Attention & interpretability: modern neural networks are hard to interpret because we haven't built tools to make it easy to analyze their decision-making processes. Part of the reason why we haven't built the tools is that it's not entirely obvious how you get a big stack of perceptual math machinery to tell you about what it is thinking in a way that is remotely useful to the untrained eye. The best thing we've been able to come up with, in the case of certain vision and language tasks, is attention where we visualize what parts of a neural network – sometimes down to an individual cell or'neuron' within it – is activating in response to. This can help us diagnose why an AI tool is responding in the way it is. This component is general, working across different neural network architectures (a first, the researchers claim), and only requires the person to fiddle with it at its input or output points.

Every day countless headlines emerge from myriad sources across the globe, both warning of dire consequences and promising utopian futures – all thanks to artificial intelligence. AI "is transforming the workplace," writes the Wall Street Journal, while Fortune magazine tells us that we are facing an "AI revolution" that will "change our lives." But we don't really understand what interacting with AI will be like – or what it should be like. It turns out, though, that we already have a concept we can use when we think about AI: it's how we think about animals. As a former animal trainer (albeit briefly) who now studies how people use AI, I know that animals and animal training can teach us quite a lot about how we ought to think about, approach and interact with artificial intelligence, both now and in the future.

Artificial Intelligence (AI) is a computer science discipline working towards replicating critical human mental faculties, and it has reached a level of capability and maturity where it could have a transformative impact on healthcare and other knowledge intensive sectors. The power, scope and scale of AI applications are increasing exponentially. Here, we draw on key insights from our book on The Future of Business and our upcoming publication on The Future of AI in Business to explore the current uses, emerging applications and future possible impact of AI in healthcare. Finally, to offer a glimpse of the sheer scale of what is over the horizon, we present a scenario outlining the revolutionary impact AI could have on the central tenets and functioning of the UK healthcare system. The technological revolution is already disrupting the healthcare industry across the globe.

I haven't said much about Workey for a few months. It has raised another $8 million and is opening up a New York office, focusing on NY as well as on San Francisco. The Israeli company says it worked with Amazon, Dell EMC, Oracle, and Yahoo during its Beta. Among Workey's features is one called social power: "Workey will promote your anonymous profile based on your connections, and suggest which friends can vouch for you for each opportunity." I monkeyed around with Workey a bit, and one interesting feature is that next to job listings, it pops up little notes indicating you're reasonably qualified, like "your proven record on Twitter puts you ahead of the pack for this opportunity."

The announcement of Apple's HomePod smart speaker at last week's WWDC event marks the latest entrant to a growing market for voice assistants, currently led by tech giants Amazon and Google. Each new launch brings the promise of a slicker user experience and a more efficient use of our time. It might take several iterations of these voice assistants to integrate seamlessly into our daily routines, but the first stage for any company looking to take advantage of this new tech will be to ask: How do users feel about it all? We recently partnered with Mindshare Futures and J. Walter Thompson Innovation Group for their Speak Easy research project to answer that very question. Our portion of the study involved observing 102 smartphone users as they carried out a selection of tasks using Amazon's Alexa, Google Assistant, text-based search, and questions directed to a real person.

Multicopter drones - drones with more than two rotors - are finally able to fly in any orientation. Normally, multicopter drones can only fly parallel to the ground, but a new hexacopter drone with six propellers can tilt 360 degrees, allowing it to fly in any orientation. The drone, dubbed Voliro, can fly sideways, upside down, diagonally and in other orientations. The Voliro was developed by a team of 11 students at the Swiss Federal Institute of Technology (ETH Zurich) and Zurich University of the Arts (ZHDK), who spent 9 months developing a prototype of the drone. Because the drone can stay stable while flying in any configuration, it's able to fly parallel to walls.

Gboard update includes new enhancements to make typing and searching easier. Google has added some fun and helpful new features to Gboard, which highlight its latest machine learning capabilities. In a blog post, Google showed off a new doodle-search function for emojis, predictive phrase suggestions for texting, and enhanced results for searches made from within the keyboard. Look for the enhanced feature to pop up in the latest update for Gboard in the Google Play Store. The feature you're likely to use most often is the new predictive phrase feature, which is integrated into the same space as word predictions.

We consider the problem of approximate Bayesian parameter inference in non-linear state-space models with intractable likelihoods. Sequential Monte Carlo with approximate Bayesian computations (SMC-ABC) is one approach to approximate the likelihood in this type of models. However, such approximations can be noisy and computationally costly which hinders efficient implementations using standard methods based on optimisation and Monte Carlo methods. We propose a computationally efficient novel method based on the combination of Gaussian process optimisation and SMC-ABC to create a Laplace approximation of the intractable posterior. We exemplify the proposed algorithm for inference in stochastic volatility models with both synthetic and real-world data as well as for estimating the Value-at-Risk for two portfolios using a copula model. We document speed-ups of between one and two orders of magnitude compared to state-of-the-art algorithms for posterior inference.

Cost partitioning is a general method for adding multiple heuristic values admissibly. In the setting of optimal classical planning, saturated cost partitioning has recently been shown to be the cost partitioning algorithm of choice for pattern database heuristics found by hill climbing, systematic pattern database heuristics and Cartesian abstraction heuristics. To evaluate the synergy of the three heuristic types, we compute the saturated cost partitioning over the combined sets of heuristics and observe that the resulting heuristic is outperformed by the heuristic that simply maximizes over the three saturated cost partitioning heuristics computed separately for each heuristic type. Our new algorithm for choosing the orders in which saturated cost partitioning considers the heuristics allows us to compute heuristics outperforming not only the maximizing heuristic but even state-of-the-art planners.

Symmetry-based state space pruning techniques have proved to greatly improve heuristic search based classical planners. Similarly, abstraction heuristics in general and pattern databases in particular are key ingredients of such planners. However, only little work has dealt with how the abstraction heuristics behave under symmetries. In this work, we investigate the symmetry properties of the popular canonical pattern databases heuristic. Exploiting structural symmetries, we strengthen the canonical pattern databases by adding symmetric pattern databases, making the resulting heuristic invariant under structural symmetry, thus making it especially attractive for symmetry-based pruning search methods. Further, we prove that this heuristic is at least as informative as using symmetric lookups over the original heuristic. An experimental evaluation confirms these theoretical results.