@inproceedings{meissner-etal-2022-debiasing,
title = "Debiasing Masks: A New Framework for Shortcut Mitigation in {NLU}",
author = "Meissner, Johannes Mario and
Sugawara, Saku and
Aizawa, Akiko",
editor = "Goldberg, Yoav and
Kozareva, Zornitsa and
Zhang, Yue",
booktitle = "Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing",
month = dec,
year = "2022",
address = "Abu Dhabi, United Arab Emirates",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.emnlp-main.517",
doi = "10.18653/v1/2022.emnlp-main.517",
pages = "7607--7613",
abstract = "Debiasing language models from unwanted behaviors in Natural Language Understanding (NLU) tasks is a topic with rapidly increasing interest in the NLP community. Spurious statistical correlations in the data allow models to perform shortcuts and avoid uncovering more advanced and desirable linguistic features.A multitude of effective debiasing approaches has been proposed, but flexibility remains a major issue. For the most part, models must be retrained to find a new set of weights with debiased behavior.We propose a new debiasing method in which we identify debiased pruning masks that can be applied to a finetuned model. This enables the selective and conditional application of debiasing behaviors.We assume that bias is caused by a certain subset of weights in the network; our method is, in essence, a mask search to identify and remove biased weights.Our masks show equivalent or superior performance to the standard counterparts, while offering important benefits.Pruning masks can be stored with high efficiency in memory, and it becomes possible to switch among several debiasing behaviors (or revert back to the original biased model) at inference time. Finally, it opens the doors to further research on how biases are acquired by studying the generated masks. For example, we observed that the early layers and attention heads were pruned more aggressively, possibly hinting towards the location in which biases may be encoded.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="meissner-etal-2022-debiasing">
<titleInfo>
<title>Debiasing Masks: A New Framework for Shortcut Mitigation in NLU</title>
</titleInfo>
<name type="personal">
<namePart type="given">Johannes</namePart>
<namePart type="given">Mario</namePart>
<namePart type="family">Meissner</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Saku</namePart>
<namePart type="family">Sugawara</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Akiko</namePart>
<namePart type="family">Aizawa</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>Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing</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>Debiasing language models from unwanted behaviors in Natural Language Understanding (NLU) tasks is a topic with rapidly increasing interest in the NLP community. Spurious statistical correlations in the data allow models to perform shortcuts and avoid uncovering more advanced and desirable linguistic features.A multitude of effective debiasing approaches has been proposed, but flexibility remains a major issue. For the most part, models must be retrained to find a new set of weights with debiased behavior.We propose a new debiasing method in which we identify debiased pruning masks that can be applied to a finetuned model. This enables the selective and conditional application of debiasing behaviors.We assume that bias is caused by a certain subset of weights in the network; our method is, in essence, a mask search to identify and remove biased weights.Our masks show equivalent or superior performance to the standard counterparts, while offering important benefits.Pruning masks can be stored with high efficiency in memory, and it becomes possible to switch among several debiasing behaviors (or revert back to the original biased model) at inference time. Finally, it opens the doors to further research on how biases are acquired by studying the generated masks. For example, we observed that the early layers and attention heads were pruned more aggressively, possibly hinting towards the location in which biases may be encoded.</abstract>
<identifier type="citekey">meissner-etal-2022-debiasing</identifier>
<identifier type="doi">10.18653/v1/2022.emnlp-main.517</identifier>
<location>
<url>https://aclanthology.org/2022.emnlp-main.517</url>
</location>
<part>
<date>2022-12</date>
<extent unit="page">
<start>7607</start>
<end>7613</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Debiasing Masks: A New Framework for Shortcut Mitigation in NLU
%A Meissner, Johannes Mario
%A Sugawara, Saku
%A Aizawa, Akiko
%Y Goldberg, Yoav
%Y Kozareva, Zornitsa
%Y Zhang, Yue
%S Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing
%D 2022
%8 December
%I Association for Computational Linguistics
%C Abu Dhabi, United Arab Emirates
%F meissner-etal-2022-debiasing
%X Debiasing language models from unwanted behaviors in Natural Language Understanding (NLU) tasks is a topic with rapidly increasing interest in the NLP community. Spurious statistical correlations in the data allow models to perform shortcuts and avoid uncovering more advanced and desirable linguistic features.A multitude of effective debiasing approaches has been proposed, but flexibility remains a major issue. For the most part, models must be retrained to find a new set of weights with debiased behavior.We propose a new debiasing method in which we identify debiased pruning masks that can be applied to a finetuned model. This enables the selective and conditional application of debiasing behaviors.We assume that bias is caused by a certain subset of weights in the network; our method is, in essence, a mask search to identify and remove biased weights.Our masks show equivalent or superior performance to the standard counterparts, while offering important benefits.Pruning masks can be stored with high efficiency in memory, and it becomes possible to switch among several debiasing behaviors (or revert back to the original biased model) at inference time. Finally, it opens the doors to further research on how biases are acquired by studying the generated masks. For example, we observed that the early layers and attention heads were pruned more aggressively, possibly hinting towards the location in which biases may be encoded.
%R 10.18653/v1/2022.emnlp-main.517
%U https://aclanthology.org/2022.emnlp-main.517
%U https://doi.org/10.18653/v1/2022.emnlp-main.517
%P 7607-7613
Markdown (Informal)
[Debiasing Masks: A New Framework for Shortcut Mitigation in NLU](https://aclanthology.org/2022.emnlp-main.517) (Meissner et al., EMNLP 2022)
ACL