@inproceedings{garbaciauskas-etal-2024-choose,
title = "Choose Your Transformer: Improved Transferability Estimation of Transformer Models on Classification Tasks",
author = "Garbaciauskas, Lukas and
Ploner, Max and
Akbik, Alan",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.757",
pages = "12752--12768",
abstract = "There currently exists a multitude of pre-trained transformer language models (LMs) that are readily available. From a practical perspective, this raises the question of which pre-trained LM will perform best if fine-tuned for a specific downstream NLP task. However, exhaustively fine-tuning all available LMs to determine the best-fitting model is computationally infeasible. To address this problem, we present an approach that inexpensively estimates a ranking of the expected performance of a given set of candidate LMs for a given task. Following a layer-wise representation analysis, we extend existing approaches such as H-score and LogME by aggregating representations across all layers of the transformer model. We present an extensive analysis of 20 transformer LMs, 6 downstream NLP tasks, and various estimators (linear probing, kNN, H-score, and LogME). Our evaluation finds that averaging the layer representations significantly improves the Pearson correlation coefficient between the true model ranks and the estimate, increasing from 0.58 to 0.86 for LogME and from 0.65 to 0.88 for H-score.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="garbaciauskas-etal-2024-choose">
<titleInfo>
<title>Choose Your Transformer: Improved Transferability Estimation of Transformer Models on Classification Tasks</title>
</titleInfo>
<name type="personal">
<namePart type="given">Lukas</namePart>
<namePart type="family">Garbaciauskas</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Max</namePart>
<namePart type="family">Ploner</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alan</namePart>
<namePart type="family">Akbik</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2024-08</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics ACL 2024</title>
</titleInfo>
<name type="personal">
<namePart type="given">Lun-Wei</namePart>
<namePart type="family">Ku</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Andre</namePart>
<namePart type="family">Martins</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vivek</namePart>
<namePart type="family">Srikumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Bangkok, Thailand and virtual meeting</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>There currently exists a multitude of pre-trained transformer language models (LMs) that are readily available. From a practical perspective, this raises the question of which pre-trained LM will perform best if fine-tuned for a specific downstream NLP task. However, exhaustively fine-tuning all available LMs to determine the best-fitting model is computationally infeasible. To address this problem, we present an approach that inexpensively estimates a ranking of the expected performance of a given set of candidate LMs for a given task. Following a layer-wise representation analysis, we extend existing approaches such as H-score and LogME by aggregating representations across all layers of the transformer model. We present an extensive analysis of 20 transformer LMs, 6 downstream NLP tasks, and various estimators (linear probing, kNN, H-score, and LogME). Our evaluation finds that averaging the layer representations significantly improves the Pearson correlation coefficient between the true model ranks and the estimate, increasing from 0.58 to 0.86 for LogME and from 0.65 to 0.88 for H-score.</abstract>
<identifier type="citekey">garbaciauskas-etal-2024-choose</identifier>
<location>
<url>https://aclanthology.org/2024.findings-acl.757</url>
</location>
<part>
<date>2024-08</date>
<extent unit="page">
<start>12752</start>
<end>12768</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Choose Your Transformer: Improved Transferability Estimation of Transformer Models on Classification Tasks
%A Garbaciauskas, Lukas
%A Ploner, Max
%A Akbik, Alan
%Y Ku, Lun-Wei
%Y Martins, Andre
%Y Srikumar, Vivek
%S Findings of the Association for Computational Linguistics ACL 2024
%D 2024
%8 August
%I Association for Computational Linguistics
%C Bangkok, Thailand and virtual meeting
%F garbaciauskas-etal-2024-choose
%X There currently exists a multitude of pre-trained transformer language models (LMs) that are readily available. From a practical perspective, this raises the question of which pre-trained LM will perform best if fine-tuned for a specific downstream NLP task. However, exhaustively fine-tuning all available LMs to determine the best-fitting model is computationally infeasible. To address this problem, we present an approach that inexpensively estimates a ranking of the expected performance of a given set of candidate LMs for a given task. Following a layer-wise representation analysis, we extend existing approaches such as H-score and LogME by aggregating representations across all layers of the transformer model. We present an extensive analysis of 20 transformer LMs, 6 downstream NLP tasks, and various estimators (linear probing, kNN, H-score, and LogME). Our evaluation finds that averaging the layer representations significantly improves the Pearson correlation coefficient between the true model ranks and the estimate, increasing from 0.58 to 0.86 for LogME and from 0.65 to 0.88 for H-score.
%U https://aclanthology.org/2024.findings-acl.757
%P 12752-12768
Markdown (Informal)
[Choose Your Transformer: Improved Transferability Estimation of Transformer Models on Classification Tasks](https://aclanthology.org/2024.findings-acl.757) (Garbaciauskas et al., Findings 2024)
ACL