He bobbed his head with the groove, and leaned way in when he wanted to play more complicated melodies, rocking and rolling with the beat of the jam. This wasn't a your average jazz band member, though--this was Shimon, a four-armed robot marimba player built by the Georgia Institute of Technology to be able to listen to music, improvise, and play along with human musicians. At a performance at Moogfest, a four-day music and technology festival in Durham, North Carolina, Gil Weinberg, the lead researcher at Georgia Tech's Center for Music Technology, demonstrated what he and his lab have been working on for the past 12 years. Their efforts have aimed at augmenting the creative capabilities of humans with robotics. That can mean robots like Shimon, which uses machine-learning programs trained on music theory and a wide range of musical styles, from chamber music to dubstep, to be able to add a superhuman element to musical performances, playing chord structures that would be physically impossible for humans to hit. But it can also mean robotic enhancements for humans: At the concert, Weinberg introduced Jason Barnes, a drummer who lost the lower part of his right arm a few years ago.

Google revealed a handful of slick new products this week that showed off its impressive artificial intelligence and machine learning tech chops. Its new conversational assistant, for example, will take its traditional search product to the next level, letting users ask it questions, find suggestions, or book services through text chat in its new messaging app Allo or voice, in its new smart speaker, Home. Google's building on many years of research, development, and data collecting here and it shows. Although Home is essentially a follow-up to Amazon's Echo speaker, Google has much more experience with voice search and natural language processing. Ask Home a question, like "How tall is Steph Curry?" and you can follow-up with "What's his jersey number?" and Home will know who you're still talking about, where Alexa can only handle single queries.

The giant hedge fund, which manages 35 billion, is as much a technology company as it is a hedge fund. It uses advanced technologies to find investment opportunities, and it just hosted its annual artificial intelligence competition. One of those technological applications involves using natural-language-processing techniques to analyze the Fed minutes, such as those set for release Wednesday afternoon. "Historically, interpretations of those minutes required art, so Fed watchers pontificated and critiqued," the firm said in a note. "Now natural language processing techniques can translate those minutes into relatively objective data."

Marco Arment, the former lead developer for Tumblr and creator of Instapaper, this weekend (May 21) argued in a blog post that Apple risks succumbing to the same fate as BlackBerry, which saw its business evaporate when Apple's iPhone changed the game. Arment contends that Apple is unprepared for the shift if consumers begin favoring services rooted in artificial intelligence being developed by other internet giants. "Today, Amazon, Facebook, and Google are placing large bets on advanced AI, ubiquitous assistants, and voice interfaces, hoping that these will become the next thing that our devices are for. Today, Apple's being led properly day-to-day and doing very well overall. But if the landscape shifts to prioritize those big-data AI services, Apple will find itself in a similar position as BlackBerry did almost a decade ago: what they're able to do, despite being very good at it, won't be enough anymore, and they won't be able to catch up."

At the top of the corporate world, Apple and Google are in a back-and-forth battle to be number one. It's not clear which of the two Silicon Valley giants will emerge on top in a contest which highlights the contrast of very different business models. Apple then regained, lost and recovered the leader position in May in a battle that appears set to continue for some time. At the close Friday, Apple was worth some 522 billion, to 496 billion for Alphabet. The two companies have both been hugely profitable in recent years, for different reasons. Apple has delivered a line of must-have iPhones and other gadgets that have set trends around the world but now "appears to be a little bit immobile," says Roger Kay, analyst at Endpoint Technologies Associates.

Really cool stuff, exactly what I was looking for.

Artificial Intelligence and machine learning are hot topics at every technology conference I go to, and the recent DLD NYC conference was no exception. Ramin Assadollahi of ExB Group, a German company dealing with cognitive computing in healthcare, focused on a variety of ways new computer techniques can help us learn "how to heal with software." Addressing many of the terms that are thrown around today, he noted that AI does not have to be cognitive computing, cognitive computing does not have to be machine learning, and big data is a separate issue entirely. Assadollahi focused on ways AI could improve the field of medicine. He noted that a pathologist looking at tissue data typically sees 200,000 samples in his or her work lifetime, but with deep learning and modern graphics cards, a computer system can process that many in two weeks.

One possible explanation is that this is another manifestation of the curse of dimensionality. As the dimension of a space increases it is well known that the volume of a hypersphere becomes increasingly concentrated at its surface. Given that the decision boundaries of a deep neural network are in a very high dimensional space it seems reasonable that most correctly classified examples are going to be close to the decision boundary - hence the ability to find a misclassified example close to the correct one, you simply have to work out the direction to the closest boundary.

Deeplearning4J is an Apache 2.0-licensed, open-source, distributed neural net library written in Java and Scala. Deeplearning4J integrates with Hadoop and Spark and runs on several backends that enable use of CPUs and GPUs. The aim is to create a plug-and-play solution that is more convention than configuration, and which allows for fast prototyping. The most recent stable release in Maven Central is 0.4-rc3.9, For more on working with snapshots, see this page.

Artificial neural networks were inspired by the human brain and simulate how neurons behave when they are shown a sensory input (e.g., images, sounds, etc). They are known to be excellent tools for image recognition, any many other problems beyond that - they also excel at weather predictions, breast cancer cell mitosis detection, brain image segmentation and toxicity prediction among many others. Deep learning means that we use an artificial neural network with multiple layers, making it even more powerful for more difficult tasks. This time they have been shown to be apt at reproducing the artistic style of many famous painters, such as Vincent Van Gogh and Pablo Picasso among many others. All the user needs to do is provide an input photograph and a target image from which the artistic style will be learned.