@inproceedings{cheng-amiri-2024-fairflow,
title = "{F}air{F}low: Mitigating Dataset Biases through Undecided Learning for Natural Language Understanding",
author = "Cheng, Jiali and
Amiri, Hadi",
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.1225/",
doi = "10.18653/v1/2024.emnlp-main.1225",
pages = "21960--21975",
abstract = "Language models are prone to dataset biases, known as shortcuts and spurious correlations in data, which often result in performance drop on new data. We present a new debiasing framework called FairFlow that mitigates dataset biases by learning to be \textit{undecided} in its predictions for data samples or representations associated with known or unknown biases. The framework introduces two key components: a suite of data and model perturbation operations that generate different biased views of input samples, and a contrastive objective that learns debiased and robust representations from the resulting biased views of samples. Experiments show that FairFlow outperforms existing debiasing methods, particularly against out-of-domain and hard test samples without compromising the in-domain performance."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="cheng-amiri-2024-fairflow">
<titleInfo>
<title>FairFlow: Mitigating Dataset Biases through Undecided Learning for Natural Language Understanding</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jiali</namePart>
<namePart type="family">Cheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hadi</namePart>
<namePart type="family">Amiri</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>Language models are prone to dataset biases, known as shortcuts and spurious correlations in data, which often result in performance drop on new data. We present a new debiasing framework called FairFlow that mitigates dataset biases by learning to be undecided in its predictions for data samples or representations associated with known or unknown biases. The framework introduces two key components: a suite of data and model perturbation operations that generate different biased views of input samples, and a contrastive objective that learns debiased and robust representations from the resulting biased views of samples. Experiments show that FairFlow outperforms existing debiasing methods, particularly against out-of-domain and hard test samples without compromising the in-domain performance.</abstract>
<identifier type="citekey">cheng-amiri-2024-fairflow</identifier>
<identifier type="doi">10.18653/v1/2024.emnlp-main.1225</identifier>
<location>
<url>https://aclanthology.org/2024.emnlp-main.1225/</url>
</location>
<part>
<date>2024-11</date>
<extent unit="page">
<start>21960</start>
<end>21975</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T FairFlow: Mitigating Dataset Biases through Undecided Learning for Natural Language Understanding
%A Cheng, Jiali
%A Amiri, Hadi
%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 cheng-amiri-2024-fairflow
%X Language models are prone to dataset biases, known as shortcuts and spurious correlations in data, which often result in performance drop on new data. We present a new debiasing framework called FairFlow that mitigates dataset biases by learning to be undecided in its predictions for data samples or representations associated with known or unknown biases. The framework introduces two key components: a suite of data and model perturbation operations that generate different biased views of input samples, and a contrastive objective that learns debiased and robust representations from the resulting biased views of samples. Experiments show that FairFlow outperforms existing debiasing methods, particularly against out-of-domain and hard test samples without compromising the in-domain performance.
%R 10.18653/v1/2024.emnlp-main.1225
%U https://aclanthology.org/2024.emnlp-main.1225/
%U https://doi.org/10.18653/v1/2024.emnlp-main.1225
%P 21960-21975
Markdown (Informal)
[FairFlow: Mitigating Dataset Biases through Undecided Learning for Natural Language Understanding](https://aclanthology.org/2024.emnlp-main.1225/) (Cheng & Amiri, EMNLP 2024)
ACL