@inproceedings{agrawal-etal-2023-kd,
title = "{KD}-Boost: Boosting Real-Time Semantic Matching in {E}-commerce with Knowledge Distillation",
author = "Agrawal, Sanjay and
Sembium, Vivek and
M S, Ankith",
editor = "Wang, Mingxuan and
Zitouni, Imed",
booktitle = "Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing: Industry Track",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.emnlp-industry.14",
doi = "10.18653/v1/2023.emnlp-industry.14",
pages = "131--141",
abstract = "Real-time semantic matching is vital to web and product search. Transformer-based models have shown to be highly effective at encoding queries into an embedding space where semantically similar entities (queries or results) are in close proximity. However, the computational complexity of large transformer models limits their utilization for real-time matching. In this paper, we propose KD-Boost, a novel knowledge distillation algorithm designed for real-time semantic matching. KD-Boost trains low latency accurate student models by leveraging soft labels from a teacher model as well as ground truth via pairwise query-product and query-query signal derived from direct audits, user behavior, and taxonomy-based data using custom loss functions. Experiments on internal and external e-commerce datasets demonstrate an improvement of 2-3{\%} ROC-AUC compared to training student models directly, outperforming teacher and SOTA knowledge distillation benchmarks. Simulated online A/B tests using KD-Boost for automated Query Reformulation (QR) indicate a 6.31{\%} increase in query-to-query matching, 2.76{\%} increase in product coverage, and a 2.19{\%} improvement in relevance.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="agrawal-etal-2023-kd">
<titleInfo>
<title>KD-Boost: Boosting Real-Time Semantic Matching in E-commerce with Knowledge Distillation</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">Vivek</namePart>
<namePart type="family">Sembium</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ankith</namePart>
<namePart type="family">M S</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-12</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing: Industry Track</title>
</titleInfo>
<name type="personal">
<namePart type="given">Mingxuan</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Imed</namePart>
<namePart type="family">Zitouni</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Singapore</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Real-time semantic matching is vital to web and product search. Transformer-based models have shown to be highly effective at encoding queries into an embedding space where semantically similar entities (queries or results) are in close proximity. However, the computational complexity of large transformer models limits their utilization for real-time matching. In this paper, we propose KD-Boost, a novel knowledge distillation algorithm designed for real-time semantic matching. KD-Boost trains low latency accurate student models by leveraging soft labels from a teacher model as well as ground truth via pairwise query-product and query-query signal derived from direct audits, user behavior, and taxonomy-based data using custom loss functions. Experiments on internal and external e-commerce datasets demonstrate an improvement of 2-3% ROC-AUC compared to training student models directly, outperforming teacher and SOTA knowledge distillation benchmarks. Simulated online A/B tests using KD-Boost for automated Query Reformulation (QR) indicate a 6.31% increase in query-to-query matching, 2.76% increase in product coverage, and a 2.19% improvement in relevance.</abstract>
<identifier type="citekey">agrawal-etal-2023-kd</identifier>
<identifier type="doi">10.18653/v1/2023.emnlp-industry.14</identifier>
<location>
<url>https://aclanthology.org/2023.emnlp-industry.14</url>
</location>
<part>
<date>2023-12</date>
<extent unit="page">
<start>131</start>
<end>141</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T KD-Boost: Boosting Real-Time Semantic Matching in E-commerce with Knowledge Distillation
%A Agrawal, Sanjay
%A Sembium, Vivek
%A M S, Ankith
%Y Wang, Mingxuan
%Y Zitouni, Imed
%S Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing: Industry Track
%D 2023
%8 December
%I Association for Computational Linguistics
%C Singapore
%F agrawal-etal-2023-kd
%X Real-time semantic matching is vital to web and product search. Transformer-based models have shown to be highly effective at encoding queries into an embedding space where semantically similar entities (queries or results) are in close proximity. However, the computational complexity of large transformer models limits their utilization for real-time matching. In this paper, we propose KD-Boost, a novel knowledge distillation algorithm designed for real-time semantic matching. KD-Boost trains low latency accurate student models by leveraging soft labels from a teacher model as well as ground truth via pairwise query-product and query-query signal derived from direct audits, user behavior, and taxonomy-based data using custom loss functions. Experiments on internal and external e-commerce datasets demonstrate an improvement of 2-3% ROC-AUC compared to training student models directly, outperforming teacher and SOTA knowledge distillation benchmarks. Simulated online A/B tests using KD-Boost for automated Query Reformulation (QR) indicate a 6.31% increase in query-to-query matching, 2.76% increase in product coverage, and a 2.19% improvement in relevance.
%R 10.18653/v1/2023.emnlp-industry.14
%U https://aclanthology.org/2023.emnlp-industry.14
%U https://doi.org/10.18653/v1/2023.emnlp-industry.14
%P 131-141
Markdown (Informal)
[KD-Boost: Boosting Real-Time Semantic Matching in E-commerce with Knowledge Distillation](https://aclanthology.org/2023.emnlp-industry.14) (Agrawal et al., EMNLP 2023)
ACL