A recent paper by Lillicrap revisits the connection between deep learning models and brain and its biological plausibility. In their work, they show that it is possible to learn with fixed feedback connections that carry modulatory signals back to the upstream neurons.

Question: Is there evidence suggesting that feedback connections in brain may be fixed and only forward synapses are updated?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.