[2604.02215] Universal Hypernetworks for Arbitrary Models

Computer Science > Machine Learning arXiv:2604.02215 (cs) [Submitted on 2 Apr 2026] Title:Universal Hypernetworks for Arbitrary Models Authors:Xuanfeng Zhou View a PDF of the paper titled Universal Hypernetworks for Arbitrary Models, by Xuanfeng Zhou View PDF HTML (experimental) Abstract:Conventional hypernetworks are typically engineered around a specific base-model parameterization, so changing the target architecture often entails redesigning the hypernetwork and retraining it from scratch. We introduce the \emph{Universal Hypernetwork} (UHN), a fixed-architecture generator that predicts weights from deterministic parameter, architecture, and task descriptors. This descriptor-based formulation decouples the generator architecture from target-network parameterization, so one generator can instantiate heterogeneous models across the tested architecture and task families. Our empirical claims are threefold: (1) one fixed UHN remains competitive with direct training across vision, graph, text, and formula-regression benchmarks; (2) the same UHN supports both multi-model generalization within a family and multi-task learning across heterogeneous models; and (3) UHN enables stable recursive generation with up to three intermediate generated UHNs before the final base model. Our code is available at this https URL. Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI) Cite as: arXiv:2604.02215 [cs.LG]   (or arXiv:2604.02215v1 [cs.LG] for this version)   https://d...