Keeping up with the latest AI research can be an odd experience. On the one hand, you’re aware that you’re looking at cutting-edge experimentation, with new papers outlining the ideas and methods that will probably (eventually) snowball into the biggest technological revolution of all time. On the other hand, sometimes what you’re looking at is just unavoidably weird and funny.

Case in point is a new paper from Google’s AI subsidiary DeepMind titled “Emergence of Locomotion Behaviours in Rich Environments.” The research explores how reinforcement learning (or RL) can be used to teach a computer to navigate unfamiliar and complex environments. It’s the sort of fundamental AI research that we’re now testing in virtual worlds, but that will one day help program robots that can navigate the stairs in your house. Don’t believe me? Just look at DeepMind’s agent go:

Everything the stick figure is doing in this video is self-taught. The jumping, the limboing, the leaping — all of these are behaviors that the computer has devised itself as the best way of getting from A to B. All DeepMind’s programmers have done is give the agent a set of virtual sensors (so it can tell whether it’s upright or not, for example) and then incentivize to move forward. The computer works the rest out for itself, using trial and error to come up with different ways of moving.

The novelty here is that the researchers are exploring how difficult environments can teach an agent complex and robust movements (i.e., using its knee to get purchase on top of a high wall). Usually, reinforcement learning produces behavior that is fragile, and that breaks down in unfamiliar circumstances, like a baby who knows how to tackle the stairs at home, but who can’t understand an escalator. This research shows that isn’t always the case, and that RL can be used to teach complex movements.

How complex? The GIF below should give you some idea: