<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="saunders-etal-2022-skeletal">
    <titleInfo>
        <title>Skeletal Graph Self-Attention: Embedding a Skeleton Inductive Bias into Sign Language Production</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Ben</namePart>
        <namePart type="family">Saunders</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Necati</namePart>
        <namePart type="given">Cihan</namePart>
        <namePart type="family">Camgöz</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Richard</namePart>
        <namePart type="family">Bowden</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2022-06</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 7th International Workshop on Sign Language Translation and Avatar Technology: The Junction of the Visual and the Textual: Challenges and Perspectives</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Eleni</namePart>
            <namePart type="family">Efthimiou</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Stavroula-Evita</namePart>
            <namePart type="family">Fotinea</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Thomas</namePart>
            <namePart type="family">Hanke</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">John</namePart>
            <namePart type="given">C</namePart>
            <namePart type="family">McDonald</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Dimitar</namePart>
            <namePart type="family">Shterionov</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Rosalee</namePart>
            <namePart type="family">Wolfe</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>European Language Resources Association</publisher>
            <place>
                <placeTerm type="text">Marseille, France</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
    </relatedItem>
    <abstract>Recent approaches to Sign Language Production (SLP) have adopted spoken language Neural Machine Translation (NMT) architectures, applied without sign-specific modifications. In addition, these works represent sign language as a sequence of skeleton pose vectors, projected to an abstract representation with no inherent skeletal structure. In this paper, we represent sign language sequences as a skeletal graph structure, with joints as nodes and both spatial and temporal connections as edges. To operate on this graphical structure, we propose Skeletal Graph Self-Attention (SGSA), a novel graphical attention layer that embeds a skeleton inductive bias into the SLP model. Retaining the skeletal feature representation throughout, we directly apply a spatio-temporal adjacency matrix into the self-attention formulation. This provides structure and context to each skeletal joint that is not possible when using a non-graphical abstract representation, enabling fluid and expressive sign language production. We evaluate our Skeletal Graph Self-Attention architecture on the challenging RWTH-PHOENIX-Weather-2014T (PHOENIX14T) dataset, achieving state-of-the-art back translation performance with an 8% and 7% improvement over competing methods for the dev and test sets.</abstract>
    <identifier type="citekey">saunders-etal-2022-skeletal</identifier>
    <location>
        <url>https://aclanthology.org/2022.sltat-1.15</url>
    </location>
    <part>
        <date>2022-06</date>
        <extent unit="page">
            <start>95</start>
            <end>102</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Skeletal Graph Self-Attention: Embedding a Skeleton Inductive Bias into Sign Language Production
%A Saunders, Ben
%A Camgöz, Necati Cihan
%A Bowden, Richard
%Y Efthimiou, Eleni
%Y Fotinea, Stavroula-Evita
%Y Hanke, Thomas
%Y McDonald, John C.
%Y Shterionov, Dimitar
%Y Wolfe, Rosalee
%S Proceedings of the 7th International Workshop on Sign Language Translation and Avatar Technology: The Junction of the Visual and the Textual: Challenges and Perspectives
%D 2022
%8 June
%I European Language Resources Association
%C Marseille, France
%F saunders-etal-2022-skeletal
%X Recent approaches to Sign Language Production (SLP) have adopted spoken language Neural Machine Translation (NMT) architectures, applied without sign-specific modifications. In addition, these works represent sign language as a sequence of skeleton pose vectors, projected to an abstract representation with no inherent skeletal structure. In this paper, we represent sign language sequences as a skeletal graph structure, with joints as nodes and both spatial and temporal connections as edges. To operate on this graphical structure, we propose Skeletal Graph Self-Attention (SGSA), a novel graphical attention layer that embeds a skeleton inductive bias into the SLP model. Retaining the skeletal feature representation throughout, we directly apply a spatio-temporal adjacency matrix into the self-attention formulation. This provides structure and context to each skeletal joint that is not possible when using a non-graphical abstract representation, enabling fluid and expressive sign language production. We evaluate our Skeletal Graph Self-Attention architecture on the challenging RWTH-PHOENIX-Weather-2014T (PHOENIX14T) dataset, achieving state-of-the-art back translation performance with an 8% and 7% improvement over competing methods for the dev and test sets.
%U https://aclanthology.org/2022.sltat-1.15
%P 95-102

Markdown (Informal)

[Skeletal Graph Self-Attention: Embedding a Skeleton Inductive Bias into Sign Language Production](https://aclanthology.org/2022.sltat-1.15) (Saunders et al., SLTAT 2022)

• Skeletal Graph Self-Attention: Embedding a Skeleton Inductive Bias into Sign Language Production (Saunders et al., SLTAT 2022)

ACL

• Ben Saunders, Necati Cihan Camgöz, and Richard Bowden. 2022. Skeletal Graph Self-Attention: Embedding a Skeleton Inductive Bias into Sign Language Production. In Proceedings of the 7th International Workshop on Sign Language Translation and Avatar Technology: The Junction of the Visual and the Textual: Challenges and Perspectives, pages 95–102, Marseille, France. European Language Resources Association.