How healthcare has evolved from the first clinically useful image to a library of images analyzed by AI In August 1980, a team from Scotland made a breakthrough in imaging. Setting the stage for the widespread use of MRI scans, they obtained the first clinically useful image of a patient's internal tissues. Almost 30 years later, breakthroughs in imaging are becoming the normal. While there are juxtaposed views around the potential of the technology, both skeptics and supporters know there's transformative potential. That's why one hospital system is pinpointing what's been holding AI back and developing the business model, platform and tools to ensure clinicians and patients can benefit from its potential. "Instead of building AI solutions in isolation, we should think about the technology the way we are growing to think about patient care – as a continuum, spanning care areas and disease states," said Mark Michalski, MD, Executive Director of the Massachusetts General Hospital and Brigham and Women's Hospital Center for Clinical Data Science.

To its devotees, it is an ancient practice used successfully for centuries to find water. And while sceptics say there is no scientific proof dowsing works, at least one important group of experts appears to be prepared to give it a try – Britain's water companies. They admit they still allow their engineers to use rods to divine for mains pipes. A Severn Trent water company technician was spotted using dowsing rods as a means to locate underground water. It has emerged many companies' engineers also use the method One company even said that'the older tried and tested methods are just as effective' as modern techniques using drones and satellites.

Microsoft's Seeing AI app, which helps blind and partially sighted people by narrating the world around them, has been released in the UK. The free program uses artificial intelligence to recognise objects, people and text via a phone or tablet's camera and describes them to the user. Seeing AI, an ongoing research project from Microsoft, is designed to help people with vision impairments complete everyday tasks and offer new levels of independence. According to the Royal National Institute of Blind People (RNIB), more than two million people in the UK live with sight loss, and almost half of blind and partially sighted people feel "moderately" or "completely" cut off from people and things around them. The RNIB estimates that sight loss costs the UK economy more than £4.3 billion in indirect costs, such as unpaid carer costs and reduced employment rates.

AI could be one of the biggest commercial opportunities for Ireland. Artificial intelligence (AI) is forecast to boost Irish GDP by 11.6pc or the equivalent of an extra €48bn, according to new research by PwC. The consulting giant recommends that the effect on jobs in the long term will at least be neutral, if not net positive, but this depends on employers putting Ireland at the forefront of the AI revolution by investing in skills and technology. 'Put an action plan in place around AI and manage with the same discipline you would put around any technology-enabled transformation. Don't wait for it to happen around you' – RONAN FITZPATRICK The analysis in the PwC report The Economic Impact of Artificial Intelligence on Ireland's Economy shows that the potential for AI to impact the Irish economy is slightly lower than the global average (13.8pc by 2030 and $15.7trn) but slightly higher than in other northern European (9.9pc by 2030 and $1.8trn) and southern European (11.5pc by 2030 and $0.7trn) economies.

Matthias Gerstgrasser University of Oxford Markus Waser Technical University of Denmark Peter Dal-Bianco Medical University of Vienna Heinrich Garn Austrian Institute of Technology Georg Dorffner Medical University of Vienna The diagnosis of Alzheimer's disease (AD) in routine clinical practice is most commonly based on subjective clinical interpretations. Quantitative electroencephalography (QEEG) measures have been shown to reflect neurodegenerative processes in AD and might qualify as affordable and thereby widely available markers to facilitate the objectivization of AD assessment. Here, we present a novel framework combining Riemannian tangent space mapping and elastic net regression for the development of brain atrophy markers. While most AD QEEG studies are based on small sample sizes and psychological test scores as outcome measures, here we train and test our models using data of one of the largest prospective EEG AD trials ever conducted, including MRI biomarkers of brain atrophy.

We tackle the problem of constructive preference elicitation, that is the problem of learning user preferences over very large decision problems, involving a combinatorial space of possible outcomes. In this setting, the suggested configuration is synthesized on-the-fly by solving a constrained optimization problem, while the preferences are learned itera tively by interacting with the user. Previous work has shown that Coactive Learning is a suitable method for learning user preferences in constructive scenarios. In Coactive Learning the user provides feedback to the algorithm in the form of an improvement to a suggested configuration. When the problem involves many decision variables and constraints, this type of interaction poses a significant cognitive burden on the user. We propose a decomposition technique for large preference-based decision problems relying exclusively on inference and feedback over partial configurations. This has the clear advantage of drastically reducing the user cognitive load. Additionally, part-wise inference can be (up to exponentially) less computationally demanding than inference over full configurations. We discuss the theoretical implications of working with parts and present promising empirical results on one synthetic and two realistic constructive problems.

Many cognitive, sensory and motor processes have correlates in oscillatory neural sources, which are embedded as a subspace into the recorded brain signals. Decoding such processes from noisy magnetoencephalogram/electroencephalogram (M/EEG) signals usually requires the use of data-driven analysis methods. The objective evaluation of such decoding algorithms on experimental raw signals, however, is a challenge: the amount of available M/EEG data typically is limited, labels can be unreliable, and raw signals often are contaminated with artifacts. The latter is specifically problematic, if the artifacts stem from behavioral confounds of the oscillatory neural processes of interest. To overcome some of these problems, simulation frameworks have been introduced for benchmarking decoding methods. Generating artificial brain signals, however, most simulation frameworks make strong and partially unrealistic assumptions about brain activity, which limits the generalization of obtained results to real-world conditions. In the present contribution, we thrive to remove many shortcomings of current simulation frameworks and propose a versatile alternative, that allows for objective evaluation and benchmarking of novel data-driven decoding methods for neural signals. Its central idea is to utilize post-hoc labelings of arbitrary M/EEG recordings. This strategy makes it paradigm-agnostic and allows to generate comparatively large datasets with noiseless labels. Source code and data of the novel simulation approach are made available for facilitating its adoption.

Networked data, in which every training example involves two objects and may share some common objects with others, is used in many machine learning tasks such as learning to rank and link prediction. A challenge of learning from networked examples is that target values are not known for some pairs of objects. In this case, neither the classical i.i.d.\ assumption nor techniques based on complete U-statistics can be used. Most existing theoretical results of this problem only deal with the classical empirical risk minimization (ERM) principle that always weights every example equally, but this strategy leads to unsatisfactory bounds. We consider general weighted ERM and show new universal risk bounds for this problem. These new bounds naturally define an optimization problem which leads to appropriate weights for networked examples. Though this optimization problem is not convex in general, we devise a new fully polynomial-time approximation scheme (FPTAS) to solve it.

We establish a new connection between value and policy based reinforcement learning (RL) based on a relationship between softmax temporal value consistency and policy optimality under entropy regularization. Specifically, we show that softmax consistent action values correspond to optimal entropy regularized policy probabilities along any action sequence, regardless of provenance. From this observation, we develop a new RL algorithm, Path Consistency Learning (PCL), that minimizes a notion of soft consistency error along multi-step action sequences extracted from both on- and off-policy traces. We examine the behavior of PCL in different scenarios and show that PCL can be interpreted as generalizing both actor-critic and Q-learning algorithms. We subsequently deepen the relationship by showing how a single model can be used to represent both a policy and the corresponding softmax state values, eliminating the need for a separate critic. The experimental evaluation demonstrates that PCL significantly outperforms strong actor-critic and Q-learning baselines across several benchmarks.

Whether it's telling you whether or not you need to take a raincoat on a walk or controlling the lights in your apartment with a simple verbal command, there are plenty of ways that smart speakers such as Amazon Echo or Google Home make our lives a little bit easier. But can they fundamentally improve people's lives? That's the question posed by a trial currently taking place in the U.K., in which a small number of people with learning disabilities are given devices like the Echo to see whether they can help make their lives easier -- and save the care sector some money in the process. The trial gave these devices to five people in Wales for a six-month period. The study will examine whether, during that time, they reduce the need to staff people's homes 24/7, by carrying out caregiver jobs including offering reminders about taking medication, attending appointments, and carrying out some household tasks.