@inproceedings{vu-etal-2018-grounded,
title = "Grounded Textual Entailment",
author = "Vu, Hoa Trong and
Greco, Claudio and
Erofeeva, Aliia and
Jafaritazehjan, Somayeh and
Linders, Guido and
Tanti, Marc and
Testoni, Alberto and
Bernardi, Raffaella and
Gatt, Albert",
booktitle = "Proceedings of the 27th International Conference on Computational Linguistics",
month = aug,
year = "2018",
address = "Santa Fe, New Mexico, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/C18-1199",
pages = "2354--2368",
abstract = "Capturing semantic relations between sentences, such as entailment, is a long-standing challenge for computational semantics. Logic-based models analyse entailment in terms of possible worlds (interpretations, or situations) where a premise P entails a hypothesis H iff in all worlds where P is true, H is also true. Statistical models view this relationship probabilistically, addressing it in terms of whether a human would likely infer H from P. In this paper, we wish to bridge these two perspectives, by arguing for a visually-grounded version of the Textual Entailment task. Specifically, we ask whether models can perform better if, in addition to P and H, there is also an image (corresponding to the relevant {``}world{''} or {``}situation{''}). We use a multimodal version of the SNLI dataset (Bowman et al., 2015) and we compare {``}blind{''} and visually-augmented models of textual entailment. We show that visual information is beneficial, but we also conduct an in-depth error analysis that reveals that current multimodal models are not performing {``}grounding{''} in an optimal fashion.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="vu-etal-2018-grounded">
<titleInfo>
<title>Grounded Textual Entailment</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hoa</namePart>
<namePart type="given">Trong</namePart>
<namePart type="family">Vu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Claudio</namePart>
<namePart type="family">Greco</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aliia</namePart>
<namePart type="family">Erofeeva</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Somayeh</namePart>
<namePart type="family">Jafaritazehjan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guido</namePart>
<namePart type="family">Linders</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marc</namePart>
<namePart type="family">Tanti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alberto</namePart>
<namePart type="family">Testoni</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Raffaella</namePart>
<namePart type="family">Bernardi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Albert</namePart>
<namePart type="family">Gatt</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2018-aug</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 27th International Conference on Computational Linguistics</title>
</titleInfo>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Santa Fe, New Mexico, USA</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Capturing semantic relations between sentences, such as entailment, is a long-standing challenge for computational semantics. Logic-based models analyse entailment in terms of possible worlds (interpretations, or situations) where a premise P entails a hypothesis H iff in all worlds where P is true, H is also true. Statistical models view this relationship probabilistically, addressing it in terms of whether a human would likely infer H from P. In this paper, we wish to bridge these two perspectives, by arguing for a visually-grounded version of the Textual Entailment task. Specifically, we ask whether models can perform better if, in addition to P and H, there is also an image (corresponding to the relevant “world” or “situation”). We use a multimodal version of the SNLI dataset (Bowman et al., 2015) and we compare “blind” and visually-augmented models of textual entailment. We show that visual information is beneficial, but we also conduct an in-depth error analysis that reveals that current multimodal models are not performing “grounding” in an optimal fashion.</abstract>
<identifier type="citekey">vu-etal-2018-grounded</identifier>
<location>
<url>https://aclanthology.org/C18-1199</url>
</location>
<part>
<date>2018-aug</date>
<extent unit="page">
<start>2354</start>
<end>2368</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Grounded Textual Entailment
%A Vu, Hoa Trong
%A Greco, Claudio
%A Erofeeva, Aliia
%A Jafaritazehjan, Somayeh
%A Linders, Guido
%A Tanti, Marc
%A Testoni, Alberto
%A Bernardi, Raffaella
%A Gatt, Albert
%S Proceedings of the 27th International Conference on Computational Linguistics
%D 2018
%8 aug
%I Association for Computational Linguistics
%C Santa Fe, New Mexico, USA
%F vu-etal-2018-grounded
%X Capturing semantic relations between sentences, such as entailment, is a long-standing challenge for computational semantics. Logic-based models analyse entailment in terms of possible worlds (interpretations, or situations) where a premise P entails a hypothesis H iff in all worlds where P is true, H is also true. Statistical models view this relationship probabilistically, addressing it in terms of whether a human would likely infer H from P. In this paper, we wish to bridge these two perspectives, by arguing for a visually-grounded version of the Textual Entailment task. Specifically, we ask whether models can perform better if, in addition to P and H, there is also an image (corresponding to the relevant “world” or “situation”). We use a multimodal version of the SNLI dataset (Bowman et al., 2015) and we compare “blind” and visually-augmented models of textual entailment. We show that visual information is beneficial, but we also conduct an in-depth error analysis that reveals that current multimodal models are not performing “grounding” in an optimal fashion.
%U https://aclanthology.org/C18-1199
%P 2354-2368
Markdown (Informal)
[Grounded Textual Entailment](https://aclanthology.org/C18-1199) (Vu et al., COLING 2018)
ACL
- Hoa Trong Vu, Claudio Greco, Aliia Erofeeva, Somayeh Jafaritazehjan, Guido Linders, Marc Tanti, Alberto Testoni, Raffaella Bernardi, and Albert Gatt. 2018. Grounded Textual Entailment. In Proceedings of the 27th International Conference on Computational Linguistics, pages 2354–2368, Santa Fe, New Mexico, USA. Association for Computational Linguistics.