In this project, we perform neural dependency parsing using the ArcEager model with both Bi-LSTM and BERT as oracles.
-
Arc-Eager: Parser and static oracle
-
Dataset: Analysis and preprocessing
-
Baseline Model: Bi-LSTM (from scratch)
-
BERT-based model (from scratch)
-
Error analysis
-
Discussion: Procedures, results, and comparison with SOTA
| Dejan Dichoski | José Chacón Mejía |