On the Statistical Benefits of Temporal Difference Learning

Authors: David Cheikhi, Daniel Russo

ICML 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Figure 3 displays the Mean Square Error (MSE) of the TD and MC estimates for these quantities when the dataset contains n = 2000 independent trajectories. MSE calculations involve 10000 Monte-Carlo replications.
Researcher Affiliation Academia David Cheikhi 1 Daniel Russo 1 1Columbia University. Correspondence to: David Cheikhi <d.cheikhi@columbia.edu>.
Pseudocode No The paper does not contain any pseudocode or clearly labeled algorithm blocks.
Open Source Code No The paper does not provide any statement or link for open-source code availability.
Open Datasets No The paper mentions using a 'batch of trajectories' and 'dataset' but does not provide concrete access information (link, DOI, repository, or formal citation) for a publicly available dataset.
Dataset Splits No The paper mentions 'training data' and a 'dataset' used for calculations (e.g., 'n = 2000 independent trajectories'), but it does not provide specific details on train/validation/test dataset splits or cross-validation setup.
Hardware Specification No The paper does not provide any specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details with version numbers (e.g., library or solver names with version numbers like Python 3.8, PyTorch 1.9, CPLEX 12.4).
Experiment Setup Yes We consider a Layered MRP with width W = 5. We focus on state s(1) 1 and s(2) 1 and study the accuracy of the estimates of their value as we vary the horizon T of the MRP. ... the dataset contains n = 2000 independent trajectories. MSE calculations involve 10000 Monte-Carlo replications.