@inproceedings{zheng-lapata-2021-compositional-generalization,
title = "Compositional Generalization via Semantic Tagging",
author = "Zheng, Hao and
Lapata, Mirella",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2021",
month = nov,
year = "2021",
address = "Punta Cana, Dominican Republic",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.findings-emnlp.88",
doi = "10.18653/v1/2021.findings-emnlp.88",
pages = "1022--1032",
abstract = "Although neural sequence-to-sequence models have been successfully applied to semantic parsing, they fail at compositional generalization, i.e., they are unable to systematically generalize to unseen compositions of seen components. Motivated by traditional semantic parsing where compositionality is explicitly accounted for by symbolic grammars, we propose a new decoding framework that preserves the expressivity and generality of sequence-to-sequence models while featuring lexicon-style alignments and disentangled information processing. Specifically, we decompose decoding into two phases where an input utterance is first tagged with semantic symbols representing the meaning of individual words, and then a sequence-to-sequence model is used to predict the final meaning representation conditioning on the utterance and the predicted tag sequence. Experimental results on three semantic parsing datasets show that the proposed approach consistently improves compositional generalization across model architectures, domains, and semantic formalisms.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zheng-lapata-2021-compositional-generalization">
<titleInfo>
<title>Compositional Generalization via Semantic Tagging</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hao</namePart>
<namePart type="family">Zheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mirella</namePart>
<namePart type="family">Lapata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2021-nov</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2021</title>
</titleInfo>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Punta Cana, Dominican Republic</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Although neural sequence-to-sequence models have been successfully applied to semantic parsing, they fail at compositional generalization, i.e., they are unable to systematically generalize to unseen compositions of seen components. Motivated by traditional semantic parsing where compositionality is explicitly accounted for by symbolic grammars, we propose a new decoding framework that preserves the expressivity and generality of sequence-to-sequence models while featuring lexicon-style alignments and disentangled information processing. Specifically, we decompose decoding into two phases where an input utterance is first tagged with semantic symbols representing the meaning of individual words, and then a sequence-to-sequence model is used to predict the final meaning representation conditioning on the utterance and the predicted tag sequence. Experimental results on three semantic parsing datasets show that the proposed approach consistently improves compositional generalization across model architectures, domains, and semantic formalisms.</abstract>
<identifier type="citekey">zheng-lapata-2021-compositional-generalization</identifier>
<identifier type="doi">10.18653/v1/2021.findings-emnlp.88</identifier>
<location>
<url>https://aclanthology.org/2021.findings-emnlp.88</url>
</location>
<part>
<date>2021-nov</date>
<extent unit="page">
<start>1022</start>
<end>1032</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Compositional Generalization via Semantic Tagging
%A Zheng, Hao
%A Lapata, Mirella
%S Findings of the Association for Computational Linguistics: EMNLP 2021
%D 2021
%8 nov
%I Association for Computational Linguistics
%C Punta Cana, Dominican Republic
%F zheng-lapata-2021-compositional-generalization
%X Although neural sequence-to-sequence models have been successfully applied to semantic parsing, they fail at compositional generalization, i.e., they are unable to systematically generalize to unseen compositions of seen components. Motivated by traditional semantic parsing where compositionality is explicitly accounted for by symbolic grammars, we propose a new decoding framework that preserves the expressivity and generality of sequence-to-sequence models while featuring lexicon-style alignments and disentangled information processing. Specifically, we decompose decoding into two phases where an input utterance is first tagged with semantic symbols representing the meaning of individual words, and then a sequence-to-sequence model is used to predict the final meaning representation conditioning on the utterance and the predicted tag sequence. Experimental results on three semantic parsing datasets show that the proposed approach consistently improves compositional generalization across model architectures, domains, and semantic formalisms.
%R 10.18653/v1/2021.findings-emnlp.88
%U https://aclanthology.org/2021.findings-emnlp.88
%U https://doi.org/10.18653/v1/2021.findings-emnlp.88
%P 1022-1032
Markdown (Informal)
[Compositional Generalization via Semantic Tagging](https://aclanthology.org/2021.findings-emnlp.88) (Zheng & Lapata, Findings 2021)
ACL
- Hao Zheng and Mirella Lapata. 2021. Compositional Generalization via Semantic Tagging. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 1022–1032, Punta Cana, Dominican Republic. Association for Computational Linguistics.