Výsledky vyhledávání

Neural network surprisal predicts the existence but not the magnitude of human syntactic disambiguation difficulty

Autor: van Schijndel M, Linzen T

The disambiguation of a syntactically ambiguous sentence in favor of dispreferred parse can lead to slower reading at the disambiguation point. This phenomenon, referred to as a garden path effect, has motivated models in which readers only maintain

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::a5dbf0453affd08cc50eeda1f6b34de0
https://psyarxiv.com/7j8d6

Zobrazit plný text záznamu

On the Realization of Compositionality in Neural Networks

Autor: Baan, J., Leible, J., Nikolaus, M., Rau, D., Ulmer, D., Baumgärtner, T., Hupkes, D., Bruni, E., Linzen, T., Chrupała, G., Belinkov, Y.

Publikováno v: BlackboxNLP@ACL
The BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP at ACL 2019: ACL 2019 : proceedings of the Second Workshop : August 1, 2019, Florence, Italy, 127-137
STARTPAGE=127;ENDPAGE=137;TITLE=The BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP at ACL 2019

We present a detailed comparison of two types of sequence to sequence models trained to conduct a compositional task. The models are architecturally identical at inference time, but differ in the way that they are trained: our baseline model is train

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::acbe70d920f8c45f7f88d5ec57bf0ca9

Zobrazit plný text záznamu

Transcoding compositionally: using attention to find more generalizable solutions

Autor: Korrel, K., Hupkes, D., Dankers, V., Bruni, E., Linzen, T., Chrupała, G., Belinkov, Y.

Publikováno v: BlackboxNLP@ACL
The BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP at ACL 2019: ACL 2019 : proceedings of the Second Workshop : August 1, 2019, Florence, Italy, 1-11
STARTPAGE=1;ENDPAGE=11;TITLE=The BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP at ACL 2019

While sequence-to-sequence models have shown remarkable generalization power across several natural language tasks, their construct of solutions are argued to be less compositional than human-like generalization. In this paper, we present seq2attn, a

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::3dfd3394d444cd6dd2e2371b1892f989

Zobrazit plný text záznamu

Quantifiers in a Multimodal World: Hallucinating Vision with Language and Sound

Autor: Testoni, A., Pezzelle, S., Bernardi, R., Chersoni, E., Jacobs, C., Lenci, A., Linzen, T., Prévot, L., Santus, E.

Publikováno v: Cognitive Modeling and Computational Linguistics: NAACL HLT 2019 : proceedings of the workshop : June 7, 2019, Minneapolis, USA, 105-116
STARTPAGE=105;ENDPAGE=116;TITLE=Cognitive Modeling and Computational Linguistics

Inspired by the literature on multisensory integration, we develop a computational model to ground quantifiers in perception. The model learns to pick, out of nine quantifiers (‘few’, ‘many’, ‘all’, etc.), the one that is more likely to d

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::51a5091fbe78ede91509e33b3d4d0865
https://doi.org/10.18653/v1/w19-2912

Zobrazit plný text záznamu

Under the Hood: Using Diagnostic Classifiers to Investigate and Improve how Language Models Track Agreement Information

Autor: Giulianelli, M., Harding, J., Mohnert, F., Hupkes, D., Zuidema, W., Linzen, T., Chrupała, G., Alishahi, A.

Publikováno v: The 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP: EMNLP 2018 : proceedings of the First Workshop : November 1, 2018, Brussels, Belgium
The 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP
BlackboxNLP@EMNLP

How do neural language models keep track of number agreement between subject and verb? We show that `diagnostic classifiers', trained to predict number from the internal states of a language model, provide a detailed understanding of how, when, and w

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::b3a8352088ef7bf3eb78f74987208539
http://arxiv.org/abs/1808.08079

Zobrazit plný text záznamu

Do Language Models Understand Anything?

Autor: Jumelet, J., Hupkes, D., Linzen, T., Chrupała, G., Alishahi, A.

Publikováno v: The 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP: EMNLP 2018 : proceedings of the First Workshop : November 1, 2018, Brussels, Belgium, 222-231
STARTPAGE=222;ENDPAGE=231;TITLE=The 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP
BlackboxNLP@EMNLP

In this paper, we attempt to link the inner workings of a neural language model to linguistic theory, focusing on a complex phenomenon well discussed in formal linguis- tics: (negative) polarity items. We briefly discuss the leading hypotheses about

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::ad16e85cd398d57858c37a10f13567d4
https://doi.org/10.18653/v1/w18-5424

Zobrazit plný text záznamu

Jump to better conclusions: SCAN both left and right

Autor: Bastings, J., Baroni, M., Weston, J., Cho, K., Kiela, D., Linzen, T., Chrupała, G., Alishahi, A.

Publikováno v: BlackboxNLP@EMNLP
The 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP: EMNLP 2018 : proceedings of the First Workshop : November 1, 2018, Brussels, Belgium, 47-55
STARTPAGE=47;ENDPAGE=55;TITLE=The 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP

Lake & Baroni (2018) recently introduced the SCAN data set, which consists of simple commands paired with action sequences and is intended to test the strong generalization abilities of recurrent sequence-to-sequence models. Their initial experiments

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::ff5f14aec597705150babd052d683722
https://doi.org/10.18653/v1/w18-5407

Zobrazit plný text záznamu

Vyhledávací nástroje:

Upřesnit hledání