Sequential Causal Imitation Learning with Unobserved Confounders

Authors: Daniel Kumor, Junzhe Zhang, Elias Bareinboim

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, we provide an efficient algorithm for determining imitability and corroborate our theory with simulations. We performed 2 experiments (for full details, refer to Kumor et al. (2021, Appendix B)), comparing the performance of 4 separate approaches to determining which variables to include in an imitating policy:
Researcher Affiliation Academia Daniel Kumor Purdue University dkumor@purdue.edu Junzhe Zhang Columbia University junzhez@cs.columbia.edu Elias Bareinboim Columbia University eb@cs.columbia.edu
Pseudocode Yes Algorithm 1 Find largest valid OX in ancestral graph of Y given G, X and target Y
Open Source Code No The paper references a technical report for full details but does not provide a direct link to open-source code for the methodology described.
Open Datasets Yes The second simulation used a synthetic, adversarial causal model, enriched with continuous data from the High D dataset (Krajewski et al., 2018) altered to conform to the causal model
Dataset Splits No The paper does not specify exact training, validation, and test dataset splits (percentages or sample counts).
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory amounts) used for running the experiments.
Software Dependencies No The paper mentions that 'A neural network was trained for each action-policy pair using standard supervised learning approaches' but does not specify any software or library names with version numbers.
Experiment Setup No The paper mentions training a neural network using 'standard supervised learning approaches' but does not provide concrete hyperparameters (e.g., learning rate, batch size, number of epochs) or other detailed system-level training settings.