DeepMind AI Learns Soccer

Artificial intelligence has learned to play soccer. By learning from decades’ worth of computer simulations, an AI took digital humanoids from flailing tots to proficient players. Researchers at the AI research company DeepMind taught the AI how to play soccer in a computer simulation through an athletic curriculum resembling a sped-up version of a human baby growing into a soccer player. The AI was given control over digital humanoids with realistic body masses and joint movements. The first phase of the curriculum trained the digital humanoids to run naturally by imitating motion-capture video clips of humans playing soccer. A second phase involved practicing dribbling and shooting the ball through a form of trial-and-error machine learning that rewarded the AI for staying close to the ball.

The first two phases represented about 1.5 years of simulation training time, which the AI sped through in about 24 hours. But more complex behaviours beyond movement and ball control began emerging after five simulated years of soccer matches. The third phase of training challenged the digital humanoids to score goals in two-on-two matches. Teamwork skills, such as anticipating where to receive a pass, emerged over the course of about 20 to 30 simulated years of matches, or the equivalent of two to three weeks in the real world. This led to demonstrated improvements in the digital humanoids’ off-ball scoring opportunity ratings, a real-world measure of how often a player gets in a favourable position on the pitch. Such simulations won’t immediately lead to flashy soccer-playing robots.

More information:

https://www.newscientist.com/article/2336132-deepmind-ai-learns-to-play-soccer-using-decades-of-match-simulations/

Mavrik-Pro VR Gun

StrikerVR makes some of the best haptic VR guns in the biz which can be found at premiere VR attractions. Though the peripheral packs and impressive punch, the pricey product is a tough sell for smaller venues and consumers. StrikerVR hopes to fix that with its latest haptic VR gun, the Mavrik-Pro which is now available for pre-order by VR attractions and expected to ship in October.

Though it’s the company’s most affordable peripheral yet, it’s still rocking a $1,000 price tag. But the company says it plans to sell the gun to consumers next year at an (unknown) reduced price point, and with a wider variety of colors. That would of course need to be backed by support for at least some consumer VR games, though there’s no word on that at this point.

More information:

https://www.roadtovr.com/strikervr-mavrik-pro-pre-order-launch-consumer-version/

3D Images from Single Exposure

Researchers have developed a camera that uses a thin microlens array and new image processing algorithms to capture 3D information about objects in a scene with a single exposure. The camera could be useful for a variety of applications such as industrial part inspection, gesture recognition and collecting data for 3D display systems. Researchers consider their camera lensless because it replaces the bulk lenses used in conventional cameras with a thin, lightweight microlens array made of flexible polymer.

Because each microlens can observe objects from different viewing angles, it can accomplish complex imaging tasks such as acquiring 3D information from objects partially obscured by objects closer to the camera. The camera learns from existing data how to digitally reconstruct a 3D scene, therefore it can produce 3D images in real time. This 3D camera could be used to give robots 3D vision as well as provide content for 3D displays used in gaming, entertainment or many other applications.

More information:

https://www.optica.org/en-us/about/newsroom/news_releases/2022/september/lensless_camera_creates_3d_images_from_single_expo/