Learning through Model Differentiation

Policy Gradient Through Model

Method that calculates policy gradients by propagating rewards through a differentiable environment model.

← Kembali