@inproceedings{jimenez-gutierrez-etal-2022-thinking,
title = "Thinking about {GPT}-3 In-Context Learning for Biomedical {IE}? Think Again",
author = "Jimenez Gutierrez, Bernal and
McNeal, Nikolas and
Washington, Clayton and
Chen, You and
Li, Lang and
Sun, Huan and
Su, Yu",
editor = "Goldberg, Yoav and
Kozareva, Zornitsa and
Zhang, Yue",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2022",
month = dec,
year = "2022",
address = "Abu Dhabi, United Arab Emirates",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.findings-emnlp.329",
doi = "10.18653/v1/2022.findings-emnlp.329",
pages = "4497--4512",
abstract = "Large pre-trained language models (PLMs) such as GPT-3 have shown strong in-context learning capabilities, which are highly appealing for domains such as biomedicine that feature high and diverse demands of language technologies but also high data annotation costs. In this paper, we present the first systematic and comprehensive study to compare the few-shot performance of GPT-3 in-context learning with fine-tuning smaller (i.e., BERT-sized) PLMs on two representative biomedical information extraction (IE) tasks: named entity recognition and relation extraction. We follow the true few-shot setting to avoid overestimating models{'} few-shot performance by model selection over a large validation set. We also optimize GPT-3{'}s performance with known techniques such as contextual calibration and dynamic in-context example retrieval. However, our results show that GPT-3 still significantly underperforms compared to simply fine-tuning a smaller PLM. In addition, GPT-3 in-context learning also yields smaller gains in accuracy when more training data becomes available. More in-depth analyses further reveal issues of in-context learning that may be detrimental to IE tasks in general. Given the high cost of experimenting with GPT-3, we hope our study provides helpful guidance for biomedical researchers and practitioners towards more practical solutions such as fine-tuning small PLMs before better in-context learning is available for biomedical IE.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="jimenez-gutierrez-etal-2022-thinking">
<titleInfo>
<title>Thinking about GPT-3 In-Context Learning for Biomedical IE? Think Again</title>
</titleInfo>
<name type="personal">
<namePart type="given">Bernal</namePart>
<namePart type="family">Jimenez Gutierrez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nikolas</namePart>
<namePart type="family">McNeal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Clayton</namePart>
<namePart type="family">Washington</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">You</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lang</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Huan</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yu</namePart>
<namePart type="family">Su</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2022-12</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2022</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yoav</namePart>
<namePart type="family">Goldberg</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zornitsa</namePart>
<namePart type="family">Kozareva</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Abu Dhabi, United Arab Emirates</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Large pre-trained language models (PLMs) such as GPT-3 have shown strong in-context learning capabilities, which are highly appealing for domains such as biomedicine that feature high and diverse demands of language technologies but also high data annotation costs. In this paper, we present the first systematic and comprehensive study to compare the few-shot performance of GPT-3 in-context learning with fine-tuning smaller (i.e., BERT-sized) PLMs on two representative biomedical information extraction (IE) tasks: named entity recognition and relation extraction. We follow the true few-shot setting to avoid overestimating models’ few-shot performance by model selection over a large validation set. We also optimize GPT-3’s performance with known techniques such as contextual calibration and dynamic in-context example retrieval. However, our results show that GPT-3 still significantly underperforms compared to simply fine-tuning a smaller PLM. In addition, GPT-3 in-context learning also yields smaller gains in accuracy when more training data becomes available. More in-depth analyses further reveal issues of in-context learning that may be detrimental to IE tasks in general. Given the high cost of experimenting with GPT-3, we hope our study provides helpful guidance for biomedical researchers and practitioners towards more practical solutions such as fine-tuning small PLMs before better in-context learning is available for biomedical IE.</abstract>
<identifier type="citekey">jimenez-gutierrez-etal-2022-thinking</identifier>
<identifier type="doi">10.18653/v1/2022.findings-emnlp.329</identifier>
<location>
<url>https://aclanthology.org/2022.findings-emnlp.329</url>
</location>
<part>
<date>2022-12</date>
<extent unit="page">
<start>4497</start>
<end>4512</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Thinking about GPT-3 In-Context Learning for Biomedical IE? Think Again
%A Jimenez Gutierrez, Bernal
%A McNeal, Nikolas
%A Washington, Clayton
%A Chen, You
%A Li, Lang
%A Sun, Huan
%A Su, Yu
%Y Goldberg, Yoav
%Y Kozareva, Zornitsa
%Y Zhang, Yue
%S Findings of the Association for Computational Linguistics: EMNLP 2022
%D 2022
%8 December
%I Association for Computational Linguistics
%C Abu Dhabi, United Arab Emirates
%F jimenez-gutierrez-etal-2022-thinking
%X Large pre-trained language models (PLMs) such as GPT-3 have shown strong in-context learning capabilities, which are highly appealing for domains such as biomedicine that feature high and diverse demands of language technologies but also high data annotation costs. In this paper, we present the first systematic and comprehensive study to compare the few-shot performance of GPT-3 in-context learning with fine-tuning smaller (i.e., BERT-sized) PLMs on two representative biomedical information extraction (IE) tasks: named entity recognition and relation extraction. We follow the true few-shot setting to avoid overestimating models’ few-shot performance by model selection over a large validation set. We also optimize GPT-3’s performance with known techniques such as contextual calibration and dynamic in-context example retrieval. However, our results show that GPT-3 still significantly underperforms compared to simply fine-tuning a smaller PLM. In addition, GPT-3 in-context learning also yields smaller gains in accuracy when more training data becomes available. More in-depth analyses further reveal issues of in-context learning that may be detrimental to IE tasks in general. Given the high cost of experimenting with GPT-3, we hope our study provides helpful guidance for biomedical researchers and practitioners towards more practical solutions such as fine-tuning small PLMs before better in-context learning is available for biomedical IE.
%R 10.18653/v1/2022.findings-emnlp.329
%U https://aclanthology.org/2022.findings-emnlp.329
%U https://doi.org/10.18653/v1/2022.findings-emnlp.329
%P 4497-4512
Markdown (Informal)
[Thinking about GPT-3 In-Context Learning for Biomedical IE? Think Again](https://aclanthology.org/2022.findings-emnlp.329) (Jimenez Gutierrez et al., Findings 2022)
ACL