Skip to main content

Loaded DiCE: Trading off Bias and Variance in Any−Order Score Function Estimators for Reinforcement Learning

Gregory Farquhar‚ Shimon Whiteson and Jakob Foerster


Gradient-based methods for optimisation of objectives in stochastic settings with unknown or intractable dynamics require estimators of derivatives. We derive an objective that, under automatic differentiation, produces low-variance unbiased estimators of derivatives at any order. Our objective is compatible with arbitrary advantage estimators, which allows the control of the bias and variance of any-order derivatives when using function approximation. Furthermore, we propose a method to trade off bias and variance of higher order derivatives by discounting the impact of more distant causal dependencies. We demonstrate the correctness and utility of our objective in analytically tractable MDPs and in meta-reinforcement-learning for continuous control.

Book Title
NeurIPS 2019: Proceedings of the Thirty−third Annual Conference on Neural Information Processing Systems