<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="henry-etal-2020-query">
    <titleInfo>
        <title>Query-Key Normalization for Transformers</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Alex</namePart>
        <namePart type="family">Henry</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Prudhvi</namePart>
        <namePart type="given">Raj</namePart>
        <namePart type="family">Dachapally</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Shubham</namePart>
        <namePart type="given">Shantaram</namePart>
        <namePart type="family">Pawar</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yuxuan</namePart>
        <namePart type="family">Chen</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>Findings of the Association for Computational Linguistics: EMNLP 2020</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Trevor</namePart>
            <namePart type="family">Cohn</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Yulan</namePart>
            <namePart type="family">He</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Yang</namePart>
            <namePart type="family">Liu</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>Low-resource language translation is a challenging but socially valuable NLP task. Building on recent work adapting the Transformer’s normalization to this setting, we propose QKNorm, a normalization technique that modifies the attention mechanism to make the softmax function less prone to arbitrary saturation without sacrificing expressivity. Specifically, we apply l2-normalization along the head dimension of each query and key matrix prior to multiplying them and then scale up by a learnable parameter instead of dividing by the square root of the embedding dimension. We show improvements averaging 0.928 BLEU over state-of-the-art bilingual benchmarks for 5 low-resource translation pairs from the TED Talks corpus and IWSLT’15.</abstract>
    <identifier type="citekey">henry-etal-2020-query</identifier>
    <identifier type="doi">10.18653/v1/2020.findings-emnlp.379</identifier>
    <location>
        <url>https://aclanthology.org/2020.findings-emnlp.379</url>
    </location>
    <part>
        <date>2020-11</date>
        <extent unit="page">
            <start>4246</start>
            <end>4253</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Query-Key Normalization for Transformers
%A Henry, Alex
%A Dachapally, Prudhvi Raj
%A Pawar, Shubham Shantaram
%A Chen, Yuxuan
%Y Cohn, Trevor
%Y He, Yulan
%Y Liu, Yang
%S Findings of the Association for Computational Linguistics: EMNLP 2020
%D 2020
%8 November
%I Association for Computational Linguistics
%C Online
%F henry-etal-2020-query
%X Low-resource language translation is a challenging but socially valuable NLP task. Building on recent work adapting the Transformer’s normalization to this setting, we propose QKNorm, a normalization technique that modifies the attention mechanism to make the softmax function less prone to arbitrary saturation without sacrificing expressivity. Specifically, we apply l2-normalization along the head dimension of each query and key matrix prior to multiplying them and then scale up by a learnable parameter instead of dividing by the square root of the embedding dimension. We show improvements averaging 0.928 BLEU over state-of-the-art bilingual benchmarks for 5 low-resource translation pairs from the TED Talks corpus and IWSLT’15.
%R 10.18653/v1/2020.findings-emnlp.379
%U https://aclanthology.org/2020.findings-emnlp.379
%U https://doi.org/10.18653/v1/2020.findings-emnlp.379
%P 4246-4253

Markdown (Informal)

[Query-Key Normalization for Transformers](https://aclanthology.org/2020.findings-emnlp.379) (Henry et al., Findings 2020)

• Query-Key Normalization for Transformers (Henry et al., Findings 2020)

ACL

• Alex Henry, Prudhvi Raj Dachapally, Shubham Shantaram Pawar, and Yuxuan Chen. 2020. Query-Key Normalization for Transformers. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 4246–4253, Online. Association for Computational Linguistics.