<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ilharco-etal-2019-large">
    <titleInfo>
        <title>Large-Scale Representation Learning from Visually Grounded Untranscribed Speech</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Gabriel</namePart>
        <namePart type="family">Ilharco</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yuan</namePart>
        <namePart type="family">Zhang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jason</namePart>
        <namePart type="family">Baldridge</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2019-11</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 23rd Conference on Computational Natural Language Learning (CoNLL)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Mohit</namePart>
            <namePart type="family">Bansal</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Aline</namePart>
            <namePart type="family">Villavicencio</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Hong Kong, China</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
    </relatedItem>
    <abstract>Systems that can associate images with their spoken audio captions are an important step towards visually grounded language learning. We describe a scalable method to automatically generate diverse audio for image captioning datasets. This supports pretraining deep networks for encoding both audio and images, which we do via a dual encoder that learns to align latent representations from both modalities. We show that a masked margin softmax loss for such models is superior to the standard triplet loss. We fine-tune these models on the Flickr8k Audio Captions Corpus and obtain state-of-the-art results—improving recall in the top 10 from 29.6% to 49.5%. We also obtain human ratings on retrieval outputs to better assess the impact of incidentally matching image-caption pairs that were not associated in the data, finding that automatic evaluation substantially underestimates the quality of the retrieved results.</abstract>
    <identifier type="citekey">ilharco-etal-2019-large</identifier>
    <identifier type="doi">10.18653/v1/K19-1006</identifier>
    <location>
        <url>https://aclanthology.org/K19-1006/</url>
    </location>
    <part>
        <date>2019-11</date>
        <extent unit="page">
            <start>55</start>
            <end>65</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Large-Scale Representation Learning from Visually Grounded Untranscribed Speech
%A Ilharco, Gabriel
%A Zhang, Yuan
%A Baldridge, Jason
%Y Bansal, Mohit
%Y Villavicencio, Aline
%S Proceedings of the 23rd Conference on Computational Natural Language Learning (CoNLL)
%D 2019
%8 November
%I Association for Computational Linguistics
%C Hong Kong, China
%F ilharco-etal-2019-large
%X Systems that can associate images with their spoken audio captions are an important step towards visually grounded language learning. We describe a scalable method to automatically generate diverse audio for image captioning datasets. This supports pretraining deep networks for encoding both audio and images, which we do via a dual encoder that learns to align latent representations from both modalities. We show that a masked margin softmax loss for such models is superior to the standard triplet loss. We fine-tune these models on the Flickr8k Audio Captions Corpus and obtain state-of-the-art results—improving recall in the top 10 from 29.6% to 49.5%. We also obtain human ratings on retrieval outputs to better assess the impact of incidentally matching image-caption pairs that were not associated in the data, finding that automatic evaluation substantially underestimates the quality of the retrieved results.
%R 10.18653/v1/K19-1006
%U https://aclanthology.org/K19-1006/
%U https://doi.org/10.18653/v1/K19-1006
%P 55-65

Markdown (Informal)

[Large-Scale Representation Learning from Visually Grounded Untranscribed Speech](https://aclanthology.org/K19-1006/) (Ilharco et al., CoNLL 2019)

• Large-Scale Representation Learning from Visually Grounded Untranscribed Speech (Ilharco et al., CoNLL 2019)

ACL

• Gabriel Ilharco, Yuan Zhang, and Jason Baldridge. 2019. Large-Scale Representation Learning from Visually Grounded Untranscribed Speech. In Proceedings of the 23rd Conference on Computational Natural Language Learning (CoNLL), pages 55–65, Hong Kong, China. Association for Computational Linguistics.