Neural Networks with Recurrent Generative Feedback

Understand

Neural networks are vulnerable to input perturbations such as additive noise and adversarial attacks.

• In contrast, human perception is much more robust to such perturbations.
• The Bayesian brain hypothesis states that human brains use an internal generative model to update the posterior beliefs of the sensory input.
• This mechanism can be interpreted as a form of self-consistency between the maximum a posteriori (MAP) estimation of an internal generative model and the external environment.