Frustration and the angst of trying and not putting together Ikea furniture may seem like an exercise in humiliation for you, but know this: the particleboard nightmare may one day lead to robots that aren’t so dumb.
In recent years, roboticists have found that building Ikea furniture is actually a great way to teach robots how to deal with the chaos of the real world. A group of researchers coded a simulator in which virtual robot arms used trial and error to assemble the chairs. Others have managed to get a different set of robot arms to build Ikea chairs in the real world, however. it took them 20 minutes. And now, a helpful robot can help a human put together an Ikea bookcase by predicting which room they’ll want next and handing it over.
“It’s one of those things that’s easy to try, even if we break a few libraries in the lab, that’s okay,” says Stefanos Nikolaidis, roboticist at the University of Southern California, co-author of ‘a recent paper describing the research, which was presented in May at the International Conference on Robotics and Automation. “It’s pretty cheap. And it’s also something we all have to do at some point in our lives.
Nikolaidis and his colleagues began by studying how different people build an Ikea bookcase. Instead of providing them with this instruction sheet with pictograms, they asked subjects to improvise the order in which they set up the support boards for the frame, as well as the shelf inserts. (This is an important distinction, because the most important research question for this experiment is not about building furniture – more on that in a second.) Based on these findings, the researchers might cluster people in types or preferences. Some tie all the shelves to one of the frames, for example. Others attached a single shelf to both frames at a time. These are called action sequences.
They then asked the subjects to re-assemble, this time with a nearby robot arm to grab parts for them. The researcher wrote down the parts (shelves or racks) that the person started with, establishing a model for the robot. “Let’s say you go in and put the first shelf,” Nikolaidis says. “OK, the robot doesn’t know much. Then you choose the second shelf. And now you start to put the third shelf. Well, it is very, very likely that you belong to that group of users who put together all six shelves in a row. It is very very unlikely that you would then suddenly change your preference. Once the robot knows a person’s preference, it hands them the part that it knows people like them have previously chosen. The experiments showed that the robot could thus quickly and precisely adapt to the style of a human, successfully transmitting the right components.
Think of it like how AI researchers develop an image recognition algorithm: if you want to detect cats, you feed a neural network with heaps of feline images. Because it has already seen so many examples, the algorithm can then generalize. If you show him a photo of a cat he has never seen before, he can use his previous knowledge to confirm that he is indeed analyzing a furry four-legged mammal with a shitty attitude.
This robot does the same, but instead of using a static image bank, it relies on examples of sequences, the order in which humans put together shelves and racks, based on their preferences. “The robot knows that the next action it needs to do is hand you the next shelf, with very, very high certainty,” Nikolaidis explains.
Ultimately, however, this research is not about developing highly specialized bots that come to you and help you build libraries. Nor is it about developing machines that can perform complex tasks like this on their own. It’s about teaching robots to collaborate with humans without even driving them. After crazy that people already are when building Ikea furniture.
Despite all the hype around the arrival of robots to steal our jobs, the reality is that you are more likely to make a machine work with you than replace you outright. For now – and probably for a while in the future – people are just going to be much better at some tasks. No machine can replicate the dexterity of the human hand or come close to solving problems like we do. Which robots are well this is raw work. Think of an automotive assembly line: The robot’s arms lift the car doors into place, but the meticulous work requires a human touch.