SimpleHGN

SimpleHGN[KDD 2021]

• paper: Are we really making much progress? Revisiting, benchmarking,and refining heterogeneous graph neural networks

Basic Idea

• The model extend the original graph attention mechanism in GAT by including edge type information into attention calculation.
• At each layer, we calculate the coefficient:

$$ \alpha_{ij} = \frac{exp(LeakyReLU(a^T[Wh_i||Wh_j||W_r r_{\psi(<i,j>)}]))}{\Sigma_{k\in\mathcal{E}}{exp(LeakyReLU(a^T[Wh_i||Wh_k||W_r r_{\psi(<i,k>)}]))}}

$$

• Residual connection including Node residual

$$ h_i^{(l)} = \sigma(\Sigma_{j\in \mathcal{N}i} {\alpha{ij}^{(l)}W^{(l)}h_j^{(l-1)}} + h_i^{(l-1)}) $$

• where $h_i$ and $h_j$ is the features of the source and the target node. $r_{\psi(e)}$ is a $d$-dimension embedding for each edge type $\psi(e) \in T_e$.

• and Edge residual:

$$ \alpha_{ij}^{(l)} = (1-\beta)\alpha_{ij}^{(l)}+\beta\alpha_{ij}^{(l-1)} $$

• Finally, a multi-head attention is used.

How to run

• Clone the Openhgnn-DGL

  # For node classification task
  python main.py -m SimpleHGN -t node_classification -d imdb4MAGNN -g 0 --use_best_config

  If you do not have gpu, set -gpu -1.

  Supported dataset

  • imdb4MAGNN

  • Number of nodes

    movie	4278
    director	2081
    actor	5257

  • Number of edges

    movie-director	4278
    movie-actor	12828

• dblp4MAGNN

  • Number of nodes

    author	4057
    paper	14328
    term	7723
    venue	20

  • Number of edges

    author-paper	19645
    paper-term	85810
    paper-venue	14328

Performance

Task: Node classification

Evaluation metric: accuracy

Dataset	HGBn-ACM	HGBn-DBLP	imdb4MAGNN	dblp4MAGNN
Macro_f1	66.64	86.31	48.78	86.79
Micro_f1	88.40	87.24	52.25	86.75

TrainerFlow: node classification flow

Hyper-parameter specific to the model

You can modify the parameters[SimpleHGN] in openhgnn/config.ini.

More

Contirbutor

Yaoqi Liu[GAMMA LAB]

If you have any questions,

Submit an issue or email to [email protected].