Semantic Ads Retrieval at Walmart eCommerce with Language Models Progressively Trained on Multiple Knowledge Domains

TL;DR

This paper presents a comprehensive approach to improve Walmart's e-commerce sponsored search by leveraging advanced language models, specialized training techniques, and human-in-the-loop methods, resulting in significant relevance and revenue gains.

Contribution

The paper introduces a novel end-to-end ads retrieval system using a pretrained BERT-like model, Siamese network structure, and progressive training, tailored for large-scale e-commerce search.

Findings

Search relevance improved by up to 16%.

Online A/B tests show increased ad revenue.

Enhanced understanding of product semantics.

Abstract

Sponsored search in e-commerce poses several unique and complex challenges. These challenges stem from factors such as the asymmetric language structure between search queries and product names, the inherent ambiguity in user search intent, and the vast volume of sparse and imbalanced search corpus data. The role of the retrieval component within a sponsored search system is pivotal, serving as the initial step that directly affects the subsequent ranking and bidding systems. In this paper, we present an end-to-end solution tailored to optimize the ads retrieval system on Walmart.com. Our approach is to pretrain the BERT-like classification model with product category information, enhancing the model's understanding of Walmart product semantics. Second, we design a two-tower Siamese Network structure for embedding structures to augment training efficiency. Third, we introduce a Human-in-the-loop Progressive Fusion Training method to ensure robust model performance. Our results demonstrate the effectiveness of this pipeline. It enhances the search relevance metric by up to 16% compared to a baseline DSSM-based model. Moreover, our large-scale online A/B testing demonstrates that our approach surpasses the ad revenue of the existing production model.