@inproceedings{gong-etal-2022-harmless,
title = "Harmless Transfer Learning for Item Embeddings",
author = "Gong, Chengyue and
Du, Xiaocong and
Choudhary, Dhruv and
Bhushanam, Bhargav and
Liu, Qiang and
Kejariwal, Arun",
editor = "Carpuat, Marine and
de Marneffe, Marie-Catherine and
Meza Ruiz, Ivan Vladimir",
booktitle = "Findings of the Association for Computational Linguistics: NAACL 2022",
month = jul,
year = "2022",
address = "Seattle, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.findings-naacl.38",
doi = "10.18653/v1/2022.findings-naacl.38",
pages = "504--516",
abstract = "Learning embedding layers (for classes, words, items, etc.) is a key component of lots of applications, ranging from natural language processing, recommendation systems to electronic health records, etc. However, the frequency of real-world items follows a long-tail distribution in these applications, causing naive training methods perform poorly on the rare items. A line of previous works address this problem by transferring the knowledge from the frequent items to rare items by introducing an auxiliary transfer loss. However, when defined improperly, the transfer loss may introduce harmful biases and deteriorate the performance. In this work, we propose a harmless transfer learning framework that limits the impact of the potential biases in both the definition and optimization of the transfer loss. On the definition side, we reduce the bias in transfer loss by focusing on the items to which information from high-frequency items can be efficiently transferred. On the optimization side, we leverage a lexicographic optimization framework to efficiently incorporate the information of the transfer loss without hurting the minimization of the main prediction loss function. Our method serves as a plug-in module and significantly boosts the performance on a variety of NLP and recommendation system tasks.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="gong-etal-2022-harmless">
<titleInfo>
<title>Harmless Transfer Learning for Item Embeddings</title>
</titleInfo>
<name type="personal">
<namePart type="given">Chengyue</namePart>
<namePart type="family">Gong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaocong</namePart>
<namePart type="family">Du</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dhruv</namePart>
<namePart type="family">Choudhary</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bhargav</namePart>
<namePart type="family">Bhushanam</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qiang</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arun</namePart>
<namePart type="family">Kejariwal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2022-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: NAACL 2022</title>
</titleInfo>
<name type="personal">
<namePart type="given">Marine</namePart>
<namePart type="family">Carpuat</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marie-Catherine</namePart>
<namePart type="family">de Marneffe</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ivan</namePart>
<namePart type="given">Vladimir</namePart>
<namePart type="family">Meza Ruiz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Seattle, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Learning embedding layers (for classes, words, items, etc.) is a key component of lots of applications, ranging from natural language processing, recommendation systems to electronic health records, etc. However, the frequency of real-world items follows a long-tail distribution in these applications, causing naive training methods perform poorly on the rare items. A line of previous works address this problem by transferring the knowledge from the frequent items to rare items by introducing an auxiliary transfer loss. However, when defined improperly, the transfer loss may introduce harmful biases and deteriorate the performance. In this work, we propose a harmless transfer learning framework that limits the impact of the potential biases in both the definition and optimization of the transfer loss. On the definition side, we reduce the bias in transfer loss by focusing on the items to which information from high-frequency items can be efficiently transferred. On the optimization side, we leverage a lexicographic optimization framework to efficiently incorporate the information of the transfer loss without hurting the minimization of the main prediction loss function. Our method serves as a plug-in module and significantly boosts the performance on a variety of NLP and recommendation system tasks.</abstract>
<identifier type="citekey">gong-etal-2022-harmless</identifier>
<identifier type="doi">10.18653/v1/2022.findings-naacl.38</identifier>
<location>
<url>https://aclanthology.org/2022.findings-naacl.38</url>
</location>
<part>
<date>2022-07</date>
<extent unit="page">
<start>504</start>
<end>516</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Harmless Transfer Learning for Item Embeddings
%A Gong, Chengyue
%A Du, Xiaocong
%A Choudhary, Dhruv
%A Bhushanam, Bhargav
%A Liu, Qiang
%A Kejariwal, Arun
%Y Carpuat, Marine
%Y de Marneffe, Marie-Catherine
%Y Meza Ruiz, Ivan Vladimir
%S Findings of the Association for Computational Linguistics: NAACL 2022
%D 2022
%8 July
%I Association for Computational Linguistics
%C Seattle, United States
%F gong-etal-2022-harmless
%X Learning embedding layers (for classes, words, items, etc.) is a key component of lots of applications, ranging from natural language processing, recommendation systems to electronic health records, etc. However, the frequency of real-world items follows a long-tail distribution in these applications, causing naive training methods perform poorly on the rare items. A line of previous works address this problem by transferring the knowledge from the frequent items to rare items by introducing an auxiliary transfer loss. However, when defined improperly, the transfer loss may introduce harmful biases and deteriorate the performance. In this work, we propose a harmless transfer learning framework that limits the impact of the potential biases in both the definition and optimization of the transfer loss. On the definition side, we reduce the bias in transfer loss by focusing on the items to which information from high-frequency items can be efficiently transferred. On the optimization side, we leverage a lexicographic optimization framework to efficiently incorporate the information of the transfer loss without hurting the minimization of the main prediction loss function. Our method serves as a plug-in module and significantly boosts the performance on a variety of NLP and recommendation system tasks.
%R 10.18653/v1/2022.findings-naacl.38
%U https://aclanthology.org/2022.findings-naacl.38
%U https://doi.org/10.18653/v1/2022.findings-naacl.38
%P 504-516
Markdown (Informal)
[Harmless Transfer Learning for Item Embeddings](https://aclanthology.org/2022.findings-naacl.38) (Gong et al., Findings 2022)
ACL
- Chengyue Gong, Xiaocong Du, Dhruv Choudhary, Bhargav Bhushanam, Qiang Liu, and Arun Kejariwal. 2022. Harmless Transfer Learning for Item Embeddings. In Findings of the Association for Computational Linguistics: NAACL 2022, pages 504–516, Seattle, United States. Association for Computational Linguistics.