Neural Dependency Parser with Biaffine Attention and BERT Embeddings

This resource contains the code of the dependency parser used in the paper: Do and Rehbein (2020). Parsers Know Best: German PP Attachment Revisited. The parser is a re-implementation of the neural dependency parser from Dozat and Manning (2017) and is extended to use the BERT language model as input features. The pre-trained models on the German dataset of the SPMRL 2014 Shared Task used to report results in the paper are also included.