanned aircraft that costs about $130 million. I don't see how President Trump can let Iran's latest attack pass without action if he expects Iran and other nations to respect the U.S. and not conclude they can attack our forces at will, without fear of retaliation.

Embedding models, which learn latent representations of users and items based on user-item interaction patterns, are a key component of recommendation systems. In many applications, contextual constraints need to be applied to refine recommendations, e.g. when a user specifies a price range or product category filter. The conventional approach, for both context-aware and standard models, is to retrieve items and apply the constraints as independent operations. The order in which these two steps are executed can induce significant problems. For example, applying constraints a posteriori can result in incomplete recommendations or low-quality results for the tail of the distribution (i.e., less popular items). As a result, the additional information that the constraint brings about user intent may not be accurately captured. In this paper we propose integrating the information provided by the contextual constraint into the similarity computation, by merging constraint application and retrieval into one operation in the embedding space. This technique allows us to generate high-quality recommendations for the specified constraint. Our approach learns constraints representations jointly with the user and item embeddings. We incorporate our methods into a matrix factorization model, and perform an experimental evaluation on one internal and two real-world datasets. Our results show significant improvements in predictive performance compared to context-aware and standard models.

Stochastic gradient descent (SGD) has been widely used in machine learning due to its computational efficiency and favorable generalization properties. Recently, it has been empirically demonstrated that the gradient noise in several deep learning settings admits a non-Gaussian, heavy-tailed behavior. This suggests that the gradient noise can be modeled by using $\alpha$-stable distributions, a family of heavy-tailed distributions that appear in the generalized central limit theorem. In this context, SGD can be viewed as a discretization of a stochastic differential equation (SDE) driven by a L\'{e}vy motion, and the metastability results for this SDE can then be used for illuminating the behavior of SGD, especially in terms of `preferring wide minima'. While this approach brings a new perspective for analyzing SGD, it is limited in the sense that, due to the time discretization, SGD might admit a significantly different behavior than its continuous-time limit. Intuitively, the behaviors of these two systems are expected to be similar to each other only when the discretization step is sufficiently small; however, to the best of our knowledge, there is no theoretical understanding on how small the step-size should be chosen in order to guarantee that the discretized system inherits the properties of the continuous-time system. In this study, we provide formal theoretical analysis where we derive explicit conditions for the step-size such that the metastability behavior of the discrete-time system is similar to its continuous-time limit. We show that the behaviors of the two systems are indeed similar for small step-sizes and we identify how the error depends on the algorithm and problem parameters. We illustrate our results with simulations on a synthetic model and neural networks.

To whom correspondence should be addressed; Email: b.roads@ucl.ac.uk. One Sentence Summary: The meaning of concepts resides in relationships across encompassing systems that each provide a window on a shared reality. Abstract Concept induction requires the extraction and naming of concepts from noisy perceptual experience. For supervised approaches, as the number of concepts grows, so does the number of required training examples. Philosophers, psychologists, and computer scientists, have long recognized that children can learn to label objects without being explicitly taught. In a series of computational experiments, we highlight how information in the environment can be used to build and align conceptual systems. Unlike supervised learning, the learning problem becomes easier the more concepts and systems there are to master. The key insight is that each concept has a unique signature within one conceptual system (e.g., images) that is recapitulated in other systems (e.g., text or audio). As predicted, children's early concepts form readily aligned systems. A typical person can correctly recognize and name thousands of objects. However, it remains unclear what mechanism makes this feat possible.

Rewards and punishments in different forms are pervasive and present in a wide variety of decision-making scenarios. By observing the outcome of a sufficient number of repeated trials, one would gradually learn the value and usefulness of a particular policy or strategy. However, in a given environment, the outcomes resulting from different trials are subject to chance influence and variations. In learning about the usefulness of a given policy, significant costs are involved in systematically undertaking the sequential trials; therefore, in most learning episodes, one would wish to keep the cost within bounds by adopting learning stopping rules. In this paper, we examine the deployment of different stopping strategies in given learning environments which vary from highly stringent for mission critical operations to highly tolerant for non-mission critical operations, and emphasis is placed on the former with particular application to aviation safety. In policy evaluation, two sequential phases of learning are identified, and we describe the outcomes variations using a probabilistic model, with closedform expressions obtained for the key measures of performance. Decision rules that map the trial observations to policy choices are also formulated. In addition, simulation experiments are performed, which corroborate the validity of the theoretical results.