Skip to content

Latest commit

 

History

History
28 lines (20 loc) · 1002 Bytes

File metadata and controls

28 lines (20 loc) · 1002 Bytes

Compositional Embeddings

Learning vectors by composing the embeddings of the constituent words.

The compositional methods rely on the distributional vectors of both the noun compounds and their constituents. They are obtained by training the distributional embeddings, and then learning the composition function.

We implemented 4 composition functions:

To train:

bash train_composition_models.sh

To predict the embeddings of the test set:

bash predict_composition_models.sh