Understanding Individual Agent Importance in Multi-Agent System via Counterfactual Reasoning

Authors: Jianming Chen, Yawen Wang, Junjie Wang, Xiaofei Xie, Jun Hu, Qing Wang, Fanjiang Xu

AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The experimental results in seven multi-agent tasks demonstrate that EMAI achieves higher fidelity in explanations than baselines and provides more effective guidance in practical applications concerning understanding policies, launching attacks, and patching policies.
Researcher Affiliation Academia 1 Institute of Software Chinese Academy of Sciences, Beijing, China 2 Science & Technology on Integrated Information System Laboratory, Beijing, China 3 State Key Laboratory of Intelligent Game, Beijing, China 4 University of Chinese Academy of Sciences, Beijing, China 5 Singapore Management University, Singapore EMAIL, EMAIL, EMAIL
Pseudocode Yes Algorithm 1: The training algorithm of EMAI.
Open Source Code No The paper does not provide an explicit statement or link to access the source code for the methodology described.
Open Datasets Yes Our experiments are conducted on three popular multi-agent benchmarks with different characteristics, selecting two to three environments from each benchmark as follows. Star Craft Multi-Agent Challenge (SMAC). SMAC (Samvelyan et al. 2019)... Google Research Football (GRF). GRF (Kurach et al. 2020)... Multi-Agent Particle Environments (MPE). MPE (Lowe et al. 2017)...
Dataset Splits No The paper mentions running experiments for a certain number of episodes (e.g., "For each experiment, we perform 500 episodes...") but does not provide specific training/test/validation dataset splits for the data used in the experiments or for training EMAI itself.
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details, such as library names with version numbers, needed to replicate the experiments. It mentions multi-agent environments but not the software stack used for implementation.
Experiment Setup No The paper mentions hyperparameters such as "β is the weight hyper-parameter of the sparsity constraints" and "λ is the weighting term to balance the two loss functions" but does not provide their concrete values or other specific experimental setup details like learning rates, batch sizes, or number of training epochs.