Robots Learn Problem-Solving From IKEA Furniture-Building Simulations

In order for robots to service humans more broadly, they’re going to be required to master more complex tasks at smaller scales. Right now, many of the most effective robot assignments involve “muscle,” like moving large car parts from a conveyor belt to a model. Humans are always called in to complete the final touches, though. Researchers at University of Southern California thought of a particular way to teach robots (at least in simulations, for now) how to approach “a multifaceted problem that can teach the machines a multitude of lessons”—in this instance, building IKEA furniture. Humans know that building IKEA furniture can be challenging, but for robots, disassembled furniture is a puzzle that allows them to master careful movements through repetitious reinforcement training with seemingly endless variability. Read more at Wired.