Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].

Asymptotically Best Causal Effect Identification with Multi-Armed Bandits

Authors: Alan Malek, Silvia Chiappa

NeurIPS 2021 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We validate our method by providing upper bounds on the sample complexity and an empirical study on artificially generated data.
Researcher Affiliation Industry Alan Malek Deep Mind London EMAIL Silvia Chiappa Deep Mind London EMAIL
Pseudocode Yes Algorithm 1 CS-LUCB; Algorithm 2 CS-SE
Open Source Code Yes The code implementing the experiments is available at github.com/deepmind/abcei_mab.
Open Datasets No The paper uses 'artificially generated data' and describes the data generation process but does not provide concrete access information to a pre-existing public or open dataset.
Dataset Splits No The paper mentions data folds (Dη, Dτ, Dσ) used internally by the estimator but does not specify overall training, validation, and test dataset splits with percentages or sample counts for the experiments.
Hardware Specification No The paper does not provide specific hardware details (e.g., CPU, GPU models, memory) used for running the experiments.
Software Dependencies No The paper mentions software components like 'ridge regression', 'logistic regression', and 'gradient-boosted regression trees' but does not specify their version numbers.
Experiment Setup Yes For the AIPW estimator, we used ridge regression and logistic regression to fit µx(Z) := Ep[Y |X = x, Z] and ex(Z) := p(X|Z) respectively. ... confidence intervals for logistic regression were approximated by a standard central limit theorem (CLT) confidence interval, h ˆσ2 + zαn/2 std(ˆσ2) n , ˆσ2 + z1 αn/2 std(ˆσ2) n i , with δ = 0.1 and αn = 6δ/πn2, following the construction described in Section 3.5.