Diffusion-based Reinforcement Learning for Edge-enabled AI-Generated Content Services

This repository is an implementation of the system design and the proposed Deep Diffusion Soft Actor-Critic (D2SAC) algorithm presented in:

"Diffusion-based Reinforcement Learning for Edge-enabled AI-Generated Content Services"

Authored by Hongyang Du, Zonghang Li, Dusit Niyato, Jiawen Kang, Zehui Xiong, Huawei Huang, and Shiwen Mao.

The paper can be found at ArXiv.

Please see docker/Dockerfile for running this project on Docker or use the prebuilt Docker image lizonghango00o1/agod:cpu from Dockerhub directly. Then, run python main.py.

If our code is useful to you, please cite:

Citation

@article{du2024diffusion,
  title={Diffusion-based Reinforcement Learning for Edge-enabled {AI}-Generated Content Services},
  author={Du, Hongyang and Li, Zonghang and Niyato, Dusit and Kang, Jiawen and Xiong, Zehui and Huang, Huawei and Mao, Shiwen},
  journal={IEEE Transactions on Mobile Computing},
  year={2024},
  volume={23},
  number={9},
  pages={8902-8918},
  doi={10.1109/TMC.2024.3356178},
  publisher={IEEE}
}