Nature Machine Intelligence reported on work that combines principles of monitoring internal physiological states of living organisms with cybernetics, reinforcement learning, and neuroscience. The goal is to create a framework for autonomous and adaptive artificial agents capable of acting within a physical body.

In living systems, changes in internal state are taken into account during interaction with complex environments and influence subsequent actions. Transferring this idea to artificial systems links their behavior not only to external signals but also to internal parameters.

The publication is based on a synopsis of metadata rather than the full text of the study. Therefore, from the available materials, it is impossible to establish which architectures, experiments, or metrics the authors used.