Sequence modeling with neural networks has lead to powerful models of symbolic music data. We address the problem of exploiting these models to reach creative musical goals, by combining with human input. To this end we generalise previous work, which sampled Markovian sequence models under the constraint that the sequence belong to the language of a given finite state machine provided by the human. We consider more expressive non-Markov models, thereby requiring approximate sampling which we provide in the form of an efficient sequential Monte Carlo method. In addition we provide and compare with a beam search strategy for conditional probability maximisation. Our algorithms are capable of convincingly re-harmonising famous musical works. To demonstrate this we provide visualisations, quantitative experiments, a human listening test and audio examples. We find both the sampling and optimisation procedures to be effective, yet complementary in character. For the case of highly permissive constraint sets, we find that sampling is to be preferred due to the overly regular nature of the optimisation based results. The generality of our algorithms permits countless other creative applications.

Google is launching new features within its free Cloud Natural Languages API (shared software technology) that will help newsrooms and other businesses sort information so that it's easier to find later. Google is seen as a mixed bag for publishers, offering a lot of traffic, but also sucking up a lot of the industry's ad revenue. Still, the features could be game-changers for newsrooms faced with the daunting task of classifying and taxonomizing hundreds of articles per day and thousands of articles in their archives. It will also make translating text from different languages much easier so that publishers can enter new markets more easily. Publications like Hearst and Vice have already been testing the new features that will be open to all newsrooms and businesses moving forward.

What is machine learning and how does it work? You've probably heard the buzz. The age of artificial intelligence has arrived. But that doesn't mean it's easy to wrap your mind around. Let's break down the basics of artificial intelligence, bots, and machine learning. Besides, there's nothing that will impact marketing more in the next five to ten years than artificial intelligence.

In Switzerland, executives from the three partner companies showcased the drone in action on Thursday, with a woman enacting the scene of loading up the drone with a bag of coffee that was flown several kilometers and landed smoothly on the rooftop of a Mercedes-Benz van. After high-fives at the successful flight, the coffee was then brewed up at a coffee cart at the ready and served for the several dozen attendees.

Amazon just dropped a whole bunch of Alexa-equipped gear on us, including a redesigned Echo speaker with a price tag of just $99. That's a bargain compared to the $189 price of its predecessor, but you can take advantage of Alexa's best new feature on your old Echo stuff right now, for free: Multi-Room Music is rolling out across the platform. Multi-Room Music allows you to play your tunes or talk radio (woo!) simultaneously across a variety of Echo devices on the same network. It works like a Sonos system, pumping audio to individual speakers throughout a house. Or, you can place several speakers around a really big room that's covered by the same network, a technique often used by gyms, studios, or other large spaces.

The last three weeks I was working on ImageMonkey - check it out here: https://imagemonkey.io/. The idea originated while I was working on another project where at some point I wanted to integrate Machine Learning into my application. With all the great Machine Learning frameworks out there, it's really easy to get your foot into the door quickly. But while I was playing a little bit with the frameworks, I somehow realized that it's really hard to get some good training data. If you are lucky then there is some (annotated) training data online, if not...well, then you have to get your hands dirty and do the tedious work yourself.

For more than a year, "Blade Runner 2049" director Denis Villeneuve and star Ryan Gosling have been working under the cover of CIA-level stealthiness. On the film's Budapest set, copies of the script for the long-awaited sequel to Ridley Scott's 1982 sci-fi neo-noir classic were held so closely -- literally locked away in safes when not in use -- that many crew members never laid eyes on one. Actors would sign out their sides on the day they were shooting a scene and be required to sign them back in before going home, lest the merest hint of a spoiler leak. On a late-September afternoon, with the film's Oct. 6 release just days away, Villeneuve and Gosling sat in a windowless conference room in a hotel in downtown Los Angeles, preparing to finally let audiences in on their secret -- and wondering what they will make of it. "When you make a movie as a filmmaker, it's like you bring people in a boat and you say, 'We will discover America together,' " said Villeneuve, the French Canadian director of such films as 2015's crime thriller "Sicario" and the 2016 cerebral sci-fi hit "Arrival."

A few days ago, I finally bought a pair of AirPods. Apple's funny-looking ear-computers have been available for about a year, provided you were willing to order them online and wait six weeks for delivery. When I saw that I could snag a pair at my local Apple store, I decided to take the plunge. I worry that makes me sound like an impulsive person. At $159, AirPods are expensive.

Artificial Intelligence has been described as "Thor's Hammer" and "the new electricity." But it's also a bit of a mystery – even to those who know it best. We'll connect with some of the world's leading experts in artificial intelligence, deep learning, and machine learning to explain how it works, how it's evolving, and how it intersects with every facet of human endeavor, from art to science. We release new episodes every week. A long-time journalist based in Silicon Valley, Michael has been in the thick of technological change since the web took hold.

Thursday morning starts with more Amazon products than you could ever possibly want. Remember how the retailer found success with its Echo speaker? Everything gets a voice assistant. Amazon might not have YouTube on Echo anymore, but the devices are coming in several new shapes and sizes. Our three-minute wrap-up video from the big event shows everything you need to know, or you can dive in to check out our impressions of the new Echo speakers (now in regular or tall size), its small-screened Echo Spot or the latest Fire TV device, which puts 4K video streaming in a dongle.