Active inference is a central aspect of how Numenta describe TBT. But I did not come across a dedicated thread on the topic. This video from 2014 has a good explanation and a demo of active inference in a robotic system https://youtu.be/_epRLGXvLTU?t=471
@mthiboust I think this might be an example of what you were hoping to find when we discussed this in the last HLC call.
Thanks Mark. I didn’t know about Max Versace and his company Neurala.
The best examples of active sensing in robotic systems I have on my side are from the Bristol Robotics Lab and the Active Touch Laboratory in Sheffield:
However, those examples are quite old now.
Some recent news on the topic (from yesterday):
To test how the MultiPrednet performs in a body, the researchers teamed up with Martin Pearson at Bristol Robotics Lab. Together they integrated it into Whiskeye , a rodent-like robot that autonomously explores its environment, using head-mounted cameras for eyes, and 24 artificial whiskers to gather tactile information.
The way these networks “train” resembles how scientists think our brain learns: By constantly generating predictions about the world, comparing them to actual sensory inputs, and then adapting the network to avoid future error signals.