Saved in:
Bibliographic Details
Main Authors: Jaspal, Amit, Dang, Qian, Ramineni, Ajantha
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2506.07261
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916784821501952
author Jaspal, Amit
Dang, Qian
Ramineni, Ajantha
author_facet Jaspal, Amit
Dang, Qian
Ramineni, Ajantha
contents Modern large-scale recommender systems employ multi-stage ranking funnel (Retrieval, Pre-ranking, Ranking) to balance engagement and computational constraints (latency, CPU). However, the initial retrieval stage, often relying on efficient but less precise methods like K-Nearest Neighbors (KNN), struggles to effectively surface the most engaging items from billion-scale catalogs, particularly distinguishing highly relevant and engaging candidates from merely relevant ones. We introduce Recall Augmentation through Deferred Asynchronous Retrieval (RADAR), a novel framework that leverages asynchronous, offline computation to pre-rank a significantly larger candidate set for users using the full complexity ranking model. These top-ranked items are stored and utilized as a high-quality retrieval source during online inference, bypassing online retrieval and pre-ranking stages for these candidates. We demonstrate through offline experiments that RADAR significantly boosts recall (2X Recall@200 vs DNN retrieval baseline) by effectively combining a larger retrieved candidate set with a more powerful ranking model. Online A/B tests confirm a +0.8% lift in topline engagement metrics, validating RADAR as a practical and effective method to improve recommendation quality under strict online serving constraints.
format Preprint
id arxiv_https___arxiv_org_abs_2506_07261
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle RADAR: Recall Augmentation through Deferred Asynchronous Retrieval
Jaspal, Amit
Dang, Qian
Ramineni, Ajantha
Information Retrieval
Machine Learning
Modern large-scale recommender systems employ multi-stage ranking funnel (Retrieval, Pre-ranking, Ranking) to balance engagement and computational constraints (latency, CPU). However, the initial retrieval stage, often relying on efficient but less precise methods like K-Nearest Neighbors (KNN), struggles to effectively surface the most engaging items from billion-scale catalogs, particularly distinguishing highly relevant and engaging candidates from merely relevant ones. We introduce Recall Augmentation through Deferred Asynchronous Retrieval (RADAR), a novel framework that leverages asynchronous, offline computation to pre-rank a significantly larger candidate set for users using the full complexity ranking model. These top-ranked items are stored and utilized as a high-quality retrieval source during online inference, bypassing online retrieval and pre-ranking stages for these candidates. We demonstrate through offline experiments that RADAR significantly boosts recall (2X Recall@200 vs DNN retrieval baseline) by effectively combining a larger retrieved candidate set with a more powerful ranking model. Online A/B tests confirm a +0.8% lift in topline engagement metrics, validating RADAR as a practical and effective method to improve recommendation quality under strict online serving constraints.
title RADAR: Recall Augmentation through Deferred Asynchronous Retrieval
topic Information Retrieval
Machine Learning
url https://arxiv.org/abs/2506.07261