<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2023-pretraining">
    <titleInfo>
        <title>Pretraining Without Attention</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Junxiong</namePart>
        <namePart type="family">Wang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jing</namePart>
        <namePart type="given">Nathan</namePart>
        <namePart type="family">Yan</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Albert</namePart>
        <namePart type="family">Gu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Alexander</namePart>
        <namePart type="family">Rush</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2023-12</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: EMNLP 2023</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Houda</namePart>
            <namePart type="family">Bouamor</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Juan</namePart>
            <namePart type="family">Pino</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Kalika</namePart>
            <namePart type="family">Bali</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Singapore</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
    </relatedItem>
    <abstract>Transformers have been essential to pretraining success in NLP. While other architectures have been used, downstream accuracy is either significantly worse, or requires attention layers to match standard benchmarks such as GLUE. This work explores pretraining without attention by using recent advances in sequence routing based on state-space models (SSMs). Our proposed model, Bidirectional Gated SSM (BiGS), combines SSM layers with a multiplicative gating architecture that has been effective in simplified sequence modeling architectures. The model learns static layers that do not consider pair-wise interactions. Even so, BiGS is able to match BERT pretraining accuracy on GLUE and can be extended to long-form pretraining of 4096 tokens without approximation. Analysis shows that while the models have similar average accuracy, the approach has different inductive biases than BERT and scales more efficiently to longer sequences.</abstract>
    <identifier type="citekey">wang-etal-2023-pretraining</identifier>
    <identifier type="doi">10.18653/v1/2023.findings-emnlp.5</identifier>
    <location>
        <url>https://aclanthology.org/2023.findings-emnlp.5</url>
    </location>
    <part>
        <date>2023-12</date>
        <extent unit="page">
            <start>58</start>
            <end>69</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Pretraining Without Attention
%A Wang, Junxiong
%A Yan, Jing Nathan
%A Gu, Albert
%A Rush, Alexander
%Y Bouamor, Houda
%Y Pino, Juan
%Y Bali, Kalika
%S Findings of the Association for Computational Linguistics: EMNLP 2023
%D 2023
%8 December
%I Association for Computational Linguistics
%C Singapore
%F wang-etal-2023-pretraining
%X Transformers have been essential to pretraining success in NLP. While other architectures have been used, downstream accuracy is either significantly worse, or requires attention layers to match standard benchmarks such as GLUE. This work explores pretraining without attention by using recent advances in sequence routing based on state-space models (SSMs). Our proposed model, Bidirectional Gated SSM (BiGS), combines SSM layers with a multiplicative gating architecture that has been effective in simplified sequence modeling architectures. The model learns static layers that do not consider pair-wise interactions. Even so, BiGS is able to match BERT pretraining accuracy on GLUE and can be extended to long-form pretraining of 4096 tokens without approximation. Analysis shows that while the models have similar average accuracy, the approach has different inductive biases than BERT and scales more efficiently to longer sequences.
%R 10.18653/v1/2023.findings-emnlp.5
%U https://aclanthology.org/2023.findings-emnlp.5
%U https://doi.org/10.18653/v1/2023.findings-emnlp.5
%P 58-69

Markdown (Informal)

[Pretraining Without Attention](https://aclanthology.org/2023.findings-emnlp.5) (Wang et al., Findings 2023)

• Pretraining Without Attention (Wang et al., Findings 2023)

ACL

• Junxiong Wang, Jing Nathan Yan, Albert Gu, and Alexander Rush. 2023. Pretraining Without Attention. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 58–69, Singapore. Association for Computational Linguistics.