[PDF] mit.edu

What should/do/can LSTMs learn when parsing auxiliary verb constructions?

M De Lhoneux, S Stymne, J Nivre - Computational Linguistics, 2021 - direct.mit.edu
… 30 epochs for all these treebanks and pick the model of the best … (2018) except for the
character BiLSTM values: We set the … We follow Belinkov (2018) and choose to work with an MLP …
被引用次数：3 相关文章 所有 13 个版本
[PDF] arxiv.org

82 treebanks, 34 models: Universal dependency parsing with multi-treebank models

…, M De Lhoneux, J Nivre, Y Shao, S Stymne - arXiv preprint arXiv …, 2018 - arxiv.org
… We present the Uppsala system for the CoNLL 2018 Shared Task on universal dependency
… a single parsing model for each treebank, we trained models with multiple treebanks for one …
被引用次数：9 相关文章 所有 9 个版本
[PDF] diva-portal.org

[PDF][PDF] Universal Dependency Parsing at Uppsala University

…, M de Lhoneux, A Smith, S Stymne - Thanks to our sponsors …, 2018 - diva-portal.org
… where one or more languages have multiple treebanks. In the 2018 CoNLL Shared Task,
we use only 34 models to … 82 treebanks, 34 models: Universal dependency parsing with multi…
相关文章

82 Treebanks, 34 Models: Universal Dependency Parsing with Cross-Treebank Models

…, B Bohnet, J Nivre, M de Lhoneux, S Stymne, Y Shao - 2018 - research.google
… a single parsing model for each treebank, we trained models with multiple treebanks for the
same language or closely related languages, greatly reducing the number of models. On the …
被引用次数：1 相关文章
[PDF] arxiv.org

Recursive subtree composition in LSTM-based dependency parsing

M De Lhoneux, M Ballesteros, J Nivre - arXiv preprint arXiv:1902.09781, 2019 - arxiv.org
… 2018). This raises the question of whether features extracted with … Data We test our models
on a sample of treebanks from … We thank Sara Stymne and Aaron Smith for many discus…
被引用次数：14 相关文章 所有 10 个版本
[PDF] diva-portal.org

[PDF][PDF] Parameter Sharing in Multilingual Dependency Parsing

M de Lhoneux - Thanks to our sponsors: Gold: Stora Skuggans …, 2018 - diva-portal.org
… Since our parser has three basic sets of model parameters, we … For all 10 languages, we
use treebanks from the Universal … I also thank Joakim Nivre, Sara Stymne and Aaron Smith for …
相关文章
[PDF] arxiv.org

Deep Contextualized Word Embeddings in Transition-Based and Graph-Based Dependency Parsing--A Tale of Two Parsers Revisited

…, M de Lhoneux, J Gontrum, E Fano, J Nivre - arXiv preprint arXiv …, 2019 - arxiv.org
… multilingual cased model provided by Google,34 which is trained … 82 treebanks, 34 models:
Universal dependency parsing … Smith, Miryam de Lhoneux, Sara Stymne, and Joakim Nivre. …
被引用次数：30 相关文章 所有 7 个版本
[PDF] arxiv.org

Parsing with pretrained language models, multiple datasets, and dataset embeddings

R Van Der Goot, M de Lhoneux - arXiv preprint arXiv:2112.03625, 2021 - arxiv.org
… We use two training setups for our baselines: 1) monolingual models (MONO) and 2)
models where all treebanks are concatenated (CONCAT). This is in contrast to Smith et al. (2018) …
被引用次数：6 相关文章 所有 7 个版本
[PDF] arxiv.org

Parameter sharing between dependency parsers for related languages

M de Lhoneux, J Bjerva, I Augenstein… - arXiv preprint arXiv …, 2018 - arxiv.org
… Since our parser has three basic sets of model parameters, we … Datasets For all 10 languages,
we use treebanks from the … sharing parameters when using larger treebanks, we find it in…
被引用次数：44 相关文章 所有 7 个版本
[PDF] arxiv.org

An investigation of the interactions between pre-trained word embeddings, character models and POS tags in dependency parsing

A Smith, M De Lhoneux, S Stymne, J Nivre - arXiv preprint arXiv …, 2018 - arxiv.org
… be related to the number of entities being embedded (Shao… 2018). In Table 4, we show
how LAS varies with a few values … character models are very effective for Finnish, a treebank …
被引用次数：17 相关文章 所有 8 个版本