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; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1]

A Theoretical Justification for Asymmetric Actor-Critic Algorithms

Authors: Gaspard Lambrechts, Damien Ernst, Aditya Mahajan

ICML 2025 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We propose such a justification for asymmetric actor-critic algorithms with linear function approximators by adapting a finite-time convergence analysis to this setting. The resulting finite-time bound reveals that the asymmetric critic eliminates error terms arising from aliasing in the agent state.
Researcher Affiliation Academia 1Montefiore Institute, University of Li ege 2Department of Electrical and Computer Engineering, Mc Gill University. Correspondence to: Gaspard Lambrechts <EMAIL>.
Pseudocode Yes Algorithm 1 m-step temporal difference learning algorithm
Open Source Code No The paper does not contain any explicit statement about providing source code for the methodology described, nor does it provide a link to a code repository.
Open Datasets No The paper is theoretical and does not describe experiments involving specific datasets. It references the 'Tiger POMDP' as an example (Figure 1) but not as a dataset used for empirical evaluation with public access information.
Dataset Splits No The paper is theoretical and does not involve empirical experiments with datasets, thus it does not specify any dataset splits.
Hardware Specification No The paper focuses on theoretical analysis and does not describe any empirical experiments, therefore no hardware specifications are provided.
Software Dependencies No The paper is theoretical and does not include empirical experiments, so no specific software dependencies with version numbers are mentioned.
Experiment Setup No The paper presents a theoretical justification for algorithms and does not include any experimental results or specific details about an experimental setup, such as hyperparameters or training configurations.