Saved in:
Bibliographic Details
Main Authors: Di, Jieming, Chen, Xiaoyu, She, Ying, Wang, Siyu, Liu, Lizzie, Wu, Fenggang, Mu, Jiaoying, Tsui, Tony, Elroumy, Amr, Tang, Hsing, Jiang, Zewei, Yang, Qiao, Qi, Lin, Lin, Haibo, Cui, Weifeng, Li, Daniel, Gupta, Kapil, Singh, Shivendra Pratap, Zheng, Jie, Overwijk, Arnold, Leng, Ling, Reddy, Sri, Malkin, Robert, Liu, Rocky
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.00324
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910183371833344
author Di, Jieming
Chen, Xiaoyu
She, Ying
Wang, Siyu
Liu, Lizzie
Wu, Fenggang
Mu, Jiaoying
Tsui, Tony
Elroumy, Amr
Tang, Hsing
Jiang, Zewei
Yang, Qiao
Qi, Lin
Lin, Haibo
Cui, Weifeng
Li, Daniel
Gupta, Kapil
Singh, Shivendra Pratap
Zheng, Jie
Overwijk, Arnold
Leng, Ling
Reddy, Sri
Malkin, Robert
Liu, Rocky
author_facet Di, Jieming
Chen, Xiaoyu
She, Ying
Wang, Siyu
Liu, Lizzie
Wu, Fenggang
Mu, Jiaoying
Tsui, Tony
Elroumy, Amr
Tang, Hsing
Jiang, Zewei
Yang, Qiao
Qi, Lin
Lin, Haibo
Cui, Weifeng
Li, Daniel
Gupta, Kapil
Singh, Shivendra Pratap
Zheng, Jie
Overwijk, Arnold
Leng, Ling
Reddy, Sri
Malkin, Robert
Liu, Rocky
contents Large-scale ranking systems depend on thousands of features derived from user behavior across multiple time horizons. Typically requires model retraining -- resulting in long iteration cycles (3--6 months), substantial GPU resource consumption, and limited rollout throughput. We introduce Intelligent Elastic Feature Fading (IEFF), a production infrastructure system that enables retrain-free feature efficiency rollouts by elastically controlling feature coverage and distribution at serving time. IEFF supports incremental feature coverage adjustments while models adapt through recurring training, eliminating dependencies on explicit retraining cycles. The system incorporates strict safety guardrails, reversibility mechanisms, and comprehensive monitoring to ensure stability at scale. Across multiple production use cases, IEFF accelerates efficiency-related rollouts by 5$\times$, eliminates retraining-related GPU overhead, and enables faster capacity recycling. Extensive offline and online experiments demonstrate that gradual feature fading prevents 50--55\% of online performance degradation compared to abrupt feature removal, while maintaining stable model behavior. These results establish elastic, system-level feature fading as a practical and scalable approach for managing feature efficiency in modern industrial ranking systems.
format Preprint
id arxiv_https___arxiv_org_abs_2605_00324
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Intelligent Elastic Feature Fading: Enabling Model Retrain-Free Feature Efficiency Rollouts at Scale
Di, Jieming
Chen, Xiaoyu
She, Ying
Wang, Siyu
Liu, Lizzie
Wu, Fenggang
Mu, Jiaoying
Tsui, Tony
Elroumy, Amr
Tang, Hsing
Jiang, Zewei
Yang, Qiao
Qi, Lin
Lin, Haibo
Cui, Weifeng
Li, Daniel
Gupta, Kapil
Singh, Shivendra Pratap
Zheng, Jie
Overwijk, Arnold
Leng, Ling
Reddy, Sri
Malkin, Robert
Liu, Rocky
Information Retrieval
Machine Learning
Large-scale ranking systems depend on thousands of features derived from user behavior across multiple time horizons. Typically requires model retraining -- resulting in long iteration cycles (3--6 months), substantial GPU resource consumption, and limited rollout throughput. We introduce Intelligent Elastic Feature Fading (IEFF), a production infrastructure system that enables retrain-free feature efficiency rollouts by elastically controlling feature coverage and distribution at serving time. IEFF supports incremental feature coverage adjustments while models adapt through recurring training, eliminating dependencies on explicit retraining cycles. The system incorporates strict safety guardrails, reversibility mechanisms, and comprehensive monitoring to ensure stability at scale. Across multiple production use cases, IEFF accelerates efficiency-related rollouts by 5$\times$, eliminates retraining-related GPU overhead, and enables faster capacity recycling. Extensive offline and online experiments demonstrate that gradual feature fading prevents 50--55\% of online performance degradation compared to abrupt feature removal, while maintaining stable model behavior. These results establish elastic, system-level feature fading as a practical and scalable approach for managing feature efficiency in modern industrial ranking systems.
title Intelligent Elastic Feature Fading: Enabling Model Retrain-Free Feature Efficiency Rollouts at Scale
topic Information Retrieval
Machine Learning
url https://arxiv.org/abs/2605.00324