no code implementations • ACL 2020 • Daniel Fern{\'a}ndez-Gonz{\'a}lez, Carlos G{\'o}mez-Rodr{\'\i}guez
Transition-based parsers implemented with Pointer Networks have become the new state of the art in dependency parsing, excelling in producing labelled syntactic trees and outperforming graph-based models in this task.
Ranked #2 on Semantic Dependency Parsing on DM
no code implementations • ACL 2020 • Daniel Fern{\'a}ndez-Gonz{\'a}lez, Carlos G{\'o}mez-Rodr{\'\i}guez
Sequence-to-sequence constituent parsing requires a linearization to represent trees as sequences.
1 code implementation • NAACL 2019 • Daniel Fern{\'a}ndez-Gonz{\'a}lez, Carlos G{\'o}mez-Rodr{\'\i}guez
We propose a novel transition-based algorithm that straightforwardly parses sentences from left to right by building n attachments, with n being the length of the input sentence.
1 code implementation • EMNLP 2018 • Daniel Fern{\'a}ndez-Gonz{\'a}lez, Carlos G{\'o}mez-Rodr{\'\i}guez
In addition, by improving the performance of the state-of-the-art in-order shift-reduce parser, we achieve the best accuracy to date (92. 0 F1) obtained by a fully-supervised single-model greedy shift-reduce constituent parser on the WSJ benchmark.
no code implementations • NAACL 2018 • Daniel Fern{\'a}ndez-Gonz{\'a}lez, Carlos G{\'o}mez-Rodr{\'\i}guez
We propose an efficient dynamic oracle for training the 2-Planar transition-based parser, a linear-time parser with over 99{\%} coverage on non-projective syntactic corpora.
no code implementations • NAACL 2018 • Daniel Fern{\'a}ndez-Gonz{\'a}lez, Carlos G{\'o}mez-Rodr{\'\i}guez
We present a novel transition system, based on the Covington non-projective parser, introducing non-local transitions that can directly create arcs involving nodes to the left of the current focus positions.
no code implementations • ACL 2017 • Daniel Fern{\'a}ndez-Gonz{\'a}lez, Carlos G{\'o}mez-Rodr{\'\i}guez
Restricted non-monotonicity has been shown beneficial for the projective arc-eager dependency parser in previous research, as posterior decisions can repair mistakes made in previous states due to the lack of information.