Batch Stationary Distribution Estimation

Authors: Junfeng Wen, Bo Dai, Lihong Li, Dale Schuurmans

ICML 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this section, we demonstrate the advantages of VPM in four representative applications. Due to space limit, experiment details are provided in Appendix D.
Researcher Affiliation Collaboration 1Department of Computing Science, University of Alberta, Edmonton, Canada 2Google Brain. Correspondence to: Junfeng Wen <junfengwen@gmail.com>.
Pseudocode Yes Algorithm 1: Variational Power Method
Open Source Code No The paper does not provide an explicit statement about the release of its source code or a link to a code repository.
Open Datasets No The paper mentions environments like Geo/Geo/1 queue, Ornstein-Uhlenbeck process, HMC, Taxi, Reacher, Half Cheetah, and Ant, but does not provide specific links, DOIs, repositories, or formal citations with author and year for public dataset access, nor does it explicitly state the datasets generated for experiments are publicly available.
Dataset Splits No The paper describes data collection and training, but does not provide specific train/validation/test splits (e.g., percentages, sample counts, or references to predefined splits).
Hardware Specification No The paper does not specify any particular hardware (e.g., GPU/CPU models, memory) used for running the experiments.
Software Dependencies No The paper mentions using 'Adam optimizer' and 'neural network' but does not provide specific version numbers for these or other software dependencies.
Experiment Setup Yes Fig. 6 shows the MMD curves for the MCMC funnel dataset in Section 5.3, using different learning rates, number of inner optimization steps and the regularization λ. Other datasets show similar trends. ... Default (lr, M, λ) = (0.001, 10, 0.5).