@inproceedings{bhattacharya-van-schijndel-2020-filler,
title = "Filler-gaps that neural networks fail to generalize",
author = "Bhattacharya, Debasmita and
van Schijndel, Marten",
editor = "Fern{\'a}ndez, Raquel and
Linzen, Tal",
booktitle = "Proceedings of the 24th Conference on Computational Natural Language Learning",
month = nov,
year = "2020",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2020.conll-1.39/",
doi = "10.18653/v1/2020.conll-1.39",
pages = "486--495",
abstract = "It can be difficult to separate abstract linguistic knowledge in recurrent neural networks (RNNs) from surface heuristics. In this work, we probe for highly abstract syntactic constraints that have been claimed to govern the behavior of filler-gap dependencies across different surface constructions. For models to generalize abstract patterns in expected ways to unseen data, they must share representational features in predictable ways. We use cumulative priming to test for representational overlap between disparate filler-gap constructions in English and find evidence that the models learn a general representation for the existence of filler-gap dependencies. However, we find no evidence that the models learn any of the shared underlying grammatical constraints we tested. Our work raises questions about the degree to which RNN language models learn abstract linguistic representations."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="bhattacharya-van-schijndel-2020-filler">
<titleInfo>
<title>Filler-gaps that neural networks fail to generalize</title>
</titleInfo>
<name type="personal">
<namePart type="given">Debasmita</namePart>
<namePart type="family">Bhattacharya</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marten</namePart>
<namePart type="family">van Schijndel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2020-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 24th Conference on Computational Natural Language Learning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Raquel</namePart>
<namePart type="family">Fernández</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tal</namePart>
<namePart type="family">Linzen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Online</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>It can be difficult to separate abstract linguistic knowledge in recurrent neural networks (RNNs) from surface heuristics. In this work, we probe for highly abstract syntactic constraints that have been claimed to govern the behavior of filler-gap dependencies across different surface constructions. For models to generalize abstract patterns in expected ways to unseen data, they must share representational features in predictable ways. We use cumulative priming to test for representational overlap between disparate filler-gap constructions in English and find evidence that the models learn a general representation for the existence of filler-gap dependencies. However, we find no evidence that the models learn any of the shared underlying grammatical constraints we tested. Our work raises questions about the degree to which RNN language models learn abstract linguistic representations.</abstract>
<identifier type="citekey">bhattacharya-van-schijndel-2020-filler</identifier>
<identifier type="doi">10.18653/v1/2020.conll-1.39</identifier>
<location>
<url>https://aclanthology.org/2020.conll-1.39/</url>
</location>
<part>
<date>2020-11</date>
<extent unit="page">
<start>486</start>
<end>495</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Filler-gaps that neural networks fail to generalize
%A Bhattacharya, Debasmita
%A van Schijndel, Marten
%Y Fernández, Raquel
%Y Linzen, Tal
%S Proceedings of the 24th Conference on Computational Natural Language Learning
%D 2020
%8 November
%I Association for Computational Linguistics
%C Online
%F bhattacharya-van-schijndel-2020-filler
%X It can be difficult to separate abstract linguistic knowledge in recurrent neural networks (RNNs) from surface heuristics. In this work, we probe for highly abstract syntactic constraints that have been claimed to govern the behavior of filler-gap dependencies across different surface constructions. For models to generalize abstract patterns in expected ways to unseen data, they must share representational features in predictable ways. We use cumulative priming to test for representational overlap between disparate filler-gap constructions in English and find evidence that the models learn a general representation for the existence of filler-gap dependencies. However, we find no evidence that the models learn any of the shared underlying grammatical constraints we tested. Our work raises questions about the degree to which RNN language models learn abstract linguistic representations.
%R 10.18653/v1/2020.conll-1.39
%U https://aclanthology.org/2020.conll-1.39/
%U https://doi.org/10.18653/v1/2020.conll-1.39
%P 486-495
Markdown (Informal)
[Filler-gaps that neural networks fail to generalize](https://aclanthology.org/2020.conll-1.39/) (Bhattacharya & van Schijndel, CoNLL 2020)
ACL
- Debasmita Bhattacharya and Marten van Schijndel. 2020. Filler-gaps that neural networks fail to generalize. In Proceedings of the 24th Conference on Computational Natural Language Learning, pages 486–495, Online. Association for Computational Linguistics.