/Sparse-Interaction-as-Anomaly

Official code for the paper "Multi-Agent Sparse Interaction Modeling is an Anomaly Detection Problem".

Primary LanguagePython

Multi-Agent Sparse Interaction Modeling is an Anomaly Detection Problem

Code for the paper "Multi-Agent Sparse Interaction Modeling is an Anomaly Detection Problem" submitted to ICASSP 2024.

This repository develops SIA algorithm on StarCraft Multi-Agent Challenge benchmark. In addition, we compare it with multiple baselines including MAVEN, QTRAN and QPLEX.

Requirements

To install requirements:

pip install -r requirements.txt

Training

To train the approach in the paper, run this command:

python main.py

You can select the training algorithm by setting --config='sia' or other available choices

Hyper-parameters

To modify the hyper-parameters of algorithms and environments, refer to:

src/config/algs/sia.yaml
src/config/default.yaml

src/config/envs/sc2.yaml

Note

This repository is developed based on PyMARL. And we have cited the SMAC paper in our work.