Saved in:
Bibliographic Details
Main Authors: Chen, Yun, Zhang, Moyu, Hu, Jinxin, Zhang, Yu, Zeng, Xiaoyi
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.20014
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908902835093504
author Chen, Yun
Zhang, Moyu
Hu, Jinxin
Zhang, Yu
Zeng, Xiaoyi
author_facet Chen, Yun
Zhang, Moyu
Hu, Jinxin
Zhang, Yu
Zeng, Xiaoyi
contents Neural Architecture Search (NAS) deployment in industrial production systems faces a fundamental validation bottleneck: verifying a single candidate architecture pi requires evaluating the deployed ensemble of M models, incurring prohibitive O(M) computational cost per candidate. This cost barrier severely limits architecture iteration frequency in real-world applications where ensembles (M=50-200) are standard for robustness. This work introduces Ensemble-Decoupled Architecture Search, a framework that leverages ensemble theory to predict system-level performance from single-learner evaluation. We establish the Ensemble-Decoupled Theory with a sufficient condition for monotonic ensemble improvement under homogeneity assumptions: a candidate architecture pi yields lower ensemble error than the current baseline if rho(pi) < rho(pi_old) - (M / (M - 1)) * (Delta E(pi) / sigma^2(pi)), where Delta E, rho, and sigma^2 are estimable from lightweight dual-learner training. This decouples architecture search from full ensemble training, reducing per-candidate search cost from O(M) to O(1) while maintaining O(M) deployment cost only for validated winners. We unify solution strategies across pipeline continuity: (1) closed-form optimization for tractable continuous pi (exemplified by feature bagging in CTR prediction), (2) constrained differentiable optimization for intractable continuous pi, and (3) LLM-driven search with iterative monotonic acceptance for discrete pi. The framework reveals two orthogonal improvement mechanisms -- base diversity gain and accuracy gain -- providing actionable design principles for industrial-scale NAS. All theoretical derivations are rigorous with detailed proofs deferred to the appendix. Comprehensive empirical validation will be included in the journal extension of this work.
format Preprint
id arxiv_https___arxiv_org_abs_2603_20014
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle AgenticRS-EnsNAS: Ensemble-Decoupled Self-Evolving Architecture Search
Chen, Yun
Zhang, Moyu
Hu, Jinxin
Zhang, Yu
Zeng, Xiaoyi
Machine Learning
Neural Architecture Search (NAS) deployment in industrial production systems faces a fundamental validation bottleneck: verifying a single candidate architecture pi requires evaluating the deployed ensemble of M models, incurring prohibitive O(M) computational cost per candidate. This cost barrier severely limits architecture iteration frequency in real-world applications where ensembles (M=50-200) are standard for robustness. This work introduces Ensemble-Decoupled Architecture Search, a framework that leverages ensemble theory to predict system-level performance from single-learner evaluation. We establish the Ensemble-Decoupled Theory with a sufficient condition for monotonic ensemble improvement under homogeneity assumptions: a candidate architecture pi yields lower ensemble error than the current baseline if rho(pi) < rho(pi_old) - (M / (M - 1)) * (Delta E(pi) / sigma^2(pi)), where Delta E, rho, and sigma^2 are estimable from lightweight dual-learner training. This decouples architecture search from full ensemble training, reducing per-candidate search cost from O(M) to O(1) while maintaining O(M) deployment cost only for validated winners. We unify solution strategies across pipeline continuity: (1) closed-form optimization for tractable continuous pi (exemplified by feature bagging in CTR prediction), (2) constrained differentiable optimization for intractable continuous pi, and (3) LLM-driven search with iterative monotonic acceptance for discrete pi. The framework reveals two orthogonal improvement mechanisms -- base diversity gain and accuracy gain -- providing actionable design principles for industrial-scale NAS. All theoretical derivations are rigorous with detailed proofs deferred to the appendix. Comprehensive empirical validation will be included in the journal extension of this work.
title AgenticRS-EnsNAS: Ensemble-Decoupled Self-Evolving Architecture Search
topic Machine Learning
url https://arxiv.org/abs/2603.20014