/JAME

Joint Autoencoders for Multi-task Network Embedding

Primary LanguagePython

How to run

  • To generate embeddings run main.py

ASONAM paper

"Multi-task Network Embedding with Adaptive Loss Weighting"

Files

  • main.py: generates embeddings
  • loss_weighting: it's an adaptive loss weighting layer in neural networks
  • link_pred: link prediction
  • classify: node classification
  • att_infer: attribute prediction

Folders

  • Cora: Cora Dataset
  • Citeseer: Citeseer Dataset
  • PubMed: PubMed Dataset
  • results: contains result files
  • loss: contains loss values and task weights