Domain specificity and data efficiency in typo tolerant spell checkers: the case of search in online marketplaces

TL;DR

This paper introduces a data augmentation approach to improve typo correction in domain-specific search queries for online marketplaces, demonstrating effective real-time spelling correction with limited annotated data.

Contribution

It presents a novel data augmentation method and a domain-specific neural embedding approach for typo tolerant spell checking in online marketplace search.

Findings

Effective typo correction in real-time API for Microsoft AppSource

Synthetic data improves domain-specific spell checking accuracy

Approach reduces reliance on large annotated datasets

Abstract

Typographical errors are a major source of frustration for visitors of online marketplaces. Because of the domain-specific nature of these marketplaces and the very short queries users tend to search for, traditional spell cheking solutions do not perform well in correcting typos. We present a data augmentation method to address the lack of annotated typo data and train a recurrent neural network to learn context-limited domain-specific embeddings. Those embeddings are deployed in a real-time inferencing API for the Microsoft AppSource marketplace to find the closest match between a misspelled user query and the available product names. Our data efficient solution shows that controlled high quality synthetic data may be a powerful tool especially considering the current climate of large language models which rely on prohibitively huge and often uncontrolled datasets.