@inproceedings{agrawal-etal-2025-rationale,
title = "Rationale-Guided Distillation for {E}-Commerce Relevance Classification: Bridging Large Language Models and Lightweight Cross-Encoders",
author = "Agrawal, Sanjay and
Ahemad, Faizan and
Sembium, Vivek Varadarajan",
editor = "Rambow, Owen and
Wanner, Leo and
Apidianaki, Marianna and
Al-Khalifa, Hend and
Eugenio, Barbara Di and
Schockaert, Steven and
Darwish, Kareem and
Agarwal, Apoorv",
booktitle = "Proceedings of the 31st International Conference on Computational Linguistics: Industry Track",
month = jan,
year = "2025",
address = "Abu Dhabi, UAE",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.coling-industry.12/",
pages = "136--148",
abstract = "Accurately classifying the relevance of Query-Product pairs is critical in online retail stores such as Amazon, as displaying irrelevant products can harm user experience and reduce engagement. While Large Language Models (LLMs) excel at this task due to their broad knowledge and strong reasoning abilities. However, their high computational demands constrain their practical deployment in real-world applications. In this paper, we propose a novel distillation approach for e-commerce relevance classification that uses {\textquotedblleft}rationales{\textquotedblright} generated by LLMs to guide smaller cross-encoder models. These rationales capture key decision-making insights from LLMs, enhancing training efficiency and enabling the distillation to smaller cross-encoder models deployable in production without requiring the LLM. Our method achieves average ROC-AUC improvements of 1.4{\%} on 9 multilingual e-commerce datasets, 2.4{\%} on 3 ESCI datasets, and 6{\%} on GLUE datasets over vanilla cross-encoders. Our 110M parameter BERT model matches 7B parameter LLMs in performance ({\ensuremath{<}} 1{\%} ROC-AUC difference) while being 50 times faster per sample."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="agrawal-etal-2025-rationale">
<titleInfo>
<title>Rationale-Guided Distillation for E-Commerce Relevance Classification: Bridging Large Language Models and Lightweight Cross-Encoders</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sanjay</namePart>
<namePart type="family">Agrawal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Faizan</namePart>
<namePart type="family">Ahemad</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vivek</namePart>
<namePart type="given">Varadarajan</namePart>
<namePart type="family">Sembium</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-01</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 31st International Conference on Computational Linguistics: Industry Track</title>
</titleInfo>
<name type="personal">
<namePart type="given">Owen</namePart>
<namePart type="family">Rambow</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Leo</namePart>
<namePart type="family">Wanner</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marianna</namePart>
<namePart type="family">Apidianaki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hend</namePart>
<namePart type="family">Al-Khalifa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Barbara</namePart>
<namePart type="given">Di</namePart>
<namePart type="family">Eugenio</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Steven</namePart>
<namePart type="family">Schockaert</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kareem</namePart>
<namePart type="family">Darwish</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Apoorv</namePart>
<namePart type="family">Agarwal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Abu Dhabi, UAE</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Accurately classifying the relevance of Query-Product pairs is critical in online retail stores such as Amazon, as displaying irrelevant products can harm user experience and reduce engagement. While Large Language Models (LLMs) excel at this task due to their broad knowledge and strong reasoning abilities. However, their high computational demands constrain their practical deployment in real-world applications. In this paper, we propose a novel distillation approach for e-commerce relevance classification that uses “rationales” generated by LLMs to guide smaller cross-encoder models. These rationales capture key decision-making insights from LLMs, enhancing training efficiency and enabling the distillation to smaller cross-encoder models deployable in production without requiring the LLM. Our method achieves average ROC-AUC improvements of 1.4% on 9 multilingual e-commerce datasets, 2.4% on 3 ESCI datasets, and 6% on GLUE datasets over vanilla cross-encoders. Our 110M parameter BERT model matches 7B parameter LLMs in performance (\ensuremath< 1% ROC-AUC difference) while being 50 times faster per sample.</abstract>
<identifier type="citekey">agrawal-etal-2025-rationale</identifier>
<location>
<url>https://aclanthology.org/2025.coling-industry.12/</url>
</location>
<part>
<date>2025-01</date>
<extent unit="page">
<start>136</start>
<end>148</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Rationale-Guided Distillation for E-Commerce Relevance Classification: Bridging Large Language Models and Lightweight Cross-Encoders
%A Agrawal, Sanjay
%A Ahemad, Faizan
%A Sembium, Vivek Varadarajan
%Y Rambow, Owen
%Y Wanner, Leo
%Y Apidianaki, Marianna
%Y Al-Khalifa, Hend
%Y Eugenio, Barbara Di
%Y Schockaert, Steven
%Y Darwish, Kareem
%Y Agarwal, Apoorv
%S Proceedings of the 31st International Conference on Computational Linguistics: Industry Track
%D 2025
%8 January
%I Association for Computational Linguistics
%C Abu Dhabi, UAE
%F agrawal-etal-2025-rationale
%X Accurately classifying the relevance of Query-Product pairs is critical in online retail stores such as Amazon, as displaying irrelevant products can harm user experience and reduce engagement. While Large Language Models (LLMs) excel at this task due to their broad knowledge and strong reasoning abilities. However, their high computational demands constrain their practical deployment in real-world applications. In this paper, we propose a novel distillation approach for e-commerce relevance classification that uses “rationales” generated by LLMs to guide smaller cross-encoder models. These rationales capture key decision-making insights from LLMs, enhancing training efficiency and enabling the distillation to smaller cross-encoder models deployable in production without requiring the LLM. Our method achieves average ROC-AUC improvements of 1.4% on 9 multilingual e-commerce datasets, 2.4% on 3 ESCI datasets, and 6% on GLUE datasets over vanilla cross-encoders. Our 110M parameter BERT model matches 7B parameter LLMs in performance (\ensuremath< 1% ROC-AUC difference) while being 50 times faster per sample.
%U https://aclanthology.org/2025.coling-industry.12/
%P 136-148
Markdown (Informal)
[Rationale-Guided Distillation for E-Commerce Relevance Classification: Bridging Large Language Models and Lightweight Cross-Encoders](https://aclanthology.org/2025.coling-industry.12/) (Agrawal et al., COLING 2025)
ACL