Beyond Model Base Retrieval: Weaving Knowledge to Master Fine-grained Neural Network Design

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Wang, Jialiang, Liu, Hanmo, Di, Shimin, Wang, Zhili, Wang, Jiachuan, Chen, Lei, Zhou, Xiaofang
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911735262216192
author Wang, Jialiang
Liu, Hanmo
Di, Shimin
Wang, Zhili
Wang, Jiachuan
Chen, Lei
Zhou, Xiaofang
author_facet Wang, Jialiang
Liu, Hanmo
Di, Shimin
Wang, Zhili
Wang, Jiachuan
Chen, Lei
Zhou, Xiaofang
contents Designing high-performance neural networks for new tasks requires balancing optimization quality with search efficiency. Current methods fail to achieve this balance: neural architectural search is computationally expensive, while model retrieval often yields suboptimal static checkpoints. To resolve this dilemma, we model the performance gains induced by fine-grained architectural modifications as edit-effect evidence and build evidence graphs from prior tasks. By constructing a retrieval-augmented model refinement framework, our proposed M-DESIGN dynamically weaves historical evidence to discover near-optimal modification paths. M-DESIGN features an adaptive retrieval mechanism that quickly calibrates the evolving transferability of edit-effect evidence from different sources. To handle out-of-distribution shifts, we introduce predictive task planners that extrapolate gains from multi-hop evidence, thereby reducing reliance on an exhaustive repository. Based on our model knowledge base of 67,760 graph neural networks across 22 datasets, extensive experiments demonstrate that M-DESIGN consistently outperforms baselines, achieving the search-space best performance in 26 out of 33 cases under a strict budget.
format Preprint
id arxiv_https___arxiv_org_abs_2507_15336
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Beyond Model Base Retrieval: Weaving Knowledge to Master Fine-grained Neural Network Design
Wang, Jialiang
Liu, Hanmo
Di, Shimin
Wang, Zhili
Wang, Jiachuan
Chen, Lei
Zhou, Xiaofang
Machine Learning
Artificial Intelligence
Databases
Designing high-performance neural networks for new tasks requires balancing optimization quality with search efficiency. Current methods fail to achieve this balance: neural architectural search is computationally expensive, while model retrieval often yields suboptimal static checkpoints. To resolve this dilemma, we model the performance gains induced by fine-grained architectural modifications as edit-effect evidence and build evidence graphs from prior tasks. By constructing a retrieval-augmented model refinement framework, our proposed M-DESIGN dynamically weaves historical evidence to discover near-optimal modification paths. M-DESIGN features an adaptive retrieval mechanism that quickly calibrates the evolving transferability of edit-effect evidence from different sources. To handle out-of-distribution shifts, we introduce predictive task planners that extrapolate gains from multi-hop evidence, thereby reducing reliance on an exhaustive repository. Based on our model knowledge base of 67,760 graph neural networks across 22 datasets, extensive experiments demonstrate that M-DESIGN consistently outperforms baselines, achieving the search-space best performance in 26 out of 33 cases under a strict budget.
title Beyond Model Base Retrieval: Weaving Knowledge to Master Fine-grained Neural Network Design
topic Machine Learning
Artificial Intelligence
Databases
url https://arxiv.org/abs/2507.15336