Dice: The infinitely differentiable monte carlo estimator

J Foerster, G Farquhar, M Al-Shedivat… - International …, 2018 - proceedings.mlr.press
… Correct Gradient Estimators with DiCE In this section, we propose the Infinitely Differentiable
Monte-Carlo Estimator (DICE), a practical algorithm for programatically generating correct …

DiCE: The Infinitely Differentiable Monte Carlo Estimator

S Whiteson - 2018 - cs.ox.ac.uk
… Correct Gradient Estimators with DiCE In this section, we propose the Infinitely Differentiable
Monte-Carlo Estimator (DICE), a practical algorithm for programatically generating correct …

A baseline for any order gradient estimation in stochastic computation graphs

J Mao, J Foerster, T Rocktäschel… - International …, 2019 - proceedings.mlr.press
… To improve the sample efficiency of DiCE, we propose a new … Instead, we must construct
Monte Carlo estimates of the … , the infinitely differentiable Monte-Carlo estimator (DiCE) (…

Automatic differentiable Monte Carlo: theory and application

SX Zhang, ZQ Wan, H Yao - Physical Review Research, 2023 - APS
… II, we review important background knowledge required to understand the general theory
of ADMC and its applications, including automatic differentiation, estimation on Monte Carlo

Loaded DiCE: Trading off bias and variance in any-order score function gradient estimators for reinforcement learning

G Farquhar, S Whiteson… - Advances in Neural …, 2019 - proceedings.neurips.cc
… pure Monte-Carlo estimates of the objective, introducing unacceptable variance in estimates
of … We compare the original DiCE estimator to Loaded DiCE, and the objective proposed by …

A Baseline for Any Order Gradient Estimation in Stochastic Computation Graphs

S Whiteson - 2019 - cs.ox.ac.uk
… To improve the sample efficiency of DiCE, we propose a new … Instead, we must construct
Monte Carlo estimates of the … , the infinitely differentiable Monte-Carlo estimator (DiCE) (…

[PDF][PDF] No DICE: An investigation of the bias-variance tradeoff in meta-gradients

R Vuorio, JA Beck, G Farquhar… - Deep RL Workshop …, 2021 - drive.google.com
estimation, implemented for example by DiCE and its variants, always add bias and can also
add variance to metagradient estimation… In our derivations we use the Monte Carlo return 고…

[BOOK][B] Monte carlo methods

A Barbu, SC Zhu - 2020 - Springer
… events, such as the tossing of a pair of dice to simulate the casino’s overall business model.
In … and can be approximated by Monte Carlo integration, as we will elaborate in Chap. 2…

[PDF][PDF] Loaded Dice in Monte Carlo

A van Hameren - arXiv preprint hep-ph/0101094, 2001 - core.ac.uk
… of the efficiency of the Quasi Monte Carlo method of numerical integration, which uses point
… with the application of the Monte Carlo method to phase space integration, and in particular …

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

S Whiteson - 2019 - cs.ox.ac.uk
estimates on the distributions they are sampled from. However, their formulation relies on …
Monte-Carlo estimates of the objective, introducing unacceptable variance in estimates of …