Parsing as pretraining
Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
Standard
Parsing as pretraining. / Vilares, David; Strzyz, Michalina ; Søgaard, Anders; Gómez-Rodrıguez, Carlos .
Proceedings of the AAAI Conference on Artificial Intelligence (AAAI 2020). AAAI Press, 2020. p. 9114-9121.Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - GEN
T1 - Parsing as pretraining
AU - Vilares, David
AU - Strzyz, Michalina
AU - Søgaard, Anders
AU - Gómez-Rodrıguez, Carlos
PY - 2020
Y1 - 2020
N2 - Recent analyses suggest that encoders pretrained for language modeling capture certain morpho-syntactic structure. However, probing frameworks for word vectors still do not report results on standard setups such as constituent and dependency parsing. This paper addresses this problem and does full parsing (on English) relying only on pretraining architectures – and no decoding. We first cast constituent and dependency parsing as sequence tagging. We then use a single feed-forward layer to directly map word vectors to labels that encode a linearized tree. This is used to: (i) see how far we can reach on syntax modelling with just pretrained encoders, and (ii) shed some light about the syntax-sensitivity of different word vectors (by freezing the weights of the pretraining network during training). For evaluation, we use bracketing F1-score and las, and analyze in-depth differences across representations for span lengths and dependency displacements. The overall results surpass existing sequence tagging parsers on the ptb (93.5%) and end-to-end en-ewt ud (78.8%).
AB - Recent analyses suggest that encoders pretrained for language modeling capture certain morpho-syntactic structure. However, probing frameworks for word vectors still do not report results on standard setups such as constituent and dependency parsing. This paper addresses this problem and does full parsing (on English) relying only on pretraining architectures – and no decoding. We first cast constituent and dependency parsing as sequence tagging. We then use a single feed-forward layer to directly map word vectors to labels that encode a linearized tree. This is used to: (i) see how far we can reach on syntax modelling with just pretrained encoders, and (ii) shed some light about the syntax-sensitivity of different word vectors (by freezing the weights of the pretraining network during training). For evaluation, we use bracketing F1-score and las, and analyze in-depth differences across representations for span lengths and dependency displacements. The overall results surpass existing sequence tagging parsers on the ptb (93.5%) and end-to-end en-ewt ud (78.8%).
U2 - 10.1609/aaai.v34i05.6446
DO - 10.1609/aaai.v34i05.6446
M3 - Article in proceedings
SP - 9114-9121.
BT - Proceedings of the AAAI Conference on Artificial Intelligence (AAAI 2020)
PB - AAAI Press
T2 - Thirty-Forth AAAI Conference on Artificial Intelligence
Y2 - 7 February 2020 through 12 February 2020
ER -
ID: 258333711