ASAP

ASAP: Adaptive Structure Aware Pooling for Learning Hierarchical Graph Representations

Source code for AAAI 2020 paper: ASAP: Adaptive Structure Aware Pooling for Learning Hierarchical Graph Representation

Overview of ASAP: ASAP initially considers all possible local clusters with a fixed receptive field for a given input graph. It then computes the cluster membership of the nodes using an attention mechanism. These clusters are then scored using a GNN. Further, a fraction of the top scoring clusters are selected as nodes in the pooled graph and new edge weights are computed between neighboring clusters. Please refer to Section 4 of the paper for details.

File Descriptions

• main.py - contains the driver code for the whole project
• asap_pool.py - source code for ASAP pooling operator proposed in the paper
• le_conv.py - source code for LEConv GNN used in the paper
• asap_pool_model.py - a network which uses ASAP pooling as pooling operator

Dependencies

• Python 3.x
• Pytorch (1.5)
• Pytorch_Scatter (2.0.4)
• Pytorch_Sparse (0.6.3)
• Pytorch_Geometric (1.4.3)

Use the following commands to install the above version of dependency:

pip install torch==1.5.0+${CUDA} -f https://download.pytorch.org/whl/torch_stable.html
pip install torch-scatter==2.0.4+${CUDA} -f https://pytorch-geometric.com/whl/torch-1.5.0.html
pip install torch-sparse==0.6.3+${CUDA} -f https://pytorch-geometric.com/whl/torch-1.5.0.html
pip install torch-geometric==1.4.3

where where ${CUDA} should be replaced by either cpu, cu92, cu101 or cu102 depending on your PyTorch installation and CUDA version.

E.g., if your CUDA version is 9.2 then run:

pip install torch==1.5.0+cu92 -f https://download.pytorch.org/whl/torch_stable.html
pip install torch-scatter==2.0.4+cu92 -f https://pytorch-geometric.com/whl/torch-1.5.0.html
pip install torch-sparse==0.6.3+cu92 -f https://pytorch-geometric.com/whl/torch-1.5.0.html
pip install torch-geometric==1.4.3

Training a model from scratch

Example for PROTEINS dataset:

python main.py -data PROTEINS -batch 128 -hid_dim 64 -dropout_att 0.1 -lr 0.01

Hyperparameters to reproduce reported scores in the paper

| Dataset | Batch Size | Hidden Dimension | Dropout | Learning rate | |---|---|---|---|---| | PROTEINS | 128 | 64 | 0.1| 0.01 | | FRANKENSTEIN | 128 | 32 | 0 | 0.001 | | NCI1 | 128 | 128 | 0 | 0.01 | | NCI109 | 128 | 128 | 0 | 0.01 | | DD | 64 | 16 | 0.3 | 0.01 |

Citation:

Please cite the following paper if you found it useful in your work.

@article{ranjan2019asap,
  title={{ASAP}: Adaptive Structure Aware Pooling for Learning Hierarchical Graph Representations},
  author={Ranjan, Ekagra and Sanyal, Soumya and Talukdar, Partha Pratim},
  journal={arXiv preprint arXiv:1911.07979},
  year={2019}
}

For any clarification, comments, or suggestions please create an issue or contact Ekagra.

Pytorch_Geometric

Available at PyG: Example