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].

Reward is enough for convex MDPs

Authors: Tom Zahavy, Brendan O'Donoghue, Guillaume Desjardins, Satinder Singh

NeurIPS 2021 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We explore this principle in our experiments. Finally, we show that choosing specific algorithms for the policy and cost players unifies several disparate branches of RL problems, such as apprenticeship learning, constrained MDPs, and pure exploration into a single framework, as we summarize in Table 1. which we explore empirically in Appendix F. Appendix F: Additional Experiments
Researcher Affiliation Industry Tom Zahavy Deep Mind, London EMAIL Brendan O Donoghue Deep Mind, London EMAIL Guillaume Desjardins Deep Mind, London EMAIL Satinder Singh Deep Mind, London EMAIL
Pseudocode Yes Algorithm 1: meta-algorithm for convex MDPs 1: Input: convex-concave payoff L : K Λ R, algorithms Algλ, Algπ, K N 2: for k = 1, . . . , K do 3: λk = Algλ(d1 π, . . . , dk 1 π ; L) 4: dk π = Algπ( λk) 5: end for 6: Return d K π = 1 K PK k=1 dk π, λK = 1 K PK k=1 λk
Open Source Code No The paper does not provide an explicit statement about releasing source code or a direct link to a code repository for the methodology described.
Open Datasets No The paper refers to
Dataset Splits No The paper does not provide specific dataset split information (percentages, sample counts, or explicit methodology) needed to reproduce data partitioning.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory amounts, or detailed computer specifications) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details, such as library or solver names with version numbers, needed to replicate the experiment.
Experiment Setup No The paper does not contain specific experimental setup details such as concrete hyperparameter values, training configurations, or system-level settings in the main text.