meettyj / rn2vec

Code & data for the CIKM'21 paper "Recipe Representation Learning with Networks".

Recipe Representation Learning with Networks

This is the official repo for the paper "Recipe Representation Learning with Networks", accepted at CIKM 2021.

Data

Please first download the data, then put them under folder 'data/'.

For the mappings between the nodes in RecipeNet and the recipe text (recipes_weighted_and_USDAmapped.json), please check the following code block:

recipeID2nodeID_dict, nodeID2recipeID_dict = torch.load('../data/recipeID2nodeID_and_nodeID2recipeID.pt')
ingreID2nodeID_dict, nodeID2ingreID_dict = torch.load('../data/ingre2nodeID_and_nodeID2ingre.pt')

Code

All codes have been put into a jupyter notebook for easy reading and executing. We further log the results in the notebook for demonstration purposes.

Citing rn2vec

If you find rn2vec useful, please cite our paper.

@inproceedings{rn2vec,
  author = {Tian, Yijun and Zhang, Chuxu and Metoyer, Ronald and Chawla, Nitesh V.},
  title = {Recipe Representation Learning with Networks},
  booktitle = {CIKM},
  year = {2021}
}