Mitigating Pooling Bias in E-commerce Search via False Negative Estimation

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Wang, Xiaochen, Xiao, Xiao, Zhang, Ruhan, Zhang, Xuan, Na, Taesik, Tenneti, Tejaswi, Wang, Haixun, Ma, Fenglong
Format:	Preprint
Published:	2023
Subjects:	Information Retrieval Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

_version_	1866910568456126464
author	Wang, Xiaochen Xiao, Xiao Zhang, Ruhan Zhang, Xuan Na, Taesik Tenneti, Tejaswi Wang, Haixun Ma, Fenglong
author_facet	Wang, Xiaochen Xiao, Xiao Zhang, Ruhan Zhang, Xuan Na, Taesik Tenneti, Tejaswi Wang, Haixun Ma, Fenglong
contents	Efficient and accurate product relevance assessment is critical for user experiences and business success. Training a proficient relevance assessment model requires high-quality query-product pairs, often obtained through negative sampling strategies. Unfortunately, current methods introduce pooling bias by mistakenly sampling false negatives, diminishing performance and business impact. To address this, we present Bias-mitigating Hard Negative Sampling (BHNS), a novel negative sampling strategy tailored to identify and adjust for false negatives, building upon our original False Negative Estimation algorithm. Our experiments in the Instacart search setting confirm BHNS as effective for practical e-commerce use. Furthermore, comparative analyses on public dataset showcase its domain-agnostic potential for diverse applications.
format	Preprint
id	arxiv_https___arxiv_org_abs_2311_06444
institution	arXiv
publishDate	2023
record_format	arxiv
spellingShingle	Mitigating Pooling Bias in E-commerce Search via False Negative Estimation Wang, Xiaochen Xiao, Xiao Zhang, Ruhan Zhang, Xuan Na, Taesik Tenneti, Tejaswi Wang, Haixun Ma, Fenglong Information Retrieval Machine Learning Efficient and accurate product relevance assessment is critical for user experiences and business success. Training a proficient relevance assessment model requires high-quality query-product pairs, often obtained through negative sampling strategies. Unfortunately, current methods introduce pooling bias by mistakenly sampling false negatives, diminishing performance and business impact. To address this, we present Bias-mitigating Hard Negative Sampling (BHNS), a novel negative sampling strategy tailored to identify and adjust for false negatives, building upon our original False Negative Estimation algorithm. Our experiments in the Instacart search setting confirm BHNS as effective for practical e-commerce use. Furthermore, comparative analyses on public dataset showcase its domain-agnostic potential for diverse applications.
title	Mitigating Pooling Bias in E-commerce Search via False Negative Estimation
topic	Information Retrieval Machine Learning
url	https://arxiv.org/abs/2311.06444