@inproceedings{wu-etal-2022-pruning,
title = "Pruning Adatperfusion with Lottery Ticket Hypothesis",
author = "Wu, Jiarun and
Chen, Qingliang and
Xiao, Zeguan and
Gu, Yuliang and
Sun, Mengsi",
editor = "Carpuat, Marine and
de Marneffe, Marie-Catherine and
Meza Ruiz, Ivan Vladimir",
booktitle = "Findings of the Association for Computational Linguistics: NAACL 2022",
month = jul,
year = "2022",
address = "Seattle, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.findings-naacl.123",
doi = "10.18653/v1/2022.findings-naacl.123",
pages = "1632--1646",
abstract = "Pre-trained language models have shown great success in multiple downstream tasks. However, they are computationally expensive to fine-tune. Thus, transfer learning with adapter modules has been introduced to alleviate this problem, helping to extract knowledge of the downstream tasks. Adapterfusion models are an example of the transformers-with-adapter-modules, which merge multiple adapters to incorporate knowledge from different tasks. However, merging multiple adapters will inevitably cause redundancies, increasing the training and inference time massively. Therefore, in this paper, we propose an approach to identify the influence of each adapter module and a novel way to prune adapters based on the prestigious Lottery Ticket Hypothesis. Experiments on GLUE datasets show that the pruned Adapterfusion model with our scheme can achieve state-of-the-art results, reducing sizes significantly while keeping performance intact.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wu-etal-2022-pruning">
<titleInfo>
<title>Pruning Adatperfusion with Lottery Ticket Hypothesis</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jiarun</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qingliang</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zeguan</namePart>
<namePart type="family">Xiao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuliang</namePart>
<namePart type="family">Gu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mengsi</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2022-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: NAACL 2022</title>
</titleInfo>
<name type="personal">
<namePart type="given">Marine</namePart>
<namePart type="family">Carpuat</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marie-Catherine</namePart>
<namePart type="family">de Marneffe</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ivan</namePart>
<namePart type="given">Vladimir</namePart>
<namePart type="family">Meza Ruiz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Seattle, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Pre-trained language models have shown great success in multiple downstream tasks. However, they are computationally expensive to fine-tune. Thus, transfer learning with adapter modules has been introduced to alleviate this problem, helping to extract knowledge of the downstream tasks. Adapterfusion models are an example of the transformers-with-adapter-modules, which merge multiple adapters to incorporate knowledge from different tasks. However, merging multiple adapters will inevitably cause redundancies, increasing the training and inference time massively. Therefore, in this paper, we propose an approach to identify the influence of each adapter module and a novel way to prune adapters based on the prestigious Lottery Ticket Hypothesis. Experiments on GLUE datasets show that the pruned Adapterfusion model with our scheme can achieve state-of-the-art results, reducing sizes significantly while keeping performance intact.</abstract>
<identifier type="citekey">wu-etal-2022-pruning</identifier>
<identifier type="doi">10.18653/v1/2022.findings-naacl.123</identifier>
<location>
<url>https://aclanthology.org/2022.findings-naacl.123</url>
</location>
<part>
<date>2022-07</date>
<extent unit="page">
<start>1632</start>
<end>1646</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Pruning Adatperfusion with Lottery Ticket Hypothesis
%A Wu, Jiarun
%A Chen, Qingliang
%A Xiao, Zeguan
%A Gu, Yuliang
%A Sun, Mengsi
%Y Carpuat, Marine
%Y de Marneffe, Marie-Catherine
%Y Meza Ruiz, Ivan Vladimir
%S Findings of the Association for Computational Linguistics: NAACL 2022
%D 2022
%8 July
%I Association for Computational Linguistics
%C Seattle, United States
%F wu-etal-2022-pruning
%X Pre-trained language models have shown great success in multiple downstream tasks. However, they are computationally expensive to fine-tune. Thus, transfer learning with adapter modules has been introduced to alleviate this problem, helping to extract knowledge of the downstream tasks. Adapterfusion models are an example of the transformers-with-adapter-modules, which merge multiple adapters to incorporate knowledge from different tasks. However, merging multiple adapters will inevitably cause redundancies, increasing the training and inference time massively. Therefore, in this paper, we propose an approach to identify the influence of each adapter module and a novel way to prune adapters based on the prestigious Lottery Ticket Hypothesis. Experiments on GLUE datasets show that the pruned Adapterfusion model with our scheme can achieve state-of-the-art results, reducing sizes significantly while keeping performance intact.
%R 10.18653/v1/2022.findings-naacl.123
%U https://aclanthology.org/2022.findings-naacl.123
%U https://doi.org/10.18653/v1/2022.findings-naacl.123
%P 1632-1646
Markdown (Informal)
[Pruning Adatperfusion with Lottery Ticket Hypothesis](https://aclanthology.org/2022.findings-naacl.123) (Wu et al., Findings 2022)
ACL
- Jiarun Wu, Qingliang Chen, Zeguan Xiao, Yuliang Gu, and Mengsi Sun. 2022. Pruning Adatperfusion with Lottery Ticket Hypothesis. In Findings of the Association for Computational Linguistics: NAACL 2022, pages 1632–1646, Seattle, United States. Association for Computational Linguistics.