Toggle light / dark theme

MIT’s deep-learning software produces videos of the future

When you see a photo of a dog bounding across the lawn, it’s pretty easy for us humans to imagine how the following moments played out. Well, scientists at MIT have just trained machines to do the same thing, with artificial intelligence software that can take a single image and use it to to create a short video of the seconds that followed. The technology is still bare-bones, but could one day make for smarter self-driving cars that are better prepared for the unexpected, among other applications.

The software uses a deep-learning algorithm that was trained on two million unlabeled videos amounting to a year’s worth of screen time. It actually consists of two separate neural networks that compete with one another. The first has been taught to separate the foreground and the background and to identify the object in the image, which allows the model to then determine what is moving and what isn’t.

According to the scientists, this approach improves on other computer vision technologies under development that can also create video of the future. These involve taking the information available in existing videos and stretching them out with computer-generated vision, by building each frame one at a time. The new software is claimed to be more accurate, by producing up to 32 frames per second and building out entire scenes in one go.

Read more

CERN introduces Large Hadron Collider’s robotic inspectors

Since the Large Hadron Collider (LHC) needs to be in tip-top shape to discover new particles, it has two inspectors making sure everything’s in working order. The two of them are called TIM, short not for Timothy, but for Train Inspection Monorail. These mini autonomous monorails that keep an eye on the world’s largest particle collider follow a pre-defined route and get around using tracks suspended from the ceiling. According to CERN’s post introducing the machines, the tracks are remnants from the time the tunnel housed the Large Electron Positron instead of the LHC. The LEP’s monorail was bigger, but not quite as high-tech: it was mainly used to transport materials and workers.

As for what the machines can do, the answer is “quite a few.” They can monitor the tunnel’s structure, oxygen percentage, temperature and communication bandwidth in real time. Both TIMs can also take visual and infrared images, as well as pull small wagons behind them if needed. You can watch them in action below — as you can see, they’re not much to look at with their boxy silver appearance. But without them, it’ll be tough monitoring a massive circular tunnel with a 17-mile circumference.

Read more

We are pleased to share an update on our research in 3D capture and algorithms

We took the technology out of the studio and into a car – making Holoportation truly mobile. To accomplish this, we reduced the bandwidth requirements by 97%, while still maintaining quality. This new mobile Holoportation system greatly increases the potential applications of real-time 3D capture and transmission.

Read more

Cormorant/AirMule UAV completes first autonomous flight

A lot has happened since we first heard about the AirMule, a prototype VTOL (Vertical Takeoff and Landing) aircraft that features internal rotor blades – these work along with the horizontal-thrust ducted fans visible at the rear. First, it made some tethered autonomous test flights. Then, it flew untethered for a short distance. Now, known as the Cormorant UAV, it’s made its first full untethered autonomous flight … although there were a couple of hiccups.

Designed by Israeli firm Tactical Robotics, the Cormorant is designed to deliver troops, civilian passengers or other cargo within tight quarters where helicopters with exposed rotor blades just can’t go. With the UAV in its name standing for Unmanned Aerial Vehicle, the idea is that it will perform these tasks either autonomously or by remote control.

The latest test took place in Israel on Nov. 3rd, lasting only about two minutes and involving low flight over uneven terrain. While the aircraft did successfully demonstrate autonomous flight modes such as takeoff, climb, acceleration, cruise, deceleration, descent, turns, hover and touchdown, it is hoped that subsequent flights will be able to smooth out the transitions between those modes.

Read more

Automatic Charging for E-Mobility

NRG-X is the world’s first fully automatic charging solution, which provides efficient, high power energy transfer, has a great range of parking tolerance and can be simply retrofitted to almost every electric vehicle. Therefore, NRG-X is the ideal solution for convenient every day charging of your electric vehicle primarily at home. The system comprises two basic components:

Read more

Coming to Grips with Artificial Intelligence’s Many Manifestations

The categories of AI.


Click here to learn more about author James Kobielus.

Artificial intelligence (AI) is all the rage these days. However, people often overlook the fact that it’s a truly ancient vogue. I can’t think of another current high-tech mania whose hype curve got going during the days when Ike was in the White House, “I Love Lucy” was on the small screen, and programming in assembly language was state of the art.

As AI’s adoption grows, we run the risk of belittling the technology’s potential if we continue to fixate on the notion that it’s “artificial.” When you think of it, all technologies are artificial, pretty much by definition. Cars are artificial transportation, houses are artificial shelters, and so on.

Read more