Why Basic Motor Skills Are So Hard To Learn by Robots

Moravec’s paradox says that “contrary to traditional assumptions, high-level reasoning requires very little computation, but low-level sensorimotor skills require enormous computational resources“. In other words, our unconscious capabilities are much more difficult to implement than our high level, conscious abilities.

humanoid robot falling
A semi-autonomous robot having balance problems at the 2015 DARPA challenge

As Wikipedia explains, “One possible explanation of the paradox, offered by Moravec, is based on evolution. All human skills are implemented biologically, using machinery designed by the process of natural selection. In the course of their evolution, natural selection has tended to preserve design improvements and optimizations. The older a skill is, the more time natural selection has had to improve the design. Abstract thought developed only very recently, and consequently, we should not expect its implementation to be particularly efficient“.

So while computers became better at chess than humans a few years ago, getting a robot to reproduce our moves will still require some years, as shown in this video from the 2015 DARPA humanoid robot challenge.

There is not reason why we won’t be able to built fully balanced robots in a few years (it took a decade from the first DARPA autonomous driving vehicle challenge to have fully functioning prototypes on the roads).

Yet it is really amazing to realize that those deeply engrained bodily functions we take for granted are the most difficult to reproduce! – whereas we think they are rather low-level functions.

Share