Synthetic Prefixes to Mitigate Bias in Real-Time Neural Query Autocomplete

TL;DR

The paper tackles presentation bias in real-time query autocomplete by augmenting training data with synthetic prefixes derived from general search logs, then trains a shallow neural learning-to-rank model under strict latency constraints. It combines real QAC engagement data with synthetic samples and employs an efficient listwise-to-pairwise loss with $O(n)$ complexity per event, achieving production-scale throughput. Offline and online evaluations show that the augmented data approach improves mean reciprocal rank ($\text{MRR}$) and user engagement, while maintaining low latency. The work demonstrates a scalable bias-mitigation strategy for low-latency ranking tasks and suggests applicability to related searches and query recommendations.

Abstract

We introduce a data-centric approach for mitigating presentation bias in real-time neural query autocomplete systems through the use of synthetic prefixes. These prefixes are generated from complete user queries collected during regular search sessions where autocomplete was not active. This allows us to enrich the training data for learning to rank models with more diverse and less biased examples. This method addresses the inherent bias in engagement signals collected from live query autocomplete interactions, where model suggestions influence user behavior. Our neural ranker is optimized for real-time deployment under strict latency constraints and incorporates a rich set of features, including query popularity, seasonality, fuzzy match scores, and contextual signals such as department affinity, device type, and vertical alignment with previous user queries. To support efficient training, we introduce a task-specific simplification of the listwise loss, reducing computational complexity from $O(n^2)$ to $O(n)$ by leveraging the query autocomplete structure of having only one ground-truth selection per prefix. Deployed in a large-scale e-commerce setting, our system demonstrates statistically significant improvements in user engagement, as measured by mean reciprocal rank and related metrics. Our findings show that synthetic prefixes not only improve generalization but also provide a scalable path toward bias mitigation in other low-latency ranking tasks, including related searches and query recommendations.

Figures (1)

• Figure 1: Model architecture of the Neural Ranker