Traditional robotics research has focused on precision, i.e. performing a
repetitive task as accurately as possible. This has resulted in very reliable
machines but with difficulties when adapting to changes in the environment.
Conversely, biological intelligence, understood as adaptation, has endorsed
reactive mechanisms that deal with unexpected circumstances to maintain their
optimal state for survival. Accidentally, these functionalities have evolved
into the backdrop of natural intelligence. How can we achieve natural
intelligence without paying attention to these basic mechanisms of perception
and control? Taking as inspiration how the brain may process sensory information
and generate control actions, my laboratory focus on developing artificial
intelligence for robotics that endorses adaptation Thus, in the long term, we
want to provide robots with human-like body perception and action. In this talk,
I will summarize the state-of-the-art of Active Inference in robotics through
examples (e.g., humanoid control, multidrone navigation), where it is explicit
that understanding human perception yields novel research directions that
improve current robotics and AI. I will end with some new applications that we
are pursuing in recently started projects, such as Metatool and Spikeference,
that will make the imagination fly into the future.
Reference: Active Inference in Robotics and Artificial Agents: Survey and
Challenges. https://arxiv.org/abs/2112.01871