Source code for the paper titled "Combining GNNs and Sentence Encoders for Knowledge-aware Recommendations"
You will find the two datasets used in this work (movielens1m and dbbook), and also the scripts:
- to learn graph embeddings
- to learn word embeddings
- to train the deep model and get the predictions.
In addition, we added the pre-trained embeddings used for this work (dict saved as .pickle files), so to skip the embedding learning step. Source code is documented, so to make it easier to run it.