@inproceedings{lv-etal-2024-scalable,
title = "Scalable Efficient Training of Large Language Models with Low-dimensional Projected Attention",
author = "Lv, Xingtai and
Ding, Ning and
Zhang, Kaiyan and
Hua, Ermo and
Cui, Ganqu and
Zhou, Bowen",
editor = "Al-Onaizan, Yaser and
Bansal, Mohit and
Chen, Yun-Nung",
booktitle = "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing",
month = nov,
year = "2024",
address = "Miami, Florida, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.emnlp-main.808",
pages = "14588--14599",
abstract = "Improving the effectiveness and efficiency of large language models (LLMs) simultaneously is a critical yet challenging research goal. In this paper, we find that low-rank pre-training, normally considered as efficient methods that will compromise performance, can be scalably effective when reduced parameters are precisely targeted. Specifically, by applying low-dimensional module only to the attention layer {---} resolves this issue and enhances both effectiveness and efficiency. We refer to this structure as *Low-dimensional Projected Attention (LPA)* and provide an explanatory analysis. Through extensive experimentation at parameter scales of 130M, 370M, and scaling up to 3B, we have validated the effectiveness and scalability of LPA. Our results show that LPA model can save up to 12.4{\%} in time while achieving an approximate 5{\%} improvement in test perplexity (ppl) and on downstream tasks compared with vanilla Transformer.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lv-etal-2024-scalable">
<titleInfo>
<title>Scalable Efficient Training of Large Language Models with Low-dimensional Projected Attention</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xingtai</namePart>
<namePart type="family">Lv</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ning</namePart>
<namePart type="family">Ding</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kaiyan</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ermo</namePart>
<namePart type="family">Hua</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ganqu</namePart>
<namePart type="family">Cui</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bowen</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2024-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yaser</namePart>
<namePart type="family">Al-Onaizan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<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">Yun-Nung</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Miami, Florida, USA</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Improving the effectiveness and efficiency of large language models (LLMs) simultaneously is a critical yet challenging research goal. In this paper, we find that low-rank pre-training, normally considered as efficient methods that will compromise performance, can be scalably effective when reduced parameters are precisely targeted. Specifically, by applying low-dimensional module only to the attention layer — resolves this issue and enhances both effectiveness and efficiency. We refer to this structure as *Low-dimensional Projected Attention (LPA)* and provide an explanatory analysis. Through extensive experimentation at parameter scales of 130M, 370M, and scaling up to 3B, we have validated the effectiveness and scalability of LPA. Our results show that LPA model can save up to 12.4% in time while achieving an approximate 5% improvement in test perplexity (ppl) and on downstream tasks compared with vanilla Transformer.</abstract>
<identifier type="citekey">lv-etal-2024-scalable</identifier>
<location>
<url>https://aclanthology.org/2024.emnlp-main.808</url>
</location>
<part>
<date>2024-11</date>
<extent unit="page">
<start>14588</start>
<end>14599</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Scalable Efficient Training of Large Language Models with Low-dimensional Projected Attention
%A Lv, Xingtai
%A Ding, Ning
%A Zhang, Kaiyan
%A Hua, Ermo
%A Cui, Ganqu
%A Zhou, Bowen
%Y Al-Onaizan, Yaser
%Y Bansal, Mohit
%Y Chen, Yun-Nung
%S Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing
%D 2024
%8 November
%I Association for Computational Linguistics
%C Miami, Florida, USA
%F lv-etal-2024-scalable
%X Improving the effectiveness and efficiency of large language models (LLMs) simultaneously is a critical yet challenging research goal. In this paper, we find that low-rank pre-training, normally considered as efficient methods that will compromise performance, can be scalably effective when reduced parameters are precisely targeted. Specifically, by applying low-dimensional module only to the attention layer — resolves this issue and enhances both effectiveness and efficiency. We refer to this structure as *Low-dimensional Projected Attention (LPA)* and provide an explanatory analysis. Through extensive experimentation at parameter scales of 130M, 370M, and scaling up to 3B, we have validated the effectiveness and scalability of LPA. Our results show that LPA model can save up to 12.4% in time while achieving an approximate 5% improvement in test perplexity (ppl) and on downstream tasks compared with vanilla Transformer.
%U https://aclanthology.org/2024.emnlp-main.808
%P 14588-14599
Markdown (Informal)
[Scalable Efficient Training of Large Language Models with Low-dimensional Projected Attention](https://aclanthology.org/2024.emnlp-main.808) (Lv et al., EMNLP 2024)
ACL