Sequential Graph Dependency Parser
arXiv:1905.10930
Abstract
We propose a method for non-projective dependency parsing by incrementally predicting a set of edges. Since the edges do not have a pre-specified order, we propose a set-based learning method. Our method blends graph, transition, and easy-first parsing, including a prior state of the parser as a special case. The proposed transition-based method successfully parses near the state of the art on both projective and non-projective languages, without assuming a certain parsing order.
arXiv commentRANLP 2019
References in corpus (6)
- Neural Message Passing for Quantum Chemistry
- Learning Deep Generative Models of Graphs
- Order Matters: Sequence to sequence for sets
- Insertion Transformer: Flexible Sequence Generation via Insertion Operations
- Insertion-based Decoding with automatically Inferred Generation Order
- Loss Functions for Multiset Prediction