ニューヨーク・タイムズのインスタグラム(nytimes) - 7月31日 04時01分

In the past, robotic hands could only do what vast teams of engineers programmed them to do. Now they can learn more complex tasks on their own. @maeryan and Cade Metz went inside the world’s top #ArtificialIntelligence labs to see how they work. Here, at the @uofwa, researchers are training robotic hands that have all the same digits and joints that our hands do. It’s not easy — an #anthropomorphic hand moves in so many different ways. So, they train their hand in simulation, a digital recreation of the real world. | Swipe to see a robotic hand called Dactyl, at OpenAI. The system can learn to spin an alphabet block through what would have been 100 years of trial and error. Dactyl learned this task largely on its own — and that’s a big deal. | Swipe to see a #robot that’ll make your bed. Researchers at @ucberkeleyofficial pulled the system together in just 2 weeks, using the latest machine learning techniques. Not long ago, this would have taken months or years. | And another system at Berkeley can push an object with a gripper and predict where it’ll go. The system learns this behavior by analyzing videos showing how objects get pushed. In this way, it can deal with the uncertainties and unexpected movements. | These are all simple tasks. But with machine learning, robot hands are evolving to do what ours can. ? ?

[BIHAKUEN]UVシールド(UVShield)

>> 飲む日焼け止め！「UVシールド」を購入する