Scalable Multi-Agent Reinforcement Learning through Intelligent Information Aggregation

Authors: Siddharth Nayak, Kenneth Choi, Wenqi Ding, Sydney Dolan, Karthik Gopalakrishnan, Hamsa Balakrishnan

ICML 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate our proposed model on four different navigation tasks by modifying the MAPE (Lowe et al., 2017b). Figure 3 shows the training performance of Infor MARL and the best-performing of the baselines mentioned above for the target task environment.
Researcher Affiliation Academia 1Massachusetts Institute of Technology, Cambridge, USA 2Stanford University, Stanford, USA.
Pseudocode No No pseudocode or algorithm blocks are present.
Open Source Code Yes Code available at https://github.com/nsidn98/Infor MARL.
Open Datasets Yes We evaluate our proposed model on four different navigation tasks by modifying the MAPE (Lowe et al., 2017b).
Dataset Splits No The paper discusses training and testing performance, but does not explicitly detail specific train/validation/test dataset splits with percentages or counts for reproducibility.
Hardware Specification No The authors would like to thank the MIT Super Cloud (Reuther et al., 2018) and the Lincoln Laboratory Supercomputing Center for providing high performance computing resources that have contributed to the research results reported within this paper.
Software Dependencies No We implemented Infor MARL by modifying the official codebase for MAPPO in Py Torch. (No version specified for PyTorch or any other library).
Experiment Setup Yes Tables 4, 5, 6, 7 show the hyperparameters for Infor MARL, MAPPO, MADDPG, MATD3, QMIX and VDN.