iCub humanoid robot meta-learning project at INSERM, France

A neuromimetic model was developped to describe neural mechanisms in the prefrontal cortex for decision-making and reinforcement learning. The model mimicks the way the prefrontal cortex uses reward information to update values associated to different possible actions, and to regulate exploration during decision-making (i.e. sometimes exploiting learned action values, and sometimes exploring by choosing suboptimal actions so as to gather new information).
Here the model is applied to a simple human-robot game. The robot has to find under which cube a star (reward) is hidden. the robot alternates between exploration (searching for the correct cube) and exploitation phases (repeating the same choice). In addition, the model mimicks the way the prefrontal cortex monitores performance and can associate some cues or task events to variations in performance. This enables the robot to learn by itself to recognize that some objects or events are associated with changes in the task and thus shall be followed by a re-exploration.

Category:

Science & Technology

Tags:

• robot
• humanoid
• reinforcement learning
• exploration
• neuroscience
• prefrontal cortex
• meta-learning
• iCub
• human-robot interaction
• HRI
• autonomous robot
• INSERM
• CNRS

[wlorenz65]

iCub does not "perceive such hand movement" as you describe in your video! What you have done is to hand-code a four-black-boxes-in-a-row detector, a vertical-plank-detector and a two-hands-in-the-middle detector. And then you have set up a big show with iCub in order to justify that your PFC model is correct.

If you were really serious then you would have implemented a complete brain and let it learn from the scratch. But the iCub body is not good enough for that, it contains too many bugs.