ToupleGDD

Implementation of "ToupleGDD: A Fine-Designed Solution of Influence Maximization by Deep Reinforcement Learning" (https://arxiv.org/abs/2210.07500)

Run the code

Train ToupleGDD model

python main.py --graph train_data \
                 --model Tripling \
                 --budget 5 \
                 --epoch 20000 \
                 --lr 0.001 \
                 --bs 16 \
                 --n_step 1

Test ToupleGDD model

python main.py --graph test_data/Wiki-2.txt \
                 --model Tripling \
                 --model_file tripling.ckpt \
                 --budget 10 \
                 --test

Train S2V-DQN model

python main.py --graph train_data \
                 --model S2V_DQN \
                 --model_file s2vdqn.ckpt \
                 --budget 5 \
                 --epoch 20000 \
                 --lr 0.001 \
                 --bs 16 \
                 --n_step 1

Test S2V-DQN model

python main.py --graph test_data/Wiki-2.txt \
                 --model S2V_DQN \
                 --model_file s2vdqn.ckpt \
                 --budget 10 \
                 --test

More instructions

Tips

Dependency requirement

• Python 3.6.13
• NumPy 1.19.5
• PyTorch 1.10.1+cu102
• PyG (PyTorch Geometric) 2.0.3
• PyTorch Scatter 2.0.9
• Tqdm 4.64.0
• SciPy 1.5.4

Code files

• main.py: load program arguments, graphs and set up RL agent and environment.
• runner.py: conduct simulation, train and test RL agent.
• models.py: define parameters and structures of S2V_DQN and ToupleGDD.
• rl_agents.py: define agents to follow reinforcement learning procedure.
• environment.py: store the process of simulation.
• utils/graph_utils.py: utility functions to load graphs, run MonteCarlo/RR to estimate influence spread.

Reference

Please cite our work if you find our code/paper is useful to your work.

@article{chen2022touplegdd,
  title={ToupleGDD: A Fine-Designed Solution of Influence Maximization by Deep Reinforcement Learning},
  author={Chen, Tiantian and Yan, Siwen and Guo, Jianxiong and Wu, Weili},
  journal={arXiv preprint arXiv:2210.07500},
  year={2022}
}

License

This project is licensed under the terms of the MIT license.