ALMA: Hierarchical Learning for Composite Multi-Agent Tasks

Authors: Shariq Iqbal, Robby Costales, Fei Sha

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate ALMA on two challenging environments, described below (more in Appendix B).In our experimental validation we aim to answer the following questions: 1) Learning Efficacy: Is ALMA effective in improving learning efficiency and asymptotic performance in comparison to state-of-the-art hierarchical and non-hierarchical MARL methods?
Researcher Affiliation Collaboration Shariq Iqbal Deepmind Robby Costales University of Southern California Fei Sha Google Research
Pseudocode No The paper describes its methods but does not include any explicitly labeled pseudocode or algorithm blocks.
Open Source Code Yes Our code is available at https://github.com/shariqiqbal2810/ALMA
Open Datasets Yes STARCRAFT Our next environment is the Star Craft multi-agent challenge (SMAC) [22]
Dataset Splits No The paper discusses training and evaluation but does not specify explicit train/validation/test dataset splits (e.g., percentages or sample counts).
Hardware Specification No The paper does not provide specific hardware details such as GPU or CPU models, or memory specifications used for running experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., library names like PyTorch or TensorFlow with their respective versions).
Experiment Setup Yes All exploration probabilities are annealed over the course of training, and the annealing schedules (along with all hyperparameters) are provided in Appendix D.