Microsoft Pix Camera uses artificial intelligence to make your pictures of people better. It uses algorithms behind the scenes to analyze the 10 frames it snaps for every picture you take, looking for sharpness, exposure and even facial expressions to make sure you get the very best shot. It even takes good data from the pictures it doesn't use to enhance the photos it chooses. The app, launched last summer and just updated, now offers new filters that can help you make your photos look like real works of art. These artsy filters may sound a lot like what standalone app, Prisma, does, but Microsoft's implementation was developed by Microsoft's Asia research lab in collaboration with Skype.

Batch Normalization is a commonly used trick to improve the training of deep neural networks. These neural networks use L2 regularization, also called weight decay, ostensibly to prevent overfitting. However, we show that L2 regularization has no regularizing effect when combined with normalization. Instead, regularization has an influence on the scale of weights, and thereby on the effective learning rate. We investigate this dependence, both in theory, and experimentally. We show that popular optimization methods such as ADAM only partially eliminate the influence of normalization on the learning rate. This leads to a discussion on other ways to mitigate this issue.

Pansharpening, which is the fusion of a multispectral (MS) image and a wideband panchromatic (PAN) image, is an important technique in remote sensing. Ideally, the fused image should contain all the spectral information from the MS image and all the spatial details from the PAN image. With advances in sensor development, the fusion of a high spatial resolution MS image and a low spatial resolution hyperspectral (HS) image (MS/HS fusion) is becoming relevant for many applications. A typical HS image contains hundreds of spectral reflectance bands, making the spectral information content very high. This allows for the identification of different materials based on their spectral signature, which is useful for applications such as classification of land cover types.

I thoroughly enjoyed Pillars of Eternity, but there's no doubt Obsidian played it safe. It was every part the Baldur's Gate successor Obsidian promised in its Kickstarter pitch, not the most generic of fantasy worlds but certainly a recognizable parallel to the Sword Coast. Pillars of Eternity II: Deadfire sees Obsidian moving away from that comfort zone into something bigger and bolder. We got a hands-off look at the game during E3 this week, and while Obsidian was fairly tight-lipped about story elements, in scope it looks like as big a shift as Baldur's Gate to Baldur's Gate II. You can see the same presentation I saw by watching the video below--though you'll lack an Obsidian developer talking you through it and adding context.

Here's some news that's not fake: Not everything you can read on the internet is true. Trouble is, it can be hard to know truths from untruths, and there's evidence untruths travel faster. Many hands have been wrung in recent months over what to do about made-up news stories created to convert social media shares into page views, ad dollars, and perhaps even political traction. The modest first results from an effort to crowdsource machine learning technology to help stem the flood of falsity are a reminder that machines may help us grapple with fake news--but only if humans take the lead. Late last year, Facebook's director of AI research Yann LeCun told journalists that machine learning technology that could squash fake news "either exists or can be developed."

After my first post on Anomaly Detection for Time Series post, I would like to continue presenting what I did during the course at for the Data Science for IoT Course at Department of Continued Education of the University of Oxford with Ajit Jaokar. In line with what I wrote previously, this second post will be about predictive maintenance. The post will conclude the initial exploration of the topics I covered at Oxford. When researching materials to cover this course, I had a general idea of what to look for. Having worked already in industrial environments, I had a good idea of what predictive maintenance should be and how it could be used.

An Australian startup revealed its flagship product, an earpiece that can interpret 8 different languages in real-time, at a United Nations event in Switzerland on Friday. Lingmo International, a startup based in West Gosford north of Sydney, launched its TranslateOne2One earpiece at the UN's Artificial Intelligence for Good Summit in Geneva, revealing that IBM Watson machine learning technology had been used for its algorithms. Traditionally, converting one language to another orally in real-time is called "interpreting" whereas the term "translation" is reserved for processing text across languages with some delay. Lingmo founder Danny May, however, describes his product as performing "translation in real-time". And what I mean by independent is that it doesn't require any connectivity to your phone by Bluetooth or wi-fi.

The US military will have more robot soldiers on the battlefield than real ones by 2025, a top British intelligence expert has claimed. John Bassett, a security consultant with a 20-year career at GCHQ, believes deadly combat robots are rapidly becoming "a reality" of modern day warfare. These state-of-the art military units will consist of human soldiers and robots and are aimed at maximizing performance on future battlefields. Combat robots will rapidly become an inherent part of US fighting forces within the next 10-15 years, defense experts say. Washington is apparently seeking to gain military edge over China, Russia and other rivals investing in research and development of robotic systems.

All data processing includes extreme measures regarding information security, and personal data, such as names, identity numbers and addresses are concealed already during data collection. Encryption is used for all data transmission. Public data has never previously been combined in the same way as in the experiment by Espoo and Tieto. The experiment is only possible thanks to the extremely precise public-sector information systems and high-quality data that exist in Finland. "The aim of the experiment is to combine information assets that had previously been separate in order to form customer and service pathways. We hope to find out whether artificial intelligence will enable us to identify residents who will need services at an earlier stage than we can now, while protecting the identities of individuals. In the future, it may be possible to allocate the city's resources more efficiently while improving the quality of life of individual residents, as they are offered preventive support," says Tomas Lehtinen, Project Manager, analyst at the Mayor's Office of the City of Espoo.

In our previous article on chatbots, it was stated that one does not simply create a chatbot. Artificial intelligence still poses too many challenges to go about building bots without sufficient research. One of the main problems is the ability of bots to recognize human requests. We've decided to tell you about some of the popular AI chatbots that handle this task well. In theory, the best chat bots were made to replace human services.