The company’s DeepMind artificial intelligence subsidiary has developed an AI that has learned how to navigate like a human being, the company announced in a blog post. Specifically, DeepMind’s AI has developed a system of spacial awareness that mimics human’s and other mammal’s grid cells–specific cells in the brain that allow for vector-based navigation, which allow us to calculate the direction and a distance to a location even if we’ve never traveled that route before. What’s most impressive about the AI’s mimicking of mammalian grid cells is that the AI did it on its own–it wasn’t programmed to mimic them. As the company explained in a blog post:
As a first step, we trained a recurrent network to perform the task of localising itself in a virtual environment, using predominantly movement-related velocity signals. This ability is commonly used by mammals when moving through unfamiliar places or in situations where it is not easy to spot familiar landmarks (e.g. when navigating in the dark).
We found that grid-like representations (hereafter grid units) spontaneously emerged within the network – providing a striking convergence with the neural activity patterns observed in foraging mammals, and consistent with the notion that grid cells provide an efficient code for space.
Now that AI has spontaneously learned how to navigate to different places, maybe keep it as far away from those running robots as possible.