Recently, impressive denoising results have been achieved by Bayesian approaches which assume Gaussian models for the image patches. This improvement in performance can be attributed to the use of per-patch models. Unfortunately such an approach is particularly unstable for most inverse problems beyond denoising. In this work, we propose the use of a hyperprior to model image patches, in order to stabilize the estimation procedure. There are two main advantages to the proposed restoration scheme: Firstly it is adapted to diagonal degradation matrices, and in particular to missing data problems (e.g. inpainting of missing pixels or zooming). Secondly it can deal with signal dependent noise models, particularly suited to digital cameras. As such, the scheme is especially adapted to computational photography. In order to illustrate this point, we provide an application to high dynamic range imaging from a single image taken with a modified sensor, which shows the effectiveness of the proposed scheme.

The world's most pampered rhino is having his mudpack applied by hand. Sudan the veteran northern white rhinoceros stands placidly and allows his personal assistant to rub dollops of wet clay into his hide, to moisturise his skin and keep insects away. 'He loves that,' says Zach, his 24-hour-a-day PA. Zach lives next door to the rhino enclosure, permanently on call in case His Lordship should require anything. And Sudan is very high maintenance – he used to have a man whose job was rubbing oil into his hooves.

An exciting new proposition from two students at the University of Toronto may hold the key to solving logistical problems in monitoring and addressing the state of Canadian rivers. Alberto Camacho and Patricio Córdova have developed a scheme named Drone River which utilises a three-pronged approach in order to provide better surveillance and analysis of river levels than ever before. Their system relies upon three different components: firstly, an automated web service handles texts, emails, phone calls and social media updates to gather information about potential concerns in different areas of the river. Secondly, a network of sensors are distributed onsite along the river to collect real-time in-situ data, augmenting their knowledge of the potential problems. After the problem zones have been identified, drones are then deployed to perform basic reconnaissance by taking pictures, videos and water samples, before returning to base for the information to be processed.

Yasmin Green leads a team at Google's parent company with an audacious goal: solving the thorniest geopolitical problems that emerge online. Jigsaw, where she is the head of research and development, is a think tank within Alphabet tasked with fighting the unintended unsavory consequences of technological progress. That means listening to fake news creators, jihadis, and cyber bullies so that she and her team can understand their motivations, processes, and goals. "We look at censorship, cybersecurity, cyberattacks, ISIS--everything the creators of the internet did not imagine the internet would be used for," Green said today at WIRED's 2017 Business Conference in New York. Last week, Green traveled to Macedonia to meet with peddlers of fake news, those click-hungry opportunists who had such a sway over the 2016 presidential election in the US.

I come here not to praise The Mummy but to re-bury it. Tom Cruise remains a movie star no matter what kind of nonsense tries to bring him down. If he can survive that thing with the couch, he can survive anything. But the failure-to-launch of The Mummy would jeopardize more than careers and studio marketing budgets. The movie is supposed to start up the Dark Universe, Universal Studios' multi-movie franchise starring its monsters of yore.

Let's start with a little challenge: which of the following tunes was composed by an AI, and which by an HI (Human Intelligence)? I'll tell you at the end of the answer which tune was composed by an AI and which by an HI. Both pieces of music are pleasing to the ear. Both could be part of the soundtrack of a Hollywood film, and you would never know that one was composed by an AI. And this is just the beginning.

The Indie Megabooth, a bastion of independent games at conventions around the world, is growing. But this time around, things are different: The Megabooth is expanding beyond games. The Megashow is the first standalone festival to be spun off from the Megabooth proper: It's a roaming, daylong fair designed to highlight local creatives and bring communities together around video games, art and music. The Megashow will make its debut at the Tabernacle in Atlanta on July 15th. "Our initial goal was to expose new audiences to indie games and show them that there was this whole other side to game creation," says Megabooth leader Kelly Wallick.

SoftBank said it would buy two firms that build walking robots from Google's parent company, Alphabet, adding to the Japanese company's growing artificial intelligence portfolio. SoftBank said it would buy Boston Dynamics and Tokyo-based Schaft, which design and manufacture robots that simulate human movement, but did not disclose the terms of the transactions. Shares of the company rose as much as 7.9 per cent after the deal was announced, hitting a 17-year high. "Smart robotics are going to be a key driver of the next stage of the information revolution, and Marc (Raibert) and his team at Boston Dynamics are the clear technology leaders in advanced dynamic robots," SoftBank chairman Masayoshi Son said in a statement on Friday. Mr Raibert is chief executive and founder of Boston Dynamics. SoftBank has embarked on an aggressive acquisition campaign to boost its research and development capabilities.

Abstract: Automating statistical modelling is a challenging problem that has far-reaching implications for artificial intelligence. The Automatic Statistician employs a kernel search algorithm to provide a first step in this direction for regression problems. However this does not scale due to its O(N3) running time for the model selection. This is undesirable not only because the average size of data sets is growing fast, but also because there is potentially more information in bigger data, implying a greater need for more expressive models that can discover finer structure. We propose Scalable Kernel Composition (SKC), a scalable kernel search algorithm, to encompass big data within the boundaries of automated statistical modelling.

Humans are not either thinking machines or feeling machines but rather feeling machines that think.