How Much Do Language Models Copy From Their Training Data? Evaluating Linguistic Novelty in Text Generation Using RAVEN

Autor: R. Thomas McCoy, Paul Smolensky, Tal Linzen, Jianfeng Gao, Asli Celikyilmaz

Publikováno v: Transactions of the Association for Computational Linguistics, Vol 11 (2023)

Externí odkaz: https://doaj.org/article/fcac0cb3289c4654a57733705ef21320

Representations of Syntax [MASK] Useful: Effects of Constituency and Dependency Structure in Recursive LSTMs

Autor: Tal Linzen, Michael A. Lepori, R. Thomas McCoy

Publikováno v: ACL

Sequence-based neural networks show significant sensitivity to syntactic structure, but they still perform less well on syntactic tasks than tree-based networks. Such tree-based networks can be provided with a constituency parse, a dependency parse,

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::199516cc5ee9baa7d0f70a6750ffee18
http://arxiv.org/abs/2005.00019

BERTs of a feather do not generalize together: Large variability in generalization across models with similar test set performance

Autor: Junghyun Min, Tal Linzen, R. Thomas McCoy

Publikováno v: BlackboxNLP@EMNLP

If the same neural network architecture is trained multiple times on the same dataset, will it make similar linguistic generalizations across runs? To study this question, we fine-tuned 100 instances of BERT on the Multi-genre Natural Language Infere

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::6049acf4f798a563ea0a0be4485c91a2
https://doi.org/10.18653/v1/2020.blackboxnlp-1.21

Does syntax need to grow on trees? Sources of hierarchical inductive bias in sequence-to-sequence networks

Autor: Tal Linzen, R. Thomas McCoy, Robert Frank

Publikováno v: Transactions of the Association for Computational Linguistics, Vol 8, Pp 125-140 (2020)

Learners that are exposed to the same training data might generalize differently due to differing inductive biases. In neural network models, inductive biases could in theory arise from any aspect of the model architecture. We investigate which archi

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::7eeaacd73a8b49beacb27bf450dfa531

TAG Parsing with Neural Networks and Vector Representations of Supertags

Autor: Owen Rambow, Bob Frank, R. Thomas McCoy, Alexis Nasr, Jungo Kasai

Publikováno v: EMNLP
Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing
Conference on Empirical Methods in Natural Language Processing
Conference on Empirical Methods in Natural Language Processing, Sep 2017, Copenhague, Denmark. pp.1712-1722

International audience; We present supertagging-based models for Tree Adjoining Grammar parsing that use neural network architectures and dense vector representation of supertags (elementary trees) to achieve state-of-the-art performance in unlabeled

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::7247ca77b841e1207857a50021d0a230
https://doi.org/10.18653/v1/d17-1180