Like all everyday miracles of technology, the longer you watch a robot perform surgery on a human being, the more it begins to look like an inevitable natural wonder. Earlier this month I was in an operating theatre at University College Hospital in central London watching a 59-year-old man from Potters Bar having his cancerous prostate gland removed by the four dexterous metal arms of an American-made machine, in what is likely a glimpse of the future of most surgical procedures. The robot was being controlled by Greg Shaw, a consultant urologist and surgeon sitting in the far corner of the room with his head under the black hood of a 3D monitor, like a Victorian wedding photographer. Shaw was directing the arms of the remote surgical tool with a fluid mixture of joystick control and foot-pedal pressure and amplified instruction to his theatre team standing at the patient's side. The surgeon, 43, has performed a thousand of such procedures, which are particularly useful for pelvic operations; those, he says, in which you are otherwise "looking down a deep, dark hole with a flashlight".

Autonomous vehicles are big business. So it's no surprise that some of the highest-valued technology companies in the world are looking to get in on the action. Giants like Uber, Tesla, Waymo, and Apple have all invested heavily in what they see as the transport of tomorrow: AI-driven cars capable of such precise steering and navigation that they'll never need a human at the wheel. If successful, such vehicles would free up time for human drivers, allowing them to sleep on long-haul journeys, finish typing up presentations on the way to work, or simply use their time as they see fit (so, Netflix). Statistically, these vehicles should also be much, much safer.

A top Defense Department tech official on Tuesday said the Pentagon is weeks away from publishing its first broad strategy for artificial intelligence. The department plans to release a report detailing its long-term plans for artificial intelligence "within weeks" as leaders increasingly stress the technology's potential to strengthen national security, said Thomas Michelli, the department's acting deputy chief information officer for cybersecurity. Though he didn't disclose specifics regarding the strategy, Michelli said the Pentagon is standing up a number of artificial intelligence capabilities across the enterprise and people should expect "a big announcement" in the weeks to come. The department's updated cyber strategy, which is scheduled to publish shortly before the artificial intelligence plan, will also discuss AI and offer more details on how the military is funneling additional resources into the technology, according to Michelli. "We have to move forward or else we'll lose [our] competitive advantage," he said at a forum hosted by GovernmentCIO Media.

Self-driving cars could be on Irish roads within the next five years – but we need to be ready for the new technology, the AA has said. The association's director of consumer affairs Conor Faughnan warned the country could miss out if we don't start preparing for the new technology. Autonomous and connected vehicles have the potential to make our roads safer, but authorities here have not yet begun to plan for them. Mr Faughnan told the Irish Sunday Mirror: "It's going to be a strange new world for people. "It will take some getting used to but long-term the prospects are fascinating.

It was not the football World Cup, but the defeat was not unlike Christiano Ronaldo and Lionel Messi losing on the same day. An artificial intelligence system recorded a 2-0 victory against elite physicians on 30 June 2018 in two rounds of a competition in Beijing to diagnose brain tumours and predict the expansion of brain hematomas, or bruises. BioMind, developed by researchers from the AI Research Centre for Neurological Disorders and Capital Medical University, made correct diagnoses in 87 per cent of 225 cases in about 15 minutes. A team of 15 doctors from top hospitals across China achieved 66 per cent accuracy in 30 minutes. The AI system also made correct predictions in 83 per cent of brain hematoma expansion cases, outperforming the physicians, who had only 63 per cent accuracy.

When you think of game development, you might imagine hundreds of people working at huge studios like Sony or EA. But independent developers, or indies, make big games too. These are small teams, often with modest budgets, who make games without creative input from investors. Popular farm simulator Stardew Valley and chaotic co-operative game Overcooked were built on desks in bedrooms - yet both have been recognised by BAFTA . We headed into the woods in central Sweden to visit Stugan, a programme that brings indies from the UK and all over the world to the Nordic countryside to work on projects side-by-side. There we asked three teams one simple question: Can anyone make games?

Syntactic search relies on keywords contained in a query to find suitable documents. So, documents that do not contain the keywords but contain information related to the query are not retrieved. Spreading activation is an algorithm for finding latent information in a query by exploiting relations between nodes in an associative network or semantic network. However, the classical spreading activation algorithm uses all relations of a node in the network that will add unsuitable information into the query. In this paper, we propose a novel approach for semantic text search, called query-oriented-constrained spreading activation that only uses relations relating to the content of the query to find really related information. Experiments on a benchmark dataset show that, in terms of the MAP measure, our search engine is 18.9% and 43.8% respectively better than the syntactic search and the search using the classical constrained spreading activation. NTRODUCTION With rapid development of the Word Wide Web and e-societies, information retrieval (IR) has many challenges in exploiting those rich and huge information resources. Whereas, the keyword based IR has many limitations in finding suitable documents for user's queries. Semantic search improves search precision and recall by understanding user's intent and the contextual meaning of terms in documents and queries.

To backpropagate the gradients through discrete stochastic layers, we encode the true gradients into a multiplication between random noises and the difference of the same function of two different sets of discrete latent variables, which are correlated with these random noises. The expectations of that multiplication over iterations are zeros combined with spikes from time to time. To modulate the frequencies, amplitudes, and signs of the spikes to capture the temporal evolution of the true gradients, we propose the augment-REINFORCE-merge (ARM) estimator that combines data augmentation, the score-function estimator, permutation of the indices of latent variables, and variance reduction for Monte Carlo integration using common random numbers. The ARM estimator provides low-variance and unbiased gradient estimates for the parameters of discrete distributions, leading to state-of-the-art performance in both auto-encoding variational Bayes and maximum likelihood inference, for discrete latent variable models with one or multiple discrete stochastic layers.

A preference order or ranking aggregated from pairwise comparison data is commonly understood as a strict total order. However, in real-world scenarios, some items are intrinsically ambiguous in comparisons, which may very well be an inherent uncertainty of the data. In this case, the conventional total order ranking can not capture such uncertainty with mere global ranking or utility scores. In this paper, we are specifically interested in the recent surge in crowdsourcing applications to predict partial but more accurate (i.e., making less incorrect statements) orders rather than complete ones. To do so, we propose a novel framework to learn some probabilistic models of partial orders as a \emph{margin-based Maximum Likelihood Estimate} (MLE) method. We prove that the induced MLE is a joint convex optimization problem with respect to all the parameters, including the global ranking scores and margin parameter. Moreover, three kinds of generalized linear models are studied, including the basic uniform model, Bradley-Terry model, and Thurstone-Mosteller model, equipped with some theoretical analysis on FDR and Power control for the proposed methods. The validity of these models are supported by experiments with both simulated and real-world datasets, which shows that the proposed models exhibit improvements compared with traditional state-of-the-art algorithms.

Algorithmic probability has shown some promise in dealing with the probability problem in the Everett interpretation, since it provides an objective, single-case probability measure. Many find the Everettian cosmology to be overly extravagant, however, and algorithmic probability has also provided improved models of subjective probability and Bayesian reasoning. I attempt here to generalize algorithmic Everettianism to more Bayesian and subjectivist interpretations. I present a general framework for applying generative probability, of which algorithmic probability can be considered a special case. I apply this framework to two commonly vexing thought experiments that have immediate application to quantum probability: the Sleeping Beauty and Replicator experiments.