Hasta La Gadget!

Anki Drive, the artificial intelligence-assistanted toy car game that debuted last fall, is back with new cars, new tracks, and new ways to play.

As you may have noticed, there are a million activity trackers out there right now. Most of them are glorified pedometers. The upper tier add altimeters, heart rate monitors, and sleep-tracking to the equation. But what if you don’t just want to be reminded to work out—you want to work out better, safer, and more efficiently? Moov might just be the AI coach you’ve been hoping for.

Read more…

Artificial intelligence has certainly improved by leaps and bounds over the years, where machine learning has been implemented to help out humanity including the possibility of discovering a potential cure for HIV. It looks like there are other avenues to take advantage of machine learning such as the realm of arts. MIT PhD student J. Nathan Matias has decided to take the popular machine-learning Android app known as Swiftkey to have it play with several of the Bard’s (that’s Shakespeare for you) more famous words, resulting in a bunch of sonnets that might just melt the heart of the party you’re trying to impress.

(more…)

Swift-speare Creates Sonnets That Actually Make Sense original content from Ubergizmo.

Google seems to have paid at least $500 million to acquire DeepMind, an artificial intelligence startup that has a number of high-profile investors, and that has demoed tech which shows computers playing video games in ways very similar to human players. Facebook reportedly also tried to buy the company, and the question on most people’s minds is “Why?”

More intelligent computing means more insightful data gathering and analysis, of course. Any old computer can collect information, and even do some basic analytics work in terms of comparing and contrasting it to other sets of data, drawing simple conclusions where causal or correlational factors are plainly obvious. But it still takes human analysts to make meaning from all that data, and to select the significant information from the huge, indiscriminate firehose of consumer data that comes in every day.

AI and machine learning expertise can help improve the efficiency and quality of data gathered by Google and other companies who rely on said information, but it can also set the company up for the next major stage in computing interaction: turning the Internet of Things into the Internet of Companions. Google is hard at work on tech that will make even more of our lives computer-centric, including driverless cars and humanoid robots to take over routine tasks like parcel delivery, but all those new opportunities for computer interaction need a better interface if they’re to become trusted and widely used.

https://twitter.com/chadcat/status/427671366163124224

Google has already been working on building software and tech than anticipates the needs of a user and acts as a kind of personal valet. Google Now parses information from your Gmail and search history to predict what you’ll ask about and provide the information in advance. Now has steadily been growing smart and incorporating more data sources, but it still has plenty of room for improvement, and there’s no better way to anticipate a human’s needs than with a computer that thinks like one.

Another key component of Google’s future strategy has to do with hardware. The company’s last high-profile acquisition was Nest Labs, which it bought for $3.2 billion in cash earlier this month. Nest’s smart thermostat also uses a significant amount of machine learning to help anticipate the schedule and needs of its users, which is something that DeepMind could assist with on a basic level. But there’s a larger opportunity, as once again a more human element could help make the Internet of Things a more accessible concept for the average user.

We’ve seen little beyond computers that can play video games from DeepMind, but that demonstration speaks volumes about what Google can do with the company. Robotics and hardware investments like those already made by the company are interesting, to be sure, but DeepMind is in many ways the thread that will draw all these separate initiatives together: There’s an adoption disconnect between technically impressive innovations, and convincing everyday end users to actually embrace them. DeepMind could help humanize tech that seems otherwise deeply impersonal (and in the case of self-driving cars, even anti-human) in a way that spurs uptake.

More human machines could be a big reason why Google has reportedly created an ethics board to supervise the use of DeepMind’s AI tech. Google probably isn’t that worried about the possibility of accidentally creating SkyNet, but when you start building computing devices that think and act like humans, you’re bound to get into fraught moral territory. Both in terms of both what said tech can learn and know about its users, as well as what, if any, responsibility we have to treat said tech differently than any standard computer.

Depending on your view of Google and what it does, the DeepMind acquisition is either troubling or exciting. Of course, it has the potential to be both, as does any potential advancement in AI and machine learning, but I can’t help but be enthralled by the possibilities of the picture Google is painting with its latest big-picture moves. More than any other, it seems to be committed to a future that lives up to the vision of the science fiction blockbusters we all grew up with, and it’s impossible to deny the allure of that kind of ambition.

Just when you thought that humans have perfected the art of being in denial, along comes this particular robot telemarketer who actually denies being a robot. TIME Washington Bureau Chief Michael Scherer received a phone call recently, with an extremely engaging voice who intended to persuade Scherer to bite on a health insurance deal, but Scherer felt that something was not quite right. When asked point blank if “she” was an actual human being or a computer-operated robot voice, “she” claimed that she was real enough, followed by some laughter. Other tests proved to be “her” undoing, however, especially when she could not understand the question, “What vegetable is found in tomato soup?”

Apart from that, when “she” was asked multiple times as to what day of the week it was, all “she” could muster up was the excuse of a bad connection. TIME reporters decided to spend the better part of the next hour to call “her” back, and the robot claimed “she” was Samantha West, with her main goal in life not to terminate John Connor, but to ask a bunch of questions concerning health coverage with hopes of closing a sale after transferring an all too eager customer to an actual person. Creepy, don’t you think so? Definitely not as helpful as the Yana and Bo robots.

Robot Telemarketer Denies Being A Robot original content from Ubergizmo.