@inproceedings{bhatt-etal-2024-experimental,
title = "An Experimental Design Framework for Label-Efficient Supervised Finetuning of Large Language Models",
author = "Bhatt, Gantavya and
Chen, Yifang and
Das, Arnav and
Zhang, Jifan and
Truong, Sang and
Mussmann, Stephen and
Zhu, Yinglun and
Bilmes, Jeff and
Du, Simon and
Jamieson, Kevin and
Ash, Jordan and
Nowak, Robert",
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",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.390",
doi = "10.18653/v1/2024.findings-acl.390",
pages = "6549--6560",
abstract = "Supervised finetuning (SFT) on instruction datasets has played a crucial role in achieving the remarkable zero-shot generalization capabilities observed in modern large language models (LLMs). However, the annotation efforts required to produce high quality responses for instructions are becoming prohibitively expensive, especially as the number of tasks spanned by instruction datasets continues to increase. Active learning is effective in identifying useful subsets of samples to annotate from an unlabeled pool, but its high computational cost remains a barrier to its widespread applicability in the context of LLMs. To mitigate the annotation cost of SFT and circumvent the computational bottlenecks of active learning, we propose using experimental design. Experimental design techniques select the most informative samples to label, and typically maximize some notion of uncertainty and/or diversity. In our work, we implement a framework that evaluates several existing and novel experimental design techniques and find that these methods consistently yield significant gains in label efficiency with little computational overhead. On generative tasks, to reach the same generalization performance, our methods save 50{\%} of the annotation cost compared to random sampling.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="bhatt-etal-2024-experimental">
<titleInfo>
<title>An Experimental Design Framework for Label-Efficient Supervised Finetuning of Large Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Gantavya</namePart>
<namePart type="family">Bhatt</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yifang</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arnav</namePart>
<namePart type="family">Das</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jifan</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sang</namePart>
<namePart type="family">Truong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Stephen</namePart>
<namePart type="family">Mussmann</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yinglun</namePart>
<namePart type="family">Zhu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jeff</namePart>
<namePart type="family">Bilmes</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Simon</namePart>
<namePart type="family">Du</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kevin</namePart>
<namePart type="family">Jamieson</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jordan</namePart>
<namePart type="family">Ash</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Robert</namePart>
<namePart type="family">Nowak</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</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Supervised finetuning (SFT) on instruction datasets has played a crucial role in achieving the remarkable zero-shot generalization capabilities observed in modern large language models (LLMs). However, the annotation efforts required to produce high quality responses for instructions are becoming prohibitively expensive, especially as the number of tasks spanned by instruction datasets continues to increase. Active learning is effective in identifying useful subsets of samples to annotate from an unlabeled pool, but its high computational cost remains a barrier to its widespread applicability in the context of LLMs. To mitigate the annotation cost of SFT and circumvent the computational bottlenecks of active learning, we propose using experimental design. Experimental design techniques select the most informative samples to label, and typically maximize some notion of uncertainty and/or diversity. In our work, we implement a framework that evaluates several existing and novel experimental design techniques and find that these methods consistently yield significant gains in label efficiency with little computational overhead. On generative tasks, to reach the same generalization performance, our methods save 50% of the annotation cost compared to random sampling.</abstract>
<identifier type="citekey">bhatt-etal-2024-experimental</identifier>
<identifier type="doi">10.18653/v1/2024.findings-acl.390</identifier>
<location>
<url>https://aclanthology.org/2024.findings-acl.390</url>
</location>
<part>
<date>2024-08</date>
<extent unit="page">
<start>6549</start>
<end>6560</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T An Experimental Design Framework for Label-Efficient Supervised Finetuning of Large Language Models
%A Bhatt, Gantavya
%A Chen, Yifang
%A Das, Arnav
%A Zhang, Jifan
%A Truong, Sang
%A Mussmann, Stephen
%A Zhu, Yinglun
%A Bilmes, Jeff
%A Du, Simon
%A Jamieson, Kevin
%A Ash, Jordan
%A Nowak, Robert
%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
%F bhatt-etal-2024-experimental
%X Supervised finetuning (SFT) on instruction datasets has played a crucial role in achieving the remarkable zero-shot generalization capabilities observed in modern large language models (LLMs). However, the annotation efforts required to produce high quality responses for instructions are becoming prohibitively expensive, especially as the number of tasks spanned by instruction datasets continues to increase. Active learning is effective in identifying useful subsets of samples to annotate from an unlabeled pool, but its high computational cost remains a barrier to its widespread applicability in the context of LLMs. To mitigate the annotation cost of SFT and circumvent the computational bottlenecks of active learning, we propose using experimental design. Experimental design techniques select the most informative samples to label, and typically maximize some notion of uncertainty and/or diversity. In our work, we implement a framework that evaluates several existing and novel experimental design techniques and find that these methods consistently yield significant gains in label efficiency with little computational overhead. On generative tasks, to reach the same generalization performance, our methods save 50% of the annotation cost compared to random sampling.
%R 10.18653/v1/2024.findings-acl.390
%U https://aclanthology.org/2024.findings-acl.390
%U https://doi.org/10.18653/v1/2024.findings-acl.390
%P 6549-6560
Markdown (Informal)
[An Experimental Design Framework for Label-Efficient Supervised Finetuning of Large Language Models](https://aclanthology.org/2024.findings-acl.390) (Bhatt et al., Findings 2024)
ACL
- Gantavya Bhatt, Yifang Chen, Arnav Das, Jifan Zhang, Sang Truong, Stephen Mussmann, Yinglun Zhu, Jeff Bilmes, Simon Du, Kevin Jamieson, Jordan Ash, and Robert Nowak. 2024. An Experimental Design Framework for Label-Efficient Supervised Finetuning of Large Language Models. In Findings of the Association for Computational Linguistics: ACL 2024, pages 6549–6560, Bangkok, Thailand. Association for Computational Linguistics.